Kenya - World Banksiteresources.worldbank.org/.../FINAL_ICA_Kenya_edited.doc · Web viewFigure 1 3 Contribution to GDP Growth by Institutional Sector and Main Activities, Average

Kenya

Investment Climate Assessment

June 2008

World BankRegional Program for Enterprise Development (RPED)

Africa Finance and Private Sector (AFTFP)

TABLE OF CONTENTSListing of Figures and Tables....................................................................................4

Acronyms.........................................................................................................................8Acknowledgments.......................................................................................................24Overview.........................................................................................................................251 Macro Environment............................................................................................302 Competitiveness of Kenyan Firms................................................................39

2.1 Overview........................................................................................................................392.2 Sample Composition......................................................................................................402.3 Labor Productivity.........................................................................................................402.4 Unit Labor Costs............................................................................................................422.5 Total Factor Productivity...............................................................................................432.6 Food Sector....................................................................................................................462.7 Textiles and Garments...................................................................................................48

3 Business Climate.................................................................................................523.1 Introduction....................................................................................................................523.2 Tax Rates.......................................................................................................................583.3 Corruption......................................................................................................................603.4 Electricity.......................................................................................................................683.5 Transportation and Customs..........................................................................................713.6 Crime.............................................................................................................................753.7 Tax Administration........................................................................................................793.8 Business Licensing and Permits....................................................................................81

4 Access to Finance...............................................................................................874.1 Access to Finance from an International Perspective....................................................874.2 Effect of Size on Access to Credit.................................................................................964.3 Characteristics of Loan Products...................................................................................984.4 Loan Applications and Rejections.................................................................................994.5 Conclusion...................................................................................................................100

5 Labor Markets and Human Capital.............................................................1025.1 Worker Skills...............................................................................................................1035.2 Labor Regulation.........................................................................................................1085.3 Wages..........................................................................................................................110

5.3.1 Cross-Country Comparisons................................................................................1105.3.2 Comparisons Across Firms in Kenya..................................................................112

5.4 Absenteeism.................................................................................................................1146 Microenterprises in Kenya.............................................................................117

6.1 Registration Characteristics.........................................................................................1176.2 Investment Climate for Microenterprises....................................................................118

7 Investment, Innovation, and Exports........................................................1247.1 Introduction..................................................................................................................1247.2 Who Invests, and How Much? Who Innovates?.........................................................1247.3 Who Exports, and How Much? What Are the Main Bottlenecks?..............................1337.4 Conclusions..................................................................................................................137

8 Recommendations............................................................................................139References...................................................................................................................151

2

Technical Appendix.........................................................................................................153Appendix 1: Access to finance................................................................................................153Appendix 2: Labor markets and human capital.......................................................................155Appendix 3: Enterprise Survey in Kenya: Sample Survey Design.........................................160

3

Listing of Figures and TablesFIGURES

Figure 1-1 Trends in Public and Private GDP Growth/Private Share in Total GDP,....................30Figure 1-2 Annual GDP Growth, Main Activities, 1996–2006 and 2001–06...............................31Figure 1-3 Contribution to GDP Growth by Institutional Sector and Main Activities, Average 2001–05.........................................................................................................................................32Figure 1-4 Herfindahl-Hirschmann Index of Product and Country–.............................................33Figure 1-5 Kenya: Export Composition by Destination Regions, Decade Averages....................34Figure 1-6 Real Lending Rates and Domestic Credit (Private and Rest)......................................34Figure 1-7 Real Gross Fixed Capital Formation, by Type of Assets, 1995–2005.........................36Figure 1-8 Trends in Rate of Change in Private and Public Wage................................................37Figure 2-1 Cross-Country Comparison of Labor Productivity......................................................41Figure 2-2 Unit Labor Costs..........................................................................................................43Figure 2-3 Total Factor Productivity Relative to South Africa.....................................................46Figure 2-4 Biggest Reported Constraints—Food Sector [[please add callout]]............................48Figure 2-5 Biggest Constraint: Textile and Garments Firms[[add text callout]]...........................51Figure 3-1 Top-Ranked Constraints by Labor Growth and Labor Productivity...........................54Figure 3-2 Indirect Costs in 2003 and 2007—Kenya Manufacturing Sector................................57Figure 3-3 Firms Reporting Tax Rate As Major or Very Severe Problem (%).............................58Figure 3-4 Total Amount of Taxes As Percentage of Profit: International Comparison...............59Figure 3-5 Breakdown of Taxes: International Comparison.........................................................60Figure 3-6 Firms Perceiving Corruption As a Severe or Major Constraint: International Comparison (%).............................................................................................................................61Figure 3-7 Kenya—Evolution of Transparency International Corruption Rating.........................62Figure 3-8 Transparency International Corruption Rating............................................................62Figure 3-9 Bribes in Public Procurement (Percent of Contract Value).........................................63Figure 3-10 Bribe Requests from Tax Inspectors—Cross-Country Comparison..........................64Figure 3-11 Percentage of Firms Requesting Licenses from Local and Central Government......65Figure 3-12 Bribes, Licenses, and Utilities: Share of Firms Requesting Informal Payments.......65Figure 3-13 Courts Malfunctioning: Percent of Firms Considering the Court System Efficient.66Figure 3-14 Court Procedures and Cost—International Comparison............................................67Figure 3-15: Time and Cost to Close a Business—International Comparison..............................67Figure 3-16 Share of Firms That Experienced Losses from Electrical Outages...........................68Figure 3-17 Sales Lost Due to Power Outages (Percent)—International Comparison.................70Figure 3-18 Days to Obtain an Electricity Connection—International Comparison....................70Figure 3-19 Inventory Holdings—International Comparison.......................................................71Figure 3-20 Inland Transportation Costs ($).................................................................................72Figure 3-21 Production Lost While in Transit—International Comparison..................................73Figure 3-22 Transportation Losses by Firm Characteristics..........................................................73Figure 3-24 Customs Costs ($)......................................................................................................74Figure 3-25 Percentage of Firms Reporting Crime As a Major or Very Severe Obstacle to Business —International Comparison...........................................................................................75Figure 3-26 Crime Costs—International Comparison (Percent of Sales).....................................76Figure 3-27 Cost of Security Services in Percent of Sales—International Comparison...............77

4

Figure 3-28 Total Costs of Crime—International Comparison (Percent).....................................77Figure 3-29 Impact of Security on Business Decisions (Percent of Firms)..................................78Figure 3-30 Time Spent in Dealing with Tax Officials—International........................................79Figure 3-31 Paying Taxes—Cross-Country Comparisons............................................................80Figure 3-32 Percentage of Firms Complaining About Business Licensing and Permits—..........81Figure 3-33 Duration (in Days), Cost (Percent of Income per Capita) and Number of Procedures Required to Get Business Licenses—International Comparison...................................................82Figure 3-34 Manager’s Time Spent Dealing with Regulations.....................................................82Figure 3-35 Duration (in Days) and Cost (Percent of Gross National Income per Capita) to Start a Business—..................................................................................................................................83Figure 3-36 Average Time to Renew a Business License—Kenya..............................................84Figure 3-37 Trade Documents Preparation Costs ($)....................................................................84Figure 3-38 Usage and Cost of Facilitators When Dealing with Licenses....................................85Figure 4-1: Perceptions About Access to Finance Are More Positive in Namibia Than in Many of the Comparator Countries, Especially Those in Southern African Customs Union.................87Figure 4-2 Nominal Cost of Borrowing Has Fallen Over the Last Three Years...........................88Figure 4-3 Manufacturing Firms in Kenya Use Trade Credit to Finance a Greater Share of Working Capital Than All Other Comparator Countries..............................................................90Figure 4-4 Manufacturing Firms in Kenya Are More Dependent on Bank Debt—Relative to Comparators—to Finance New Investment...................................................................................91Figure 4-5 The Median Annual Real Cost of Borrowing in Kenya Is Low and Loan Duration Terms Are Reasonable...................................................................................................................93Figure 4-6 Collateral Requirements...............................................................................................94Figure 4-7 Both Subjective and Objective Indicators Suggest That Access to Credit Is a More Serious Obstacle for Micro- And Small Enterprises Than for Medium-Sized and Large Enterprises.....................................................................................................................................96Figure 5-1 Manufacturing Firms Are at the Bottom of the Pack Concerning Perception of Skills in the Labor Force........................................................................................................................102Figure 5-2 Percent of Firms Providing Training/Percent Workers Trained................................105Figure 5-3 Manufacturing Firms in Kenya Are in the Middle of the Pack Regarding Labor Regulations..................................................................................................................................108Figure 5-4 Labor Regulation Does Not Appear To Be Particularly Burdensome in Kenya.......109Figure 5-5 Median Monthly Wages for Production Workers Are Higher in Kenya Than They Are in China and India and the Other East African Economies.........................................................110Figure 5-6 Median Monthly Wages in the Food and Garments Sectors.....................................111Figure 5-7 Unionization Rates in Kenya Are Moderate..............................................................113Figure 5-8: Absenteeism Is Lower Than in the Rest of East Africa but Higher Than in South Africa...........................................................................................................................................114Figure 5-9: The Disease Environment and Changes in Worker Absenteeism[[please add callout]].....................................................................................................................................................115Figure 6-1 Business Constraints: Percent Ranking Problem To Be Major or Severe.................117Figure 6-2 The Shares of Bank Debt and Trade Credit Are Increasing Functions of Firm Size; Use of Internal Resources Is a Declining Function of Firm Size................................................118Figure 6-3 Microenterprise Financial Characteristics.................................................................119Figure 6-4 Percentage of Income Reported for Tax Purposes: Microenterprises.......................121Figure 6-5 Perceived Reasons for Choosing Informality............................................................121

5

Figure 6-6 Median Number of Visits/Required Meetings with Tax Officials per Year..............122Figure 7-1 Sources of Technological Innovation........................................................................125Figure 7-2 Investment and Lack of Finance................................................................................130Figure 7-3 Investment and Macroeconomic Instability...............................................................130

TABLESTable 1-1 Ratio of Gross Fixed Capital Formation to GDP,.........................................................35Table 1-2 Gross Capital Formation Real Growth and Structure...................................................36Table 1-3 Kenya—Real Average Annual Wage per Employee (KSh of 2000)............................37Table 2-1 Sample Distribution in Kenya, by Sector and Location................................................40Table 2-2 Productivity Indicators..................................................................................................42Table 2-3 Pooled Value-Added Production Functions: ICA Survey 2003....................................44Table 2-4 Total Factor Productivity: Panel Estimation.................................................................45Table 2-5 Food Sector: Sample Characteristics (%)......................................................................47Table 2-6 Food Sector: Technology Characteristics.....................................................................47Table 2-7 Food Sector Performance Characteristics: Median Performance Measures.................48Table 2-8 Textile and Garments Sector: Sample Characteristics (%)...........................................49Table 2-9 Textile and Garments Sector: Technology Characteristics...........................................50Table 2-10 Textile and Garments: Performance Characteristics...................................................50Table 3-1 Firms Reporting Major or Very Severe Business Constraints:.....................................52Table 3-2 Ranking and Rating of Business Constraints in Kenya (% of Firms)...........................53Table 3-3 Kenyan Firms in Manufacturing Sector Reporting Major or Very Severe Constraints,.......................................................................................................................................................55Table 3-4 Firms Reporting Major or Very Severe Constraints: International Comparison.........56Table 3-5 Indirect Costs: All Formal Sectors, Kenya, 2007 (%)...................................................56Table 3-6 Indirect Costs, All Formal Firms: International Comparison (%)................................57Table 3-7 Frequency and Duration of Power Outages and Power Generator Ownership in Kenya.......................................................................................................................................................69Table 3-8 Power Outages and Usage of Electrical Generators......................................................69Table 4-1 Median Interest Rates and Loan Duration by Firm Size...............................................98Table 4-2 Credit Line/Loan Providers...........................................................................................98Table 4-3 Loan Characteristics.....................................................................................................98Table 4-4 Reasons for Loan Rejections.........................................................................................99Table 4-5 Reasons for Not Applying for a Loan or Line of Credit.............................................100Table 5-1 Percent of Firms Reporting Skills Shortage as a Major or Severe Constraint...........104Table 5-2 Do Reports of Skill Constraints Vary by Worker Education......................................105Table 5-3 Do Reports of Skill Constraints Vary By Training or Employment Growth?............105Table 5-4 Percent of Firms Saying that the Average Worker in the Firm Has..........................106Table 5-5 Firm-Based Training: Prevalence and Percent of Workers Trained...........................108Table 5-6 Median Monthly Wages by Occupation in 2005 U.S. Dollars....................................113Table 6-1 Probit Results: Likelihood of Having Access to Banking Sector...............................121Table 7-1 Summary Statistics: Full Sample and by Sector.........................................................125Table 7-2 Who Invests? Who Innovates?....................................................................................127Table 7-3 Investment and Innovation Regressions......................................................................128Table 7-4 Objective Investment Climate Constraints and Investment........................................132Table 7-5 Investment Regressions Based on Pooled 2002 and 2006 Data..................................133

6

Table 7-6 Who Exports? (Manufacturing Only).........................................................................134Table 7-7 Characteristics of Exporters (Manufacturing Only)....................................................135Table 7-8 Modeling the Decision to Export (Manufacturing Only)............................................136Table 7-9 Objective Investment Climate Constraints and Exports.............................................137

Appendix Table 1 Correlates of Access to Finance................................................................154Appendix Table 2 Training Determinants: Firm Level[[please provide callouts for all tables]] 155Appendix Table 3 Training Determinants: Individual Level......................................................156Appendix Table 4 Determinants of Average Wages: Firm Level...............................................157Appendix Table 5 Determinants of Worker Earnings.................................................................158Appendix Table 6 Population Size by Stratum and Sampling Region........................................161Appendix Table 7 Target Sample Size by Stratum and Sampling Region..................................162Appendix Table 8 Final Sample Size by Stratum and Sampling Region....................................162Appendix Table 9 Effective Sample—Panel by Stratum and Sampling Region.........................162

7

Acronyms AIDS Acquired immune deficiency syndromeCOMESA Common Market for Eastern and Southern AfricaCOMTRADE Common Format for Transient Data ExchangeDFID U.K. Department for International DevelopmentEPZ Export processing zoneFSD Financial Sector DeepeningFY Fiscal yearGDP Gross domestic productGoK Government of KenyaHCDA Horticultural Corps Development AuthorityHIV human immunodeficiency virusHS Harmonized SystemICA Investment Climate AssessmentIFC International Finance CorporationISO International Organization for StandardizationIT Information technologyKIPPRA Kenya Institute for Public Policy Research and AnalysisKPLC Kenya Power and Lighting Co. Ltd.K Sh Kenya shillingkWh Kilowatt hourLCSPS LAC Public Sector GroupLEGJR Legal and Judicial Reform Practice GroupMSME Micro, Small, and Medium Enterprise (program)NATTET National Association for Technology Transfer and Entrepreneurial TrainingPSDS Private Sector Development StrategyRPED Regional Program for Economic DevelopmentSME Small and medium enterpriseSMLE Small, medium, and large enterpriseSSA Sub-Saharan AfricaTFP total factor productivityVAT value-added tax

8

EXECUTIVE SUMMARY

The central objective of this Investment Climate Assessment (ICA) is to identify the main impediments to productivity growth faced by Kenyan firms. This objective is achieved through the analysis of firm level data directly collected by the World Bank in 2007. This report complements the doing business indicators and provides a solid analytical foundation for private-sector development policy dialogue and design. The last Kenya ICA (2004) indeed served as one of the key analytical tools to inform the Government of Kenya (GoK) of its reform efforts over the past few years. It showed that the business environment in Kenya was characterized by poor infrastructure, complex and bureaucratic administrative and regulatory regimes, poor governance, poor service delivery, insecurity, and unsuitable financial instruments.

This ICA arrives at a critical juncture when the government has committed to improving the investment climate, even further convinced that growth can only be achieved through a prosperous private sector. Based on the view that prosperity requires a thriving industrial sector, private-sector-led growth is central to the government’s Economic Recovery Strategy and its recent “Vision 2030.” In early 2007 the Government of Kenya launched its first-ever Private Sector Development Strategy. This strategy is based on five pillars: improving Kenya’s business environment, accelerating institutional transformation, facilitating growth through greater trade expansion, improving productivity of enterprises, supporting entrepreneurship, and small and medium enterprise (SME)[[as meant?]] development. All of them are linked to the analytical goal of the ICA.

The ICA uses a robust and standardized methodology that has been applied to many countries around the globe. The ICA is based on a representative sample of 657 formal firms and 124 informal establishments. The sample was drawn in four locations (Nairobi, Mombasa, Nakuru, and Kisumu) and covers both manufacturing and services. Weights were used in the analysis of the data to ensure full representativeness of the results. Although the sample is quite large, sample nonresponse could invalidate the results of the analysis—especially for more sensitive questions on corruption, taxes, or sales. In order to reduce such possibilities, strict quality control procedures have been applied during the data collection process. These controls led to an overall response rate above 90 percent. The analysis is based on both perception questions as well as objective indicators. While perception questions are used as starting point of the study, because of their inherent limitations objective questions are used throughout the report to confirm or reject what perception questions appear to indicate. The use of objective questions allows also for more meaningful cross-country comparisons. Consequently, international comparisons are done not only based on perception questions, but most importantly based on objective indicators such as indirect costs, prevalence of generator usage, etc. The use of objective indicators is the preferred approach to identify binding constraints. Finally, data from additional sources (e.g., Doing Business, Connecting People, Transparency International, etc.) are used to validate the conclusions drawn from the surveys results.

Although Kenya has recorded some improvements in the last four years, including an increase in productivity, Kenyan firms still face an adverse business environment. As a

9

matter of fact, the total losses incurred by businesses because of power outages, theft and breakage during transport, payments of bribes, and protection payments are much higher than those for the middle-income countries in Africa, China, and India.

The top constraints identified by the Kenyan managers were tax rates, access to finance, corruption, security, infrastructure services (electricity and transportation), and business licensing.

In Kenya, complaints about tax rate top all other constraints, becoming the most reported bottleneck since 2003. Kenya has reduced the corporate tax rates in recent years. Nevertheless, objective indicators of fiscal pressure suggest that the tax burden in Kenya remains higher than in most comparator countries. Kenyan firms are still required to pay half of their corporate income in taxes, an overall amount that is lower than in China and India but it is much higher than in the other African comparator countries. The high tax burden faced by Kenyan firms is mainly due to the profit tax rate (32.5 percent), which is the highest of all comparator countries, including China and India. Although a more detailed analysis of the tax burden in Kenya is recommended, one potential impact of a high tax regime is higher evasion, as well as the presence of a larger informal sector. As a matter of fact, our analysis shows that one of the top three main reasons for informality is the negative perception associated with the tax burden.

Access to credit is significantly more difficult for microenterprises and small enterprises. Consistent with improvements in the banking sector over the last few years, the proportion of firms constrained by access to finance in Kenya declined from 75 percent in 2003 to approximately 36 percent in 2007. Notwithstanding a favorable lending regime with low real costs of debt and a high proportion of firms with good quality information, 90 percent of microenterprises and 60 percent of small firms in Kenya declare they need loans, compared to 40 percent of medium and large firms. Microenterprises are also priced out of the market on account of collateral requirements—43 percent of microfirms and 12 percent of small firms, compared to only 7 percent of medium and large firms, report that collateral requirements discouraged loan applications. The complexity of the application process is another impediment for micro and small firms.

Although its ranking improved over the last four years, corruption remains one of the top bottlenecks for firms in Kenya. In general, 75 percent of firms in Kenya reported having to make informal payments to “get things done” with rules and regulations. Corruption costs Kenyan firms approximately 4 percent of annual sales, which is a considerable amount by international standards. Furthermore, Kenyan firms are required to pay approximately 12 percent of the value of a public contract in informal payments. This is again higher than all comparator countries. Bribes to tax inspectors are also common in Kenya, as is the request for informal payments for licensing and utility hookups. Finally, one particular aspect of corruption that seems to be unique to Kenya is the common practice of the police to request payments from trucks in transit.

Security remains a major constraint to firms in Kenya. Today, approximately one-third of Kenyan managers rate crime as a major constraint. Crime can add significantly to the cost of doing business in Kenya, both directly through theft and indirectly through security measures

10

employed to prevent violence. Overall, these costs amount to approximately 9 percent of sales, which is considerably higher than in all comparator countries, including South Africa, where it reaches only 1.5 percent. This data is based on a survey conducted before the unrest following the 2007 elections. Consequently, these figures must be considered conservative since recent conversations with businesspersons in the country have highlighted even more the issue of lack of security.

Electricity and transport are the main infrastructure bottlenecks affecting Kenyan firms. Close to 80 percent of firms in Kenya experience losses because of power interruptions. This is the highest value of all comparator countries. As a consequence, almost 70 percent of firms have generators, which are costly to obtain and operate. Power disruption costs Kenyan firms approximately 7 percent of sales. In a cross-country comparison these losses are among the highest. Similarly, Kenyan companies lose 2.6 percent of their sales because of spoilage and theft during transportation. This is the highest value of all comparator countries.

Although Kenya has recently reduced the number of tax payments, tax administration remains a major burden for firms in Kenya. Approximately one-third of firms rate it as a major bottleneck. Approximately 75 percent of firms in Kenya report having been visited by tax officials in 2007. All our comparator countries but China experience a much lower degree of visits by tax administration officials. Moreover, the tax filing system in Kenya is cumbersome. Kenyan firms spend about 430 hours in preparing forms and filing and paying taxes. On the other hand, value-added tax (VAT) refunds are relatively efficient in Kenya.

Notwithstanding recent reforms, business licensing remains an important constraint for Kenyan firms. Approximately 20 percent of managers interviewed place licenses among the top three constraints and more firms complain about them than in all comparator countries. The Kenyan government has taken this problem seriously and a number of reforms have been implemented whose effect will be felt in the next few years. Reforms notwithstanding, Kenya does not perform as well as comparator countries in such areas as starting a new business, renewing licenses, and the costs of licenses. Hence the reform program must persist.

While formalization would facilitate access to finance for informal firms, the financial burden of registration and taxation plus the minimum capital requirements are the main reasons why firms do not choose formality.

To address these constraints, the following recommendations are suggested.

11

Policy Matrix

Problem Action Taken Recommendations

1. Taxes

High taxes are the most reported bottleneck in Kenya. Objective indicators of fiscal pressure suggest that the tax burden in Kenya remains higher than in most comparator countries. Kenyan firms are still required to pay half of their corporate income in taxes, an overall amount that is much higher than in the other African comparator countries.

Kenya has recently reduced the tax rates faced by corporations. The most important reforms in corporate income taxes have focused mainly on lowering rates in efforts to combat global competition. Rates have been reduced from a peak of 45 percent in 1990 to around 30 percent today.

Conduct an in-depth study of the effective marginal rate of taxation to determine the extent of excessive taxation—taking into account rebates and fiscal incentives—across different sectors

2. Finance

Although we observed a decline in the proportion of firms constrained by access to finance since 2003—from 75 to 36 percent—access to credit is significantly more difficult for smaller firms. A total of 90 percent of microenterprises and 60 percent of small firms declare they need loans, compared to 40 percent of medium and large firms. Hence, firm size is an important determinant of access to credit. Among microenterprises, only 3 percent have access to an overdraft facility compared to 66 percent among medium and large enterprises. Similarly, only 27 percent

Regulations for private credit bureau operators are expected to be gazetted in June 2008, and the Central Bank of Kenya is preparing to license the first operator(s)—supported by the bank’s Financial and Legal Sector technical assistance project. In parallel, IFC is working with the Kenya Bankers Association on ensuring effective participation of banks in the new credit bureau(s).

A reform program for the companies register has been developed and is being implemented by the Registrar General supported by the Bank’s Financial and Legal Sector technical assistance project.

Enhance credit information infrastructure. With the new regulations to be issued by the Central Bank of Kenya, support will be given to enabling private bureaus to operate in Kenya. Several international and local companies are reporting keen interest in applying for the license once regulations are issued. The International Finance Corporation (IFC) is providing assistance to the Kenya Bankers Associationon the code of conduct and strategy for bank participation in the private bureaus.

Upgrade corporate registries, collateral registries, and public record systems. The scope of financial information infrastructure should include efficient access to corporate information, registries of secured lending charges and court records, etc.

Computerize property registration process; simplify taxes and fees.

12

of small firms report having any credit products, compared with 55 percent of medium and large firms. Together with collateral requirements, the application process itself is also considered a major problem by both micro and small firms.

Focus on land registration and transfer issues has intensified following the 2007 elections.

Efficient land registries and the ability to easily perfect and transfer land titles are an important vehicle for providing property owners with access to collateralized financing. Backlog and paper-based records necessitate that all history of transactions relevant to the property must be checked every time.

Promotion of new products is being undertaken by the private sector—as in the revolution in m-banking created by Safaricom’s M-Pesa and Equity Bank’s vast increase in client outreach. New product development is also being supported by the U.K. Department for International Development/World Bank Financial Deepening Trust in such areas as weather insurance, warehouse receipts, and payments system innovation.

Such capacity building is being supported by the Bank’s Micro, Small, and Medium Enterprises project, which promotes lending to SMEs and business development services.

Promote the application of innovative products and technology to expand access to finance. Capacity building for banks and microfinance institutions in the use of different lending technologies, secured lending, leasing, mortgage finance, and, in the longer run, the promotion of new products such as warehouse receipts or weather insurance are likely to have high impact on financial depths.

In order to promote improved access by small businesses to the products and services of commercial banks, facilitate the provision of capacity building for small businesses to better understand the requirements of banks (how to approach banks for business loans and how to use bank services) and prepare them for a relationship with a commercial bank. The training would be organized in collaboration with local-training business-development service providers and training institutes, and should be sponsored by local banks.

Increase transparency regarding interest rates and noninterest charges and fees (such as negotiation, commitment, legal, evaluation, processing, and insurance) on checking and current accounts.

Establish a clear time table for the

13

Financial Sector Deepening Kenya is supporting the Central Bank in developing improved disclosure of bank interest rates and fees to all consumers in order to encourage stronger competition in financial markets to ensure that the benefits from increasing productivity and efficiency in the banking sector give rise to benefits in pricing among consumers. Very preliminary indications suggest that price-based competition is leading to downward pressure on prices.

creation of credit bureau

Facilitate capacity building for banks in order to develop and market new products

3. Corruption

Although its ranking has improved over the last four years, Kenyan firms still place corruption among the most important constraints to their businesses. Nearly 70 percent of firms that reported corruption as a binding constraint ranked it as a top constraint. Corruption takes many different forms, from making payments for utility hookups to informal payments in public procurement. In general, three-fourths of firms in Kenya reported having to make informal payments to “get things done” with rules and regulations. This costs Kenyan firms approximately 4 percent of annual sales. The Enterprise Survey data

The GoK now posts on the ministries’ Web sites all

1) Conduct an in depth study of corruption in the country

2) Give prosecutorial power to the Anti Corruption Authority and publicize better the successful anti corruption cases

3) Tax administration: continue reforms aimed at

Minimizing human contact between taxpayer and officials and make the process more transparent by relying heavily on information technology to file tax returns;

Establishing independent internal and external audits; and

Introducing organizational changes of the Revenue Authority: incentives for high performers, sanctions for corrupt behavior, career development, and competitive

14

allows us to identify the many aspects of a business that creates opportunity for illegal payments. For instance, Kenyan firms are required to pay approximately 12 percent of the value of a public contract as informal payments. One-third of the surveyed firms reported being the subject of informal payment requests from tax inspectors visiting them. This is high by international standards. Licensing represents yet another opportunity for informal payments to take place. When dealing with licenses, Kenyan firms are requested informal payments approximately one-fourth of the time. Furthermore, one particular aspect of corruption that seems to be unique to Kenya is the common practice of the police requesting payments from trucks in transit. Finally, the share of managers concerned about the functioning of the courts—out of those that actually used them—rises to 33 percent, on par with crime and tax administration.

information on contracts, including names of contractors, decisions of the Procurement Appeals Board, bidders and tender outcomes, and contractors’ performance. Contracts above 5 million are posted on the Web site hosted at Treasury. Plans are at an advanced stage for local hosting at PPOA.

The GoK is proposing to blacklist companies found to have been involved in cases of corruption in accordance with the new procurement law. No requests for blacklisting have been received so far from any of the procuring entities.

GoK is taking steps to accelerate implementation of a more coordinated and prioritized e-government initiative, with public access to procurement as one of the highest near-term priorities.

salaries.

4) Public procurement: continue reforms aimed at

Reviewing procurement rules with the goal of simplifying tender documents, reducing minimum value of contract for single sources, and introducing anticorruption laws, performance standards, and sanctions;

Improving transparency in public-private interactions through e-procurement, publication of tender documents and tenders received, and public participation in negotiations;

Introducing a vetting system (conducted by an international firm, possibly with involvement of civil society) to prequalify companies interested in bidding for government contracts to address conflict of interests and fraudulent companies;

Establishing an independent tender evaluation and auditing and monitoring of unit rates; and

Supporting greater level of integrity and professionalism among multinationals and domestic companies through professional associations, codes of conduct, monitoring and benchmarking, and integrity pacts.

5) Police

15

GoK is taking steps to establish mobile and visiting courts in sparsely populated areas. Visiting courts at Mpekotoni, Archers Post, Wamba, Loitokitok, Dadaab, Kakuma, and Marimanti upgraded to fully pledged courts.

GoK taking steps to incorporate alternative dispute resolution mechanisms and provision of legal aid schemes. The Rules Committee considering experiences learned from a study tour with a view of coming up with a pilot project in Milimani Commercial Court.

GoK is taking steps to launch comprehensive wireless-based public information hubs in districts and constituencies, with public access to government a high priority.

Restructuring and privatization of Telkom Kenya is ongoing.

Have observers join the trucks to monitor the request for bribes. Use recording systems to monitor traveling time and illegal behavior.

Establish computerized checkpoints to make the process more transparent and quicker with less interaction between truck drivers and police officials. Educating truck drivers about the automated system will also reduce the harassment they face.

Install electronic weighing stations.

Involve associations engaged in trucking operations in sensitizing truck drivers to comply with the rules and regulations.

Establish an independent police complaints commission entrusted with following up on the implementation of the reform program.

Reduce the discretionary power of police

Conduct effective educational campaigns of traffic rules to reduce ability of police to extort bribes

6) Utility

Complete the liberalization of fixed line telephony.

Privatize some forms of service delivery (utility hookups).

Use citizen report cards to

16

assess the performance and quality of services and monitor progress. Publish progress reports periodically based on customer surveys and timely audits.

4. Electricity

Close to 80 percent of firms in Kenya experience losses because of power interruptions. This is the highest value of all comparator countries. As a consequence, almost 70 percent of firms have generators, which are costly to obtain and operate. Power disruption costs Kenyan firms approximately 7 percent of sales. In a cross-country comparison, these losses are among the highest.

Since June 2006, Kenya Power and Lighting Co. Ltd. (KPLC) has been managed by an international management services contractor.

KPLC made a profit in both fiscal year (FY) 2005/6 and FY 2006/7. During FY 2007/8 KPLC’s performance has continued to improve—e.g.,network losses reduced. At present, government is providing a nontargeted subsidy to electricity consumers of K Sh .60 per kWh.

The conversion of the Electricity Regulatory Board to the Energy Regulatory Commission on July 7, 2007, is an important step in the right direction. By taking this action, the government has moved the power sector one step closer to being overseen by an independent regulatory entity with the clear legal authority for performing the universal tasks of a regulatory entity: setting tariffs and quality of service standards, and licensing operators.

The establishment of the Rural Electrification Authority in 2007 has transferred the responsibility for rural

Increase public investment in energy generation, transmission, and distribution to increase connectivity;

Encourage increased private financing and investment in the energy sector—today, private sector accounts for 12 percent of power supply;

Establish clear rules for private generators’ “open access” to transmission network, the concept of which was established in the Energy Policy;

Ensure electricity pricing maintains the financial viability of power companies, while protecting the most vulnerable consumers.

Develop the legal framework for investments in energy

Consider using the least cost development plan to increase investments in energy

17

electrification from KPLC to the Rural Electrification Authority. The authority will manage the Rural Electrification Fund, with an expected annual turnover of K Sh 4 billion (US$60 million) of government funds plus any donor funds made available to it.

5. Transport

Managers identified transportation, together with electricity, as the two leading infrastructure constraints to doing business in Kenya. The strong discontent of Kenyan firms is echoed by the high direct and indirect costs they have to bear because of quality of the transportation infrastructure. Even worse, shipping a 40-foot container costs Kenyan firms much more than firms in all other comparator countries, except Uganda. Unfortunately, when we look at indirect costs Kenya does not perform any better. Kenyan companies lose 2.6 percent of their sales to spoilage and theft during transportation. This is the highest value of all comparator countries.

Road

Kenya’s transport system is important not only for Kenya itself but also for its regional partners. Kenya requires about US$1,500 million to bring the primary road network back into good condition, while the Road Maintenance Fuel Levy yields about $200 million per annum.

The Kenya Roads Board finalized the overall road sector expenditure strategy and investment plan but this now awaits governmental adoption.

The government passed the Kenya Roads Act and has established three roads authorities, namely, the Kenya National Highways Authority; Kenya Rural Roads Authority; and the Kenya Urban Roads Authority in order to efficiently manage the entire road network in Kenya.

The chair and members of the board of the three authorities have been appointed and the authorities will become operational as soon as the three chief executive officers and senior staff are appointed. The government has also adopted a detailed road sector policy and

1) Roads

The Ministry of Finance should establish a system for ensuring proper investment planning and management. This would, among other things, involve:

a. Issuing guidelines for a minimum level of preparation of projects before they are submitted for budget requests, including compatibility with overall sector strategy and development plans, economic analysis, confirmation of having prepared detailed designs based on field investigations and the required bidding documents, and readiness for implementation.

b. Strengthening institutional structure for implementing the guidelines. A special unit could be set up to screen projects submitted for budget funds. Such a unit would have a close working relationship with the medium-term expenditure framework and business and sustainable development[[defined as meant?]] units in the Ministry of Finance and would be the repository of a

18

strategy paper that will form the basis for future programs and reforms in the sector.

Ministry of Roads has drafted a policy paper on the involvement of the private sector in the management of truck weigh stations and axle load control.

Aviation

Kenya Airports Authority and Kenya Civil Aviation Authority have been given financial autonomy and now retain the revenue generated from their operations, which had been previously remitted to the Treasury.

The responsibility for passenger, baggage, and mail security screening at the airports has been transferred from the police to Kenya Airports Authority, allowing for better monitoring, control, and training of security staff.

The regulations for safety and security have been harmonized and adopted by all four member countries of the East African Community.

multiyear rolling investment program containing an inventory of appraised and priority-ranked projects for budgetary consideration in the future.

Ongoing reforms in the roads sector should be expedited. This would involve

a. Expediting the operationalization of the Kenya National Highways Authority, the Rural Roads Authority, and the Urban Roads Authority.

b. Strengthening the residual Ministry of Roads to perform its overall policy, planning, and coordination role.

c. Promoting the use of long-term output and performance-based contracting and concessioning for maintenance and management of the major road network by the private sector, starting with the Northern Corridor.

The government should improve governance in the road sector:

a. Strengthen the Engineers’ Registration Board and empower it further to discipline and sanction engineers and firms who perform poorly and violate its charter with regard to professional conduct and ethics. The same would be true for the Association of Consulting Engineers.

19

b. Assist the construction industry in establishing a professional body for construction contractors (national construction council, or a contractors’ registration board) and strengthen it to engage in self-regulation.

c. Develop a comprehensive construction industry development policy and establish a dedicated construction industry development board to implement the policy to enhance the performance of the construction industry.

d. Ensure regular updating of contractors’ qualifications and capacity; facilitate training in different aspects of construction and supervision techniques; and reprimand poor performance.

e. Approve policy on private sector participation in the management of weigh stations and control of axle load regulations.

Improve public transportation system

Facilitate more private involvement in transport

2) Port and Maritime

Expedite conversion of Kenya Ports Authority to a landlord authority;

Concession the Mombasa container terminal(s), the dockyard

20

and marine services, and the bulk oil terminals;

Streamline cargo clearance procedures and remove the police escort system for transit cargo by road (except for hazardous and military supplies);

Introduce risk-based targeting for cargo inspection and verification;

Implement a harmonized customs clearance system and one-stop border posts in accordance with Common Market for Eastern and Southern Africa protocols; and

Review and ensure compatibility of local maritime regulations with the International Maritime Organization treaties.

3) Aviation

Expedite safety and security enhancement at Jomo Kenyatta International Airport and strengthen the Kenya Civil Aviation Authority to obtain International Air Safety Association and United States Transportation Security Administration Category 1 clearance to operate direct flights to and from the United States.

4) Railways

Expedite putting in place the independent multi-sector regulatory body, in particular, for the railway sector;

Convert the residual Kenya Railways Corporation into an

21

asset holding company which would also monitor and evaluate the performance of the concession.

6. Licensing and Regulatory Governance

Approximately 20 percent of managers interviewed place licenses among the top three constraints, and more firms complain about them than in all comparator countries. Reforms notwithstanding, Kenya does not perform as well as comparator countries in such areas as starting a new business, renewing licenses, and the costs of licenses

Since 2005, Foreign Investment Advisory Service/International Finance Corporation-World Bank, with support from development partners, have provided technical assistance to GoK (the Business Regulatory Reform Unit at Treasury and other parts of government) on licensing and regulatory reforms.

The Kenyan government has recognized the importance of the licensing burden and a number of reforms directed at reducing the number of licenses were approved in 2006 and 2007. The reform program has identified, for the first time, 1,325 active business licenses and eliminated 315, simplified 379, and cut both the time and cost of getting building permits. Notably, 23 out of a priority list of 26 problematic licenses identified by businesses have been eliminated or simplified. The still ongoing program will eventually eliminate or simplify at least 900 more of the country’s 1,300 licenses.

In the next stages, the regulatory reform and capacity-building project will assist GoK in preparing and implementing a regulatory reform strategy. The strategy and its supporting implementation projects will continue to support the

Follow up with the implementation of the licensing reforms;

Reduce the overall burden of licenses imposed on businesses, including a reduction in time and costs of obtaining a license to undertake business operations;

Continue establishment of an electronic register of licenses;

Adopt a regulatory reform strategy to serve as a framework for licensing and other regulatory reforms, and to ensure their sustainability;

Reduce the burden imposed on businesses by on-site inspections;

Tackle licensing and regulatory reforms at the local government level;

Introduce a system for vetting proposed regulations to ensure that they do not place an undue burden on businesses;

Reduce the cost of trade documents;

Reduce minimum capital requirement to register a company;

Reduce the costs to start a business;

Reduce the time taken to start a business;

Reduce time for VAT refund;

22

licensing reforms (including setting up an electronic registry of all valid licenses), streamline inspections procedures, introduce a system for vetting new licenses, address regulatory reforms at the local government level, and build capacity of stakeholders to ensure the sustainability of the licensing reforms.

Reduce number of payments for social security contribution and for VAT payments; and

Establish online filing as already done in South Africa and Mauritius.

Harmonize the different tax identification numbers (PIN, VAT, etc.) into one universal number

Identify clear responsibilities to continue licensing reforms

Improve information and transparency on regulatory reforms and outcomes

Reduce time for VAT refunds by allowing firms to use it as credit toward next payment

Reduce number of licenses by local authority and clarify the legal status of the ‘circular’

23

Acknowledgments

This report was prepared by a team led by Giuseppe Iarossi (AFTFP) and comprising Magueye Dia (AFTFP), Leonardo Garrido (AFTP2), Ricardo Gonçalves (AFTFP), James Habyarimana (AFTFP), Alemayehu Kuma (AFTFP), Manju Kedia Shah (AFTFP), Sofia Silva (AFTFP), and Mans Soderbom (AFTFP). Hayk Sargsyan (AFTFP) and David Shiferaw (AFTFP) provided invaluable research assistance. Anil S. Bhandari (AFTTR), Lisa Bhansali (AFTPR), Christina Biebesheimer (LEGJR), Colum Garrity (PRMPS), Paivi Koljonen (AFTEG), Sahr Kpundeh (AFTPR), Carolina Rendon (LCSPS), and Barry Walsh (AFTPR) contributed to the policy recommendations. Comments and suggestions were received by Matilde Bordon (CICRS), Gerardo Corrochano (AFTFP), Jacqueline Coolidge (CICAF), Vyjayanti T. Desai (IISEC), Michael J. Fuchs (AFTFP), Lars Nikolajs Grava (CICRS), Praveen Kumar (AFTP2), Peter Ladegaard (CICRS), Melanie Mbuyi (AFTFP), Vincent Palmade (AFTFP), Sanjay Pradhan (PRMPS), Giovanni Tanzillo (AFTFP), Dileep Wagle (AFTFP), Susan E. Wilder (WBIFP), and the participants to the concept note review and ICA review sessions. Official peer reviewers were Qimiao Fan (WBIFP), Najy Benhassine (MNSED), Prof. Peter Kimuyu (University of Nairobi), Moses Ikiara (KIPPRA), Betty Maina (KAM), Frank Matsaert (DFID-Kenya), and Joseph Ngumi (National Association for Technology Transfer and Entrepreneurial Training—NATTET). We are grateful to DFID-Kenya for sponsoring both the cost of data collection and report writing.

24

OverviewDuring the 1980s and 1990s, Kenya’s economic performance was far below the country’s potential for several reasons, such as inadequate economic management, widespread corruption, periodic drought, poor infrastructure, and a lack of investments. Since December 2002 a number of important efforts have been made to promote economic growth under the Economic Recovery Strategy for Wealth and Employment Creation, 2003–07. Nevertheless, the private sector in Kenya was still facing a number of serious obstacles that slowed down the country’s growth potential. An assessment of the investment climate in 2004 showed that the business environment in Kenya was characterized by poor infrastructure, complex and bureaucratic administrative and regulatory regimes, poor governance, poor service delivery, insecurity, and unsuitable financial instruments.

Over the last few years, the Kenyan government has taken strides in creating a favorable business environment through a number of reforms in (1) access to finance (Financial Sector Deepening Trust for capacity building of financial institutions and a greenfield small- and medium-enterprise risk capital fund), (2) capacity building and sectoral linkages (matching grant for value chain in a number of sectors to promote competitiveness and linkages across sectors, develop 100 business cases to incorporate in the curriculum of three Kenyan business schools, restructuring the levy scheme), and (3) improving the business environment (streamlining the business registration process and introducing a simplified taxation regime).

Despite these efforts and the fact that the country’s real gross domestic product (GDP) growth has shown an impressive growth of around 6 percent over the last three years, Kenya still remains a low investment country by international standards. Based on the view that prosperity requires a thriving industrial sector, private-sector-led growth is central to the government’s Economic Recovery Strategy and its recent “Vision 2030,” which positions Kenya to reach middle-income status by 2030. In early 2007,1 the Government of Kenya launched its first-ever Private Sector Development Strategy (PSDS) and will soon finalize, jointly with the development partners, the PSDS Implementation Plan. This strategy is based on five pillars: improving Kenya’s business environment, accelerating institutional transformation, facilitating growth through greater trade expansion, improving productivity of enterprises, supporting entrepreneurship, and SME development.

A better knowledge of the impediments to investment, productivity, and growth at the firm level is essential to pinpoint areas of reform. The Investment Climate Assessment (ICA) is part of this effort. Based on firm level data on approximately 650 establishments, this report complements the doing-business indicators and provides a solid analytical foundation for private-sector development policy dialogue and design. The last Kenya ICA (2003) indeed served as one of the key analytical tools to inform the Government of Kenya of its reform efforts over the past few years. This ICA arrives at a critical juncture, when the government has committed to improving the investment climate even further convinced that growth can only be achieved through a prosperous private sector.

1 “Private Sector Development Strategy, 2006–2010,” Ministry of Trade and Industry, Government of Kenya.

25

Productivity in Kenya has increased by approximately 4 percent per annum over the past four years. At the same time, labor costs per worker have increased, but more slowly than the increase in productivity, leading to a decline in unit labor costs. Unit labor costs in Kenya have fallen from 31 percent of value added in 2002 to 25 percent of value added in 2006. While this indicates a move toward efficiency, strategic competitors such as China and India also have improved. In both of these countries, labor costs per worker are lower than 20 percent of value added per worker. For Kenya to catch up with its Asian competitors, further growth in productivity must be achieved.

The role of an adverse business environment is particularly important in this regard. Enterprises in Kenya continue to face an adverse business climate: total losses incurred by businesses because of power outages, theft and breakage during transport, payments of bribes, and protection payments are much higher than those for the middle-income countries in Africa. In Kenya, a substantial part of sales are lost because of these indirect costs, compared to a much smaller percent in China and India.

The top three constraints identified by the Kenyan managers were tax rates, access to finance, and corruption. They were followed by complaints about infrastructure services (electricity and transportation), crime, and practices of competitors in the informal sector. Over the last four years, the perceived constraints identified by Kenyan firms have somewhat changed. In 2003 manufacturing firms were concerned primarily about finance (mainly cost), corruption, crime, and taxes. In 2007 we recorded an improvement in corruption, whereas tax rates reached the top position of perceived constraints. Similarly, in 2007 we recorded major improvement in macro instability and telecom, once among the major constraints to Kenyan firms but today no longer an issue.

These perceived constraints have a significant impact on indirect costs. The survey data shows that firms in Kenya have to bear indirect costs that amount to approximately 20 percent of their total sales. Of these, electricity (production lost because of power outages) is the main component (7.1 percent of total sales). Crime and bribes are also relevant.

All around the world, businesses tend to complain about tax levels. Nevertheless in Kenya, complaints about tax rates top all other constraints. This perception has improved over the last four years, falling from 68 percent to 57 percent of firms perceiving it as a major problem. Kenya has reduced the corporate tax rates in recent years by making it more comparable to its neighbors in East Africa. Nevertheless, objective indicators of fiscal pressure suggest that the tax burden in Kenya remains higher than in most comparator countries. As a matter of fact, Kenyan firms are required to pay half (50.9 percent) of their corporate income in taxes. This amount is lower than China’s and India’s but much higher than the other African comparator countries’ tax burden. More specifically, the high tax burden faced by Kenyan firms is due mainly to the profit tax rate (32.5 percent), which is the highest of all comparator countries, including China and India. The profit tax in China and India is less than 20 percent and in South Africa is less than 25 percent. Kenya has profit tax rates over 7 percentage points higher than in the major comparator countries. On the other hand, labor taxes, and contributions in Kenya are lower than in most comparator countries. One potential impact of a high tax regime is higher evasion, as well as the presence of a larger informal sector. In fact, our analysis shows that one of the top three main reasons for informality is the negative perception associated with the tax burden. Given that the

26

informal sector constitutes 72 percent of the working population, further reforms in this area will discourage tax evasion as well as favor the formalization of more firms.

Consistent with changes in the performance of the banking sector over the last five years, we observed a decline in the proportion of firms constrained by access to finance. In 2003, approximately 75 percent of firms reported being constrained by it. In 2007, only 36 percent of the same set of firms reported access to finance as a major or severe impediment. This decline is large and suggests a significant improvement in the external financing regime. Nevertheless, finance remains the second-most-important obstacle identified by Kenyan firms.

While aggregate measures of access to credit in Kenya are relatively good, firm size is an important determinant of access to credit. Access to credit is significantly more difficult for microenterprises than for small enterprises and considerably more difficult for small enterprises than for medium and large enterprises. Firms with fewer than 20 employees are twice as likely as firms with more than 20 employees to report finance as a major constraint. Microenterprises (with fewer than five employees) are less likely to have a bank account. Only 41 percent of microenterprises have a bank account compared to 90 percent of small and 99 percent of medium and large firms. While loan application rates are similar between micro and small firms, the proportion of small firms with a loan is double the proportion of microfirms that have a loan. The differences in access to an overdraft facility are even more glaring. Among microenterprises only 3 percent have access to an overdraft facility compared to 66 percent among medium and large enterprises. Similarly, only 27 percent of small firms report having any credit products, compared with 55 percent of medium and large firms. Small firms are about 17 percentage points more likely to report that access to finance is a serious obstacle compared to large firms, ceteris paribus. Finally, small firms are about 46 percentage points less likely to have overdraft facilities and about 42 percentage points less likely to have a loans and overdrafts than large firms, holding all other factors constant.

Measures of loan application rates provide important information about impediments to access to finance. Microenterprises are less likely to report “no need for loan” as a reason for nonapplication. Around 90 percent of microenterprises say they need loans, compared to 62 percent of small and 40 percent of medium and large firms. This corroborates the evidence that access to credit, particularly for micro and small firms, is much worse relative to medium and large firms. Microenterprises are also more likely to be priced out of the market on account of collateral requirements. Over 40 percent of micro firms compared to 12 and 7 percent of small and medium and large firms, respectively, report that collateral requirements discouraged loan applications. Given the preponderance of fixed assets as collateral and the size and scope of microfirms, it is not surprising that collateral requirements are an impediment to accessing finance for microfirms. In fact, ownership of land is associated with a 19-percentage-point increase in securing an overdraft or loan. In addition, firms that own land are nearly 13 percentage points less likely to have a loan application rejected. Together with collateral requirements, the application process itself is considered a major problem by both micro and small firms, while small firms also complain about interest rates.

Although its ranking has improved over the last four years, Kenyan firms still place corruption among the most important constraints to their businesses. Nearly 70 percent of firms that

27

reported corruption as a binding constraint ranked it as a top constraint. The results of this report are confirmed by other data sources. Both the Transparency International Corruption Perception Index and the World Bank’s Governance Indicators show an improvement in Kenya’s ranking over the last few years. Nevertheless, Kenya’s corruption rating remains the worst among all comparator countries.

Corruption takes many different forms, from making payments for utility hookups to informal payments in public procurement. In general, three-fourths of firms in Kenya reported having to make informal payments to “get things done” with rules and regulations. This costs Kenyan firms approximately 4 percent of annual sales. The Enterprise Survey data allows us to identify the many aspects of a business that creates opportunity for illegal payments. For instance, Kenyan firms are required to pay approximately 12 percent of the value of a public contract as informal payments. This is higher than all comparator countries. Bribes to tax inspectors are also common in Kenya. One-third of the surveyed firms reported being the subject of informal payment requests from tax inspectors visiting them. This is high by international standards. Licensing represents yet another opportunity for informal payments to take place. When dealing with licenses, Kenyan firms are requested informal payments approximately one-fourth of the time. Furthermore, one particular aspect of corruption that seems to be unique to Kenya is the common practice of the police to request payments from trucks in transit. Although not widespread, this phenomenon is significant, with 21 percent of firms reporting having to pay such payments. The average amount paid is approximately 2.5 percent of sales and is born more by the service sector than by the manufacturing industry. Finally, another often-forgotten aspect of corruption relates to the functioning of the courts. The share of managers concerned about the functioning of the courts—out of those that actually used them—rises to 33 percent, at par with crime and tax administration.

Kenyan firms do not indicate power as a major constraint, although 85 percent of them report experiencing power outages. This is explained by the fact that two out of three firms own a generator. Additional survey evidence, however, shows how serious the problem of power is. Close to 80 percent of firms in Kenya experience losses because of power interruptions. This is the highest value of all comparator countries. The impact of unreliable power supply on production costs is in the order of 7 percent of sales lost. From an international perspective, the losses suffered by Kenyan firms are among the highest, as well as being the highest component of all indirect costs considered.

Finally, the ability of a country to connect firms, suppliers, and consumers to global supply chains efficiently is essential to their competitiveness. Using seven measure of performance, a recent assessment of the logistics gap across countries has ranked Kenya 76th out of 150 economies, well behind South Africa (ranked 24th), China (30th), and India (39th). In the Enterprise Survey respondents identified transportation, together with electricity, as the two leading infrastructure constraints to doing business in Kenya. While 31 percent of firms rated it as a major bottleneck, one-quarter of respondents ranked it as one of the top three constraints. The strong discontent of Kenyan firms is echoed by the high direct and indirect costs they have to bear because of the quality of the transportation infrastructure. Inland direct transport costs in Kenya are much higher than in China and India—where they are just a fraction of what they are in Kenya—and are among the highest of all comparator countries. Even worse, shipping a 40-

28

foot container costs Kenyan firms much more than firms in all other comparator countries, except Uganda. Unfortunately, when we look at indirect costs Kenya does not perform any better. Kenyan companies lose 2.6 percent of their sales to spoilage and theft during transportation. This is the highest value of all comparator countries.

In conclusion, this report identified the main constraints to private sector development in Kenya. Tax rates, finance, and corruption remain the three most important impediments, followed by infrastructure bottlenecks (mainly in electricity and transport). In order to address these bottlenecks we propose detailed recommendations on tax administration, public procurement, police, courts, utility services, customs, finance, energy, road transport, port, aviation, and licensing.

29

1 Macro Environment

After more than a decade of stagnation, Kenya’s economy shows signs of strengthening, with continued growth in per capita GDP, at rates of 1.6 in 2004, 2.6 percent in 2005, an estimated 2.9 percent in 2006, and an expected 3.1 percent in 2007,2 all within an environment of perceived moderately enhanced macro stability.

Recent growth acceleration in Kenya has been largely driven, from the institutional standpoint, by the private sector; from the demand side, by private consumption and exports; and, from the supply side, by a broad array of activities producing tradable and nontradable goods and services, including horticulture, telecommunication, wholesale and retail trade, manufacturing, and transportation.

All private main economic activities exhibited positive growth during the period 2001–05, led, in terms of overall contribution, by agriculture, transport and communication, manufacturing, and wholesale and retail trade, which together explained more than three quarters (3.1 percent points) of private growth during the period (3.9 percent). Private sector generated almost 86 percent of the value added between 2001 and 2005, more than 1 percentage point above its average during the 1990s (Figure 1-1).

Figure 1-1 Trends in Public and Private GDP Growth/Private Share in Total GDP,1978–2005

0%

1%

2%

3%

4%

5%

6%

7%

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

% A

nnua

l Gro

wth

GDP

82%83%83%84%84%85%85%86%86%87%87%

Shar

e on

Rea

l GDP

Share in private GDP / Total GDP Private GDP Growth Rate trendPublic GDP Growth Rate trend

A sector analysis shows a picture of broad-based growth between 2003 and 2006, but mainly led by the services sector. Restaurant and hotel services, telecommunications, transport, and storage together grew at around 9 percent per year. Growth in services resulted from a combination of a

2 IMF 2007.

30

spurt in tourism and telecommunications. Visitor arrivals increased by about 60 percent between 2002 and 2006, mobile telephony and other telecommunication products expanded greatly, and transit transportation activities to neighboring countries increased moderately. Given the large size of the services sector (about 53 percent of the total economy in 2006), its growth contributed more than half of the total growth in the economy from 2003 to 2006. Agricultural output (about 24 percent of the total economy in 2006) also grew in an impressive manner, led by maize, coffee, and livestock products. Export of flowers, fruits, and vegetables, mainly to Europe, became the top Kenyan agricultural activity, displacing exports of traditional products such as coffee and tea. Horticulture grew at annual rate of 5 percent between 2000 and 2005 and now represents nearly 30 percent of total agriculture GDP. The growth in the manufacturing sector was less inspiring and resulted largely from increased supply of agricultural inputs for agro-based activities, an increase in clothing production, and cement output. Finally, a rebound in construction during the last two years driven by cheaper private loans for construction and, to some extent, government spending on construction activities (stalled projects) has helped the secondary sector to grow. Reconstruction supplies to Sudan appeared to have boosted aggregate demand.

Figure 1-2 Annual GDP Growth, Main Activities, 1996–2006 and 2001–06

-5% 0% 5% 10% 15% 20% 25%

Agricultural & Animal Husbandry Serv.Electricity Supply

Public Admin. & DefenseForestry and Logging

EducationPriv. Households with Employed Persons

Mining and QuarryingRenting & Business Serv.

Financial IntermediationHealth and Social Work

Other Community, Social & Personal Serv.Dwellings, Owner Occupied and Rented

All Other ManufacturingFarming of Animals

Growing of Crops and HorticultureWater Supply

Transport and StorageConstruction

Manuf. of Food, Beverages & TobaccoWholesale, Retail Trade & Repairs

FishingPost and Telecommunications

Hotels and Restaurants

Avg. Growth 2003-2006Avg. Growth 1996-2006

31

Figure 1-3 Contribution to GDP Growth by Institutional Sector and Main Activities, Average 2001–05

Primary

Secondary

Tertiary

Tertiary

-0.5%

0.0%

0.5%1.0%

1.5%

2.0%

2.5%3.0%

3.5%

4.0%

Private Public

Perc

ent p

oint

s

Exports of goods and services was, by far, the fastest-growing component of demand between 2001 and 2005 (8.1 percent per year) but was second to private consumption in the overall contribution to GDP growth (1.7 percentage points in the former versus 2.8 percentage points in the latter).3

Kenyan exports were stimulated by favorable price and income effects and by some competitive gains in non traditional sectors over the period 2002–05. The nonoil export price index increased 1.8 percent in 2004 and 7.3 percent in 2005, led by gains in prices of food and live animals, machinery and transport equipment, animal and vegetable oils and fats, and beverages and tobacco. Furthermore, foreign income of countries importing Kenya’s products increased at an average 13 percent per year during the period 2000–05 (it increased 22 percent in 2004 and decreased 7 percent in 2005).4 Moreover, the behavior of the Real Effective Exchange Rate seems to not to have affected the exportable sector since, according to the International Monetary Fund, the Kenya shilling’s appreciation has been broadly in line with economic fundamentals. In turn, a trade analysis shows that Kenya gained competitiveness over 2002–05. This assessment is said to be consistent with the private sector’s view that although the shilling appreciation may have had some negative impact on such exports as cut flowers, other elements such as weak infrastructure remains the main constraint in Kenya’s competitiveness.5

3 Aggregate demand obtained as the sum of total GDP plus exports. It is equal to aggregate supply, which includes public and private consumption, exports of goods and services, and gross capital formation (which, in turn, equals gross fixed capital formation plus changes in inventories). Sometimes, the national accounting process yields discrepancies between total output generated as the sum of value added by productive activities and that estimated from the demand side. In this case the discrepancy (output minus demand) should be deducted from the demand component(s) suspected to be the source of difference. In Kenya’s case, this type of discrepancy has been obtained since at least 1996.4 We calculate “YProd,” a measure of the foreign income, by product, of countries importing from Kenya. Using Common Format for Transient Data Exchange (COMTRADE) HS 4-digit data, we built YProd for the period 1990–2005 as the weighted average of total GDP level (real US$ of 2000) by product “i” imported from Kenya, using as country weights their share on imports from Kenya of product “i.” Thus, YProd represents the average income level of the countries importing a given product from Kenya, not to be confused with the “Prody” developed by Hausman et al. (2005), which is the average per capita GDP of countries exporting a given product and thus which represents the income level associated with that product.5 IMF (2007).

32

Kenya has made no discoveries of new products in the last two decades,6 and has not significantly further diversified its basket of existing exportable products during the last 10 years (following the significant diversification registered between the mid-1980s and mid-1990s).7 In fact, six main groups of products accounted for more than two-thirds of Kenya’s average export values from 2000–05: Tea and Mate (19.7 percent of total exports); Horticulture (19.6 percent)8; Petroleum Products (11.2 percent); Clothing, except Fur Clothing (7.3 percent); Coffee (5.9 percent); and Fish, Fresh and Simply Preserved (3.6 percent).

On the other hand, Kenya has “discovered” a handful of new export destinations, both within Sub-Saharan Africa (SSA) and outside the region9; however, the index of export diversification by country has slightly increased over the last decade indicating a higher concentration of exports, by destination, in the last years (Figure 1-4). By destination, two-thirds of Kenya’s exports between 2000 and 2005 went to Uganda (15.2 percent of total exports), United Kingdom (14.8 percent), United States (9.8 percent), Netherlands (9.2 percent), Pakistan (6.5 percent), Tanzania (4.9 percent), and Germany (4.3 percent). SSA increased its share in Kenya’s exports by about 20 percentage points at the expense of the European Union. In the meantime, exports to Southeast Asia remained roughly constant during the last two decades at about 14.1 percent of total Kenya exports (Figure 1-5).

Figure 1-4 Herfindahl-Hirschmann Index of Product and Country–Destination Diversification, 1980–2005

6 We use Klinger and Lederman (2004) methodology to identify discoveries in Kenya, based on COMTRADE HS 4-digit disaggregated data on exports and imports by product and country for the period 1969–2005. A new product (discovery) is recorded when the 10-year average exports (up to year “t”) jumps from less than 10,000 US$ to more than 1 million US$ per year in t+10 (10 years later). Lowering the threshold for the jump of exports to 500,000 US$ yields two new products, using HS 4-digit data: “Unmilled Barley” and “Chocolate and other food preparations containing chocolate.”7 The degree of diversification of the export basket is measured by the Herfindahl-Hirschmann Index, scaled to 0–1. A lower index indicates higher diversification of the export basket, by products.8 Includes exports of products in groups 1401 to 1404 (Crude Vegetables Panting Materials and Planting Products), groups 701 to 709 (Fresh vegetables), and groups 801 to 810 (Fresh fruits) in COMTRADE HS 4-digit classification.9 A similar methodology and criteria to that used to identify product discoveries was used for country discoveries. Discovered “new” export destinations for Kenya include Belgium, Poland, Kazakhstan, Russian Federation, and Luxembourg, plus a few Sub-Saharan African economies.

33

Figure 1-5 Kenya: Export Composition by Destination Regions, Decade Averages

After more than a decade of deceleration in the growth rate of private consumption, this demand component raised 2.4 percent in the years 2003 and 2004, each; and 7.1 percent in 2005. Given its large contribution to output (79 percent in the 2000s) it maintains its role of principal contributor to GDP growth. The pickup in private consumption has been favored by the increased credit supply (as fraction of GDP) and a decline in real interest rates; and catalyzed by a real appreciation in the shilling that reduces the price of importable products (Figure 1-6).

Figure 1-6 Real Lending Rates and Domestic Credit (Private and Rest) Ratio to Nominal GDP, 1980–2006

A growth diagnostics approach shows that investment rate is low in Kenya. This is so for two broad reasons; First, returns to investment are low and risks to appropriation of returns high; second, access to finance is limited and costs are high for certain categories of borrowers, such as rural and small entrepreneurs. Further, returns to investment are low mainly because business costs—other than the cost of labor and capital—are high. These nonfactor costs take various forms. They include high transportation costs and high costs of energy. They also include opportunity costs resulting from delays of shipments, as well as direct payments in the form of bribes. The net impact of these nonfactor costs on a business is either reduced sales revenue—

34

and hence reduced profitability and productivity (as measured)—or high total costs of production. Macroeconomic and political risks have receded considerably since the 2002 elections, but they remain important. In addition, crime and the security situation remain deterrents to investment.

Kenya is a low-investment economy by international standards, but an average one among low-income countries within SSA. Kenya’s average investment to GDP was 20 percent (4 percent points) below that registered in high-income countries between 1965 and 2005 and more than 5 percent points below that of middle-income countries. In the 1990s Kenya—along with SSA—underperformed even its peers from the low-income countries group (Table 1-1).

While a part of the decline in the 1980s and 1990s was the result of declining public investment, which represented more than 46 percent of the total investment in 1978 (and 11.5 percent of GDP) and about 38 percent (6.4 percent of GDP) in 2002, private investment stagnated as well. This last aggregate showed a zero growth rate in the 1980s, which fell and recovered in the first half of the 1990s and stagnated again in the second half of the 1990s until 2002 Microevidence from manufacturing firms confirms the stagnation. In 2002/03 only 15 percent of the existing manufacturing firms had investment rates (investment expenditure as a share of replacement value of capital stock) above 10 percent, and one-third of all firms reported zero investment. RPED data from the mid-1990s also shows that approximately half of the firms undertook no investment and the majority of the firms reported modest investment rates (Soderbom 2001).

Table 1-1 Ratio of Gross Fixed Capital Formation to GDP,Current US$ (Percent Points)

Most of the recent pickup in gross capital formation has been based on the substantial growth of investment in transport equipment and other machinery and equipment, much of it linked to the importation of machines for the manufacturing sector (Table 1-2).

35

Table 1-2 Gross Capital Formation Real Growth and Structure Gross Capital Formation Real Growth and Structure

% Annual Chg. 2000-03

% Annual Chg. 2003-06

Composition 2006

Real Gross Fixed Capital Formation -1.5% 17.0% 100.0% +Buildings and Structures 6.2% 5.8% 38.1% +Transport Equipment 7.7% 34.1% 29.3% +Other Machinery and Equipment -15.4% 20.7% 32.2% +Cultivated Assets 14.1% 0.1% 0.4% +Intangible Assets -51.8% 21.0% 0.0%Source: Staff calculations based on Economic Survey

Figure 1-7 Real Gross Fixed Capital Formation, by Type of Assets, 1995–2005

0

50,000

100,000

150,000

200,000

250,000

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

MM

KSh

. of 2

001

Buildings and Structures Transport EquipmentOther Machinery and Equipment Cultivated Assets

Kenya’s consumer price index inflation accelerated from a low 2 percent per year in 2002 up to 15 percent in July 2005, slowed down back to an annual 10 percent at the end of that year and has picked up to 14.5 percent by the end of 2006, driven mainly by food price increases (12.3 percent in 2005 and 20.4 percent in 2006), and, to a much lesser degree, by fuel prices.10

Yet, Kenya consumer prices benefited from the stability of the exchange rate and appreciation of the Kenya shilling. Nonoil import prices fell 16.5 percent in 2003, increased 1.7 percent in 2004, and rose up to 6.8 percent in 2005. In 2005, the index price of imported food and live animals fell 2.9 percent. Imported mineral fuels exhibited price increases of 23.8 percent in 2004 and 33 percent in 2005, contributing to a larger increase of overall import prices (8.1 percent in 2004 and 14.3 percent in 2005).

Real wages continued their increasing trend begun in 1995, but at a somewhat slower pace since 2003. Overall, real wages increased 3.0 percent in 2003, 4.5 percent in 2004, and 4.2 percent in 10 Food, Drink, and Tobacco items represent over 53 percent of the value of the consumer price index basket in Kenya. In Nairobi, the items represent 57.5 percent of the cost of the basket for low-income households and 33.3 percent of that of middle- and high-income households.

36

2005. The maintained increases in earnings from wage employees were entirely sustained by the private sector, which experienced real pay rises over 5 percent from 2003–05, as opposed to stagnation of real public wages in 2003 and 2004 and a 1.2 percent decrease in 2005 (See Figure1-8 and Table 1-3).

Figure 1-8 Trends in Rate of Change in Private and Public Wage(Wage Employment Only), 1981–2005

Table 1-3 Kenya—Real Average Annual Wage per Employee (KSh of 2000)Kenya: Real Average Annual Wage per Employee (KSh. of 2000)

2000 2001 2002 2003 2004 2005Real Average Wage per Employee 164,621 179,224 202,022 207,997 217,306 226,486

Private Sector 161,627 176,568 198,428 208,781 220,748 236,385 -Agriculture and Forestry 67,073 70,556 77,326 80,001 80,512 81,925 -Mining and Quarry 89,348 97,307 110,075 112,318 114,675 120,173 -Manufacturing 99,518 103,160 109,582 113,843 118,730 128,327 -Electricity and Water 280,867 305,942 344,663 364,458 384,551 399,685 -Building and Construction 156,885 166,132 186,036 195,200 200,614 209,608 -Wholesale and Retail Trade, Restaurants and Hotels 251,350 275,705 315,251 333,113 347,486 366,952 -Transport and Communications 266,562 304,601 355,518 377,819 409,479 441,384 -Finance, Insurance, Real State and Business Services 320,701 354,339 404,078 429,007 453,935 486,303 -Community, Social and Personal Services 187,985 207,956 236,703 250,514 263,091 278,162

Public Sector 168,956 183,333 207,700 206,726 211,525 209,033 -Agriculture and Forestry 105,611 113,210 130,158 130,580 134,874 131,224 -Mining and Quarry 157,429 171,964 186,745 152,637 150,362 141,891 -Manufacturing 158,405 136,066 149,392 140,432 142,550 137,662 -Electricity and Water 223,156 231,836 266,024 266,382 272,359 265,990 -Building and Construction 131,913 163,196 188,372 188,611 194,186 189,373 -Wholesale and Retail Trade, Restaurants and Hotels 198,344 219,424 287,948 295,247 310,021 313,347 -Transport and Communications 302,101 241,952 284,375 291,063 311,563 312,767 -Finance, Insurance, Real State and Business Services 443,105 443,362 516,145 511,179 552,967 560,017 -Community, Social and Personal Services 156,909 180,786 203,043 201,419 203,336 199,933Source: Authors Calculations based on Economic Survey.

37

Total employment has increased at annual rates of 6.6 percent between 2000 and 2005, driven by the secondary and tertiary private sector, which saw yearly increases of 6.7 percent and 8.0 percent during the same period, respectively. In turn, public employment decreased 0.8 percent per year during the same years. In fact, secondary and tertiary private employment grew faster relative to private GDP between 2000 and 2005; however, a breakdown of employment by main sectors shows that most of the new workers went to the informal sector (8.4 percent increase per year versus a mere 1.4 percent annual increase in modern employment).

Kenya does not keep periodical records on unemployment. These figures can only be extracted from population census or labor force surveys, which are conducted once per decade, or estimated from other household survey data. The Labor Force Surveys—the last one corresponding to 1998/99—show a decade-by-decade increase in overall unemployment up to 14.6 percent of the labor force in 1998/99, with an additional underemployment rate of 4.1 percent. Urban unemployment was recorded at 25.1 percent during that year.

38

2 Competitiveness of Kenyan Firms2.1 Overview

The previous ICA survey conducted in 2003 reported that Kenyan firms had only a weak competitive advantage compared to strategic competitors such as Tanzania and Uganda, and were at a severe disadvantage compared to firms in China and India. Kenyan plants and equipment were outdated, overvalued, and inefficiently used; investment levels were low and declining. Productivity growth had been zero or negative since the 1990s. Enterprises were adversely impacted by the negative business environment, especially the burden of bribes, costly infrastructure, and higher regulatory burden.

How has Kenyan manufacturing changed over the past four years? How does Kenya now compare regionally and internationally? This chapter addresses these issues by examining data from 396 formal manufacturing enterprises surveyed in mid-2007. The chapter compares performance from this survey to the respondents interviewed in the previous ICA, and also benchmarks Kenyan firms to those in other countries.

After briefly presenting the structure of the sample, we first look at the efficiency of key inputs used in the production process: labor and capital. We then examine total factor productivity, which measures overall productivity of an enterprise after controlling for differences resulting from factor usage. Changes over time are examined using our panel data. Pooled cross-country data is used to benchmark Kenya against its competitors.

Our results on labor productivity indicate that

1. Kenyan firms have become 15 to 20 percent more productive over the past four years. 2. Kenyan firms have a significantly more productive labor force than those of Tanzania

and Uganda. 3. Productivity is not much lower than that of firms in China, and is much higher than that

of firms in India. 4. Most productivity gains have accrued to domestic, smaller enterprises.

Our results on unit labor costs, measured as a ratio of labor costs to value added, indicate that all labor productivity gains are not offset by higher worker wages and allowances. Labor costs per worker have increased, but more slowly than the increase in productivity, leading to a decline in unit labor costs. Unit labor costs in Kenya have fallen from 31 percent of value added in 2002 to 25 percent of value added in 2006. While this indicates a move toward efficiency, strategic competitors such as China and India also have improved. In both of these countries, labor costs per worker are lower than 20 percent of value added per worker. For Kenya to catch up with its Asian competitors, further increases in labor productivity must be achieved.

Total factor productivity (TFP) in Kenyan manufacturing has increased over the past four years. Depending on the measures used, productivity has improved by approximately 4 percent per annum during this period. Given the insignificant or even negative growth rates observed in

39

earlier periods, this positive trend is encouraging. The opening of regional markets, improvements in the business environment, and governance have had a positive impact on enterprise performance. Cross-country TFP estimates indicate that firms in Kenya now are more productive than firms in Tanzania and Uganda; however, they still lag far behind enterprises in SSA middle-income countries such as Botswana and Namibia.

2.2 Sample Composition

The World Bank’s 2007 Enterprise Survey in Kenya was administered to 781 firms in 4 locations. Table 2-4 shows the sample distribution across cities and sectors. The sampling approach followed was stratified simple random sampling for the formal economy, and simple random sampling for the microfirms. Close to 60 percent of the formal sample is represented by manufacturing firms, within which manufacturing food (17 percent), garments (12 percent), and other manufacturing (31 percent) represent individual strata. Outside the manufacturing sector, the retail sector account for 19 percent of the sample and fewer than one-quarter of the firms belong to the rest of the services stratum.

Table 2-4 Sample Distribution in Kenya, by Sector and Location

Nairobi Mombasa Nakuru KisumuTota

lManufacturing 274 51 34 37 396 Food and beverages 73 13 8 16 110 Garments 62 13 2 5 82 Other manufacturing 139 25 24 16 204Retail 59 16 26 25 126Rest of the universe 69 20 22 24 135Total—Formal 402 87 82 86 657Micro (4 employees or less) 64 20 20 20 124Total 466 107 102 106 781

Source: World Bank, Kenya 2007 Enterprise Survey.

In terms of geographical distribution, the capital city has the highest number of firms (60 percent) in the sample, while the rest is distributed across the other three locations, Mombasa, Nakuru, and Kisumu. Finally, 124 microfirms (with fewer than five employees) are also included in the sample. They are equally split between Nairobi and the rest of the country.11

2.3 Labor Productivity

Labor productivity is measured by manufacturing value added per worker. Earlier studies of Kenyan manufacturing have found that labor productivity in Kenya is comparable to that of China and India, and is much higher than that in Tanzania and Uganda (Figure 2-9). High labor productivity in Kenya, however, was associated with much higher capital intensity compared to

11 Note that the sample was not stratified by size. The appendix includes a detailed description of the sampling methodology followed.

40

firms in China and India. Once we controlled for capital, total factor productivity was lower in Kenya than in East Asia.

The Kenya Institute of Manufactures conducted its own Economic Survey in 2006. It found that labor productivity, measured as value added per worker in manufacturing, had increased from K Sh 456,000 in 2001 to K Sh 612,347 in 2005––an increase of 34 percent at current prices.

The present survey looks at new data from all four comparators versus Kenya (Figure 2-9). Patterns of labor productivity (measured in constant 2005 US$) remain the same across comparators. Kenyan workers are still far more productive than firms in Tanzania and Uganda (except for large Tanzanian firms, whose productivity is far higher than others). Kenyan firms have only marginally lower productivity than firms in China and much higher productivity than workers in India.

Figure 2-9 Cross-Country Comparison of Labor Productivity

Examining across firm size, we see that for most countries, labor productivity increases with firm size, except in Kenya, where labor productivity is relatively flat across different size classes. Large firms in Kenya are significantly less productive than firms in Tanzania, and only slightly more productive than firms in Uganda.

Differences in labor productivity across sectors are driven primarily by differences in capital intensity. Metal and allied products, which include the manufacture of iron and steel products in which Kenya has a significant advantage compared to Tanzania and Uganda, have the highest capital intensity and labor productivity, and corresponding lowest unit labor costs. Labor productivity and capital intensity are both higher for exporters compared to domestic firms, and between foreign versus local enterprises. These patterns are similar to those found in other countries; however, unlike in most other countries in SSA, in Kenya, capital intensity and labor productivity do not change significantly across size classes. In particular, small firms are quite capital intensive, with $8,000 worth of capital per worker (Table 2-5).

41

Table 2-5 Productivity Indicators

Number of firms

Capital productivity

Capital/labor ratio ($)

Labor productivity

($)

Labor costs per worker

($)Unit labor

costsFull Sample 395 1.04 7,786.04 7,250.75 1,791.27 0.25Small 191 1.04 8,110.46 6,843.2 1,703.2 0.26Medium 63 1.21 6,292.6 7,509.68 1,946.51 0.29Large 141 0.96 8,065.65 7,568.95 1,880.15 0.25Domestic 332 1.02 7,418.1 7,046.46 1,775.57 0.26Foreign 63 1.21 10,851.79 8,424.74 2,100.55 0.23Nonexporter 244 1.09 7,156.29 6,523.25 1,653.29 0.26Exporter 151 0.99 9,124.27 8,424.74 2,020.6 0.24Food 110 0.98 6,550.75 7,396.24 1,700.07 0.23Chemical 26 0.87 10,353.78 7,907.7 2,433.14 0.30Metalworking 30 1.07 12,805.99 13,729.85 2,561.2 0.19Textile and Garment 110 1.20 4,760.49 5,100.23 1,577.33 0.32Wood and Furniture 31 1.42 5,289.43 6,351.11 1,465.75 0.26Other Manufacturing 88 1.04 9,033.93 7,568.95 1,880.15 0.25

2.4 Unit Labor Costs

Labor productivity per se cannot be used to benchmark the competitiveness of a country’s labor force. Even though labor has low productivity, it can be offset by lower wages paid to workers, rendering workers competitive. We examine this issue by looking at unit labor costs, which measures the ratio of labor costs per worker to value added per worker. Figure 2-10 presents the differences in unit labor costs across the three African countries, China, and India.

We see that Kenya’s unit labor costs are lower than in Tanzania, and much lower than in Uganda. There is no significant difference across size classes; however, labor cost in Kenya is 25 percent of value added, compared to only 15 percent of value added in China. Higher productivity in China is not offset by higher wages, rendering China very competitive internationally. In India, while labor productivity is low, it is offset by much lower labor costs, making firms in India, particularly the larger ones, more competitive than those in Kenya (Figure2-10).

42

Figure 2-10 Unit Labor Costs

Both labor productivity and unit labor costs, however, are only partial measures of manufacturing competitiveness: they ignore the contribution of capital to the production process. Differences in labor productivity could be driven entirely by differences in the machinery and equipment use per worker. We examine this next by looking at total factor productivity.

2.5 Total Factor Productivity

Although the measures of firm productivity such as labor productivity provide useful information on firm performance, they can be misleading when considered in isolation. To get an overall assessment of productivity, it is necessary to take both capital and labor use into account. This can be done by calculating TFP. Differences in TFP are differences in output that cannot be explained by differences in the use of labor, capital, and other intermediate inputs. Differences in TFP can be due to the quality of workers, quality of management, technology used (so long as it is not embodied in capital), or firm organization. Firms for which TFP is higher are more efficient than other firms because they produce more with less capital and fewer workers.

TFP is calculated by estimating a Cobb-Douglas production function, using data for enterprises from all manufacturing subsectors, and looking at the residuals and coefficients on various variables in an augmented production function. To compare TFP between Kenya and comparator countries, we pool the observations for all comparator countries into a single regression. The production function is estimated using a log-linear (ordinary least squares) approach.

The dependent variable is the natural log of value added, and all regressions control for the enterprises’ use of capital and workers.12 All models include country dummies to pick up

12 Following Caves (1990), value added rather than sales is used as the dependent variable. Since managers’ decisions on raw material inputs and outputs are determined simultaneously, the endogeneity problem is much more acute for intermediates than other factors of production. A more detailed discussion of this, including alternate approaches to addressing the endogeneity problem, is provided in the Technical Appendix.

43

differences in TFP among the different countries and to reduce problems associated with exchange rates. The regressions also include a full set of sector dummies. The augmented production function is

where VA is value-added in firm i in sector j in country k. Labor and capital are the number of workers and the book value of machinery and equipment. The coefficients on labor and capital, β and γ, are assumed to vary among sectors. Sector and country dummies, α and μ, are included to allow for systematic differences in productivity across countries and sectors. In some specifications, a series of enterprise-level controls (such as dummies indicating whether the firm exports and has foreign ownership) are included.

The 2003 Investment Climate Assessment for Kenya reported that there had been no visible productivity improvement visible for the average firm between 1999–2000 and 2002–03. Regression analysis showed no significant change in TFP between the two periods (Table 2-6).

Table 2-6 Pooled Value-Added Production Functions: ICA Survey 2003

(1) Full Sample (2) Subsample

Coefficient Abs. t-value Coefficient Abs. t-valueFactor Inputs

log Physical Capital 0.359** 8.89 0.361 8.24**log Employment 0.696** 10.13 0.685 8.86**

Firm AgeFirm Age (years) -0.003 0.87 -0.005 1.57

TimeYear Dummy 2002 0.067 0.52 0.099 0.77

Note: Location and sector dummies included but not reported.

Has this pattern been reversed? Are Kenyan firms more productive today than four years ago? Using the identical methodology to allow comparison with the earlier results, we estimate productivity changes over time by including a time trend dummy to the unbalanced and balanced panel of firms from the 2003 and 2007 surveys. These results are presented in Table 2-7. All values have been converted to constant 2005 dollars, using the GDP deflator, and the average annual exchange rates reported by the IMF. Our results show that TFP, measured in value added terms, has increased by 26 percent over the four-year period, indicating an average annual increase of 7 percent.

Results of TFP growth, however, given only one time series component, are sensitive to model specification. Table 2-7 also presents the results for a matched panel of firms between 2003 and 2007. Again, model (2) shows a 23 percent increase in productivity in Kenya over the four-year period. Nevertheless, by simply adding capacity utilization as an additional explanatory variable, the growth rate declines from a 7 percent annual rate to a 4 percent annual rate, and this is no longer significant. We interpret this result as follows: the use of existing capacity (rather than new investments) accounts for the increase in productivity over time.

44

Table 2-7 Total Factor Productivity: Panel Estimation(1) (2) (3)

Variable Pooled—unbalanced Balanced—model A Balanced—model BConstant 4.43*** 5.61*** 5.24***

(0.329) (0.48) (0.54)Log(Capital) 0.45*** 0.35*** 0.36***

(0.029) (0.04) (0.04)Log(Labor) 0.57*** 0.65*** 0.61***

(0.042) (0.06) (0.06)Mombasa 0.01 -0.15 -0.10

(0.119) (0.16) (0.16)Nakuru 0.16 0.06 0.09

(0.145) (0.19) (0.19)Kisumu 0.04 -0.51 -0.38

(0.146) (0.31) (0.33)Food 0.07 0.07 0.08

(0.112) (0.17) (0.17)Textile and Garment -0.18 -0.40** -0.37**

(0.117) (0.17) (0.17)Wood and Furniture -0.18 0.09 0.16

(0.169) (0.24) (0.24)Metalworking 0.16 0.03 0.05

(0.148) (0.18) (0.18)Chemical 0.17 0.10 0.22

(0.171) (0.24) (0.25)Time Dummy 0.26*** 0.23** 0.17

(0.090) (0.12) 0.13)Capacity Utilization 0.007**

(0.205)Number of firms 531 238 238Adjusted R-sq 0.7569 0.68 0.68

Further to our analysis, we examine differences jointly in TFP over time and in a broader regional context by pooling data from investment climate surveys over time in Kenya, Tanzania, and Uganda and by adding higher-income countries such as Botswana, Namibia, Senegal, and South Africa. Again we see that Kenyan firms have become 15 percent more efficient during this four-year period––an increase of approximately 4 percent annually.

We also included middle-income comparators such as Namibia and Botswana, and controlled for capital and labor inputs, sectoral differences, and differences resulting from capital usage, measured by capacity utilization. The results showed that enterprises in Kenya are far less productive than firms in Namibia, where productivity is almost double that of Kenya; and in Botswana, where firms are 22 percent more efficient than Kenyan firms (Figure 2-11).

45

Figure 2-11 Total Factor Productivity Relative to South Africa

What explains the lower productivity of Kenyan firms compared to middle-income countries in SSA? Several factors could drive productivity differentials. The role of an adverse business environment is particularly important in this regard. As shown in the next chapter, our data indicate that enterprises in Kenya continue to face an adverse business climate: total losses incurred by businesses because of power outages, transport losses resulting from theft and breakage, bribe payments, and protection payments are much higher than those for the middle-income countries in Africa, and compared to China. In Kenya, a substantial part of sales are lost because of these indirect costs, compared to a much smaller percent in China and India.

2.6 Food Sector

The Kenya Association of Manufactures, in its 2006 survey, reports that the food sector in Kenya is the largest component of manufacturing: “The sector makes the largest contribution (18%) to the number of manufacturing sector enterprises in Kenya. With a production turnover of Ksh 232bn, it is overwhelmingly the dominant sector in manufacturing, with a 70% contribution. It contributed 23% of the economy’s GDP in 2003.The formal public and private sector industries employed 82,098 employees in 2003, representing a 26% of total formal country employment.”

The Association’s report found that the main constraint facing this sector was poor infrastructure: “Poor infrastructure in crop growing areas results in massive wastage of harvested produce. The Government therefore needs to invest more in rural roads and electricity connections, while the agro-processors need to invest in cooling facilities. Currently, it is only Horticultural Corps Development Authority (HCDA) which has invested in cold storage facilities and insulated transport for horticultural produce.”

It also noted that this sector is faced with regulatory constraints such as “excise duties, non-compliance with COMESA Rules of Origin by partner states, inefficiencies at the customs points of entry, clearing and forwarding of raw materials and inputs, and infrastructure generally.” These factors and a host of others have contributed significantly to the cost of doing business, exacerbated by local authorities looking at the industry as a source of revenue for them instead of making effort to attract industries in

46

their jurisdiction through provision of sustainable services. It also noted that the food sector firms in Kenya had the highest cost of doing business in the region, leading to higher-priced products, rendering them uncompetitive in regional markets.

In this section, we reexamine the constraints on operation as reported by enterprises in the food sector; their technology and learning channels that impact competitiveness; and their investment, growth, and performance characteristics. We compare Kenya with Uganda and Tanzania, its main regional competitors. Sample characteristics of firms in the food sector are presented in Table 2-8.

Table 2-8 Food Sector: Sample Characteristics (%)Kenya Tanzania Uganda

Small 49.09 57.97 75.56Medium 13.64 17.39 14.44Large 37.27 24.64 10Percent of firms importing inputs 23.64 34.78 14.44Percent exporters 27.27 13.04 24.44Number of firms 110 69 90

We see that the majority of firms in all three countries are in the small size class, sourcing raw materials locally. A total of 27 percent of Kenyan firms export some part of their output, compared to 24 percent of firms in Uganda and 13 percent in Tanzania.

Table 2-9 Food Sector: Technology CharacteristicsKenya Tanzania Uganda

Percent with foreign technology 10.91 18.84 8.89Percent with ISO 13.64 23.19 16.85

In the last three years: Percentage of firms that Introduced a new production process 69.09 56.52 51.69Introduced a new product 65.45 62.32 56.18Became part of global network 5.45Acquired information on new technology 68.18Acquired technological innovations 62.73

Examining technology characteristics, we see that very few firms in Kenya have foreign technology or International Organization for Standardization (ISO) certification. Only 5 percent have become part of a global production network, where intermediate goods from various countries are assembled into a final product in another country. The majority of firms, however, report that they have introduced new products and processes and sought information on new technology.

Table 2-10 Food Sector Performance Characteristics: Median Performance MeasuresKenya Tanzania Uganda

Average employment growth (last three years) 0.09 0.07 0.06Capacity utilization 75.50 75.00 80.00Investment/capital ratio 0.10 0.07 0.08Labor productivity 7,943.48 5,289.13 4,508.97Capital labor ratio 5,561.64 3,165.08 1,544.60

47

Unit labor costs 0.24 0.33 0.26

Comparing food sector firms to those in Tanzania and Uganda, we see that enterprises in Kenya have grown marginally faster in terms of employment, and have higher investment rates than those in Tanzania and Uganda. Labor productivity in Kenya is higher than in Tanzania and Uganda—this is not offset by higher wages, leading to lower unit labor costs. Firms in Kenya, however, are much more capital intensive than firms in Tanzania and Uganda. The biggest reported constraints for food sector firms in Kenya are as follows:

Figure 2-12 Biggest Reported Constraints—Food Sector [[please add callout]]

As also reported in the KAM study, we see that transport problems are the biggest constraint reported by 23 percent of firms, followed by electricity, which is reported as the biggest constraint by 18 percent of firms. Other big constraints include access and cost of finance, and tax rates.

2.7 Textiles and Garments

The Kenya Association of Manufactures reports that “the Textile sector is the fourth largest sector of industry contributing 11% to the number of manufacturing enterprises. At Ksh 6.2bn, it contributed 2% of the turnover from manufacturing in 2004. Textile and Garments sector also employs the second highest number of workers after Food and Beverage at 42, 646 in 2004. This is 17.5% of total formal employment in manufacturing and about 2.5% of the total country’s employment. In 2004, the sector contributed Ksh 619.12mn or about 2.9% of government revenue collected from customs duties. In terms of export earnings, the sector (excluding EPZ enterprises) contributed Ksh 1.8bn in 2003, or about 1.3% of the country’s goods export earnings.

48

The Sector has more than 200 registered enterprises and contributes 4% to manufactured export earnings.Records at the Registrar of Industries indicate that there are 255 licensed manufacturers of yarn, fabrics, and garments. The number of registered industries includes the ones operating under the EPZ scheme. There is a large informal segment in the sector and it is estimated that there are at least 1,000 establishments. It is now estimated that there are only six integrated textile mills and two spinners in operation.”

The report also notes that “the bulk of labor used in the sector is semi-skilled. On average, wages for this labor category cost US$80 per month, which is higher than in competitor countries like China where the cost is about US$30. The cost of labor constitutes about 48.6% to total value of output in the sector, which is much higher than in competitor countries like China (15.1%), India (10.1%), Malaysia (8.8%), and Mauritius (10.7%).

In Kenya many garment investors had been attracted by the incentive to manufacture for developed countries offering textile quotas to developing countries. With the end of quotas, Kenya has a challenge to retain investors and attract buyers.

Kenya suffers from lengthy lead times for imports of raw materials and exports. The country is located far from the US market (it takes 35–40 days to ship garments from the port of Mombasa to the US), while at the same time, the country is far from the raw material sources (fabrics sourced from South East Asia).”

The current ICA sample of firms in textile and garments (predominantly garments) is distributed as shown in Table 2-8 (because of small sample size, we exclude Uganda from our analysis):

Table 2-11 Textile and Garments Sector: Sample Characteristics (%)

Kenya TanzaniaSmall 58.18 92.59Medium 10.91 1.85Large 30.91 5.56Percent of firms importing inputs 38.18 7.41Percent exporters 31.82 9.26N 110 54

Technology characteristics for this sector are presented in Table 2-12:

Table 2-12 Textile and Garments Sector: Technology CharacteristicsKenya Tanzania

Foreign technology 6.36 7.41ISO 10.91 11.11

In the last three years, percentage of firms that:Introduced new production process 61.82 59.26Introduced new product 70 75.93Became part of global network 8.18

49

Acquired information on new technology 60.91Acquired technological innovations 50.91

We see that very few firms have foreign technology or ISO certification. Most firms, however, report introducing new production processes and products in the past three years.

Firm performance characteristics are presented in Table 2-13,

Table 2-13 Textile and Garments: Performance CharacteristicsKenya Tanzania

Average employment growth 0.10 0.10Capacity utilization 70.00 77.50Investment/capital ratio 0.09 0.10Labor productivity 1709.72 1567.75Capital labor ratio 1283.63 212.59Unit labor costs 0.43 0.40

We see that garment firms in both Tanzania and Kenya have very similar performance characteristics, with low labor productivity, high unit labor costs, and low capital labor ratio. On average, firms have been growing at 10 percent per annum over the last three years.

Examining enterprise constraints, we see that competition from the informal sector is reported as the most severe constraint by more than 22 percent of enterprises in Kenya. As noted by the KAM report, more than 1,000 enterprises operate in this sector, providing unfair competition to the formal firms. Bringing these firms into the formal sector will level the playing field for enterprises serving the domestic market.

Figure 2-13 Biggest Constraint: Textile and Garments Firms[[add text callout]]

50

3 Business Climate3.1 Introduction

In 2007 a firm-level survey of approximately 650 establishments was conducted in Kenya. During the interview, firms were asked to identify the major constraints to their business activity using two types of questions. One question asked them to rate a set of approximately 20 potential bottlenecks. Table 3-14 shows the percentage of firms perceiving each constraint as major or very severe.13 The second question asked the managers to rank the top three bottlenecks to their business among the same list of constraints (Table 3-15). Both questions show a consistent picture. The top three constraints identified by the Kenyan managers were tax rates, access to finance, and corruption. Over half of the sample and one-third of respondents, respectively, identified these three as major bottlenecks. They were followed by complaints about infrastructure services (electricity and transportation), crime, and practices of competitors in the informal sector.

Table 3-14 Firms Reporting Major or Very Severe Business Constraints:All Formal Firms, Kenya, 2007 (%)[[delete hyphen in “mark-up” and close up]]

All firmsTotal Small Medium Large Dom Foreign Out. Nairobi Nairobi Manuf Retail Other Non exp Exp

Tax rates 58 62 51 52 60 31 65 54 57 63 54 58 57Access to finance (avail. and cost) 42 53 25 13 44 19 50 36 36 52 35 43 20Competitors in the inform. sect. 41 45 37 28 42 23 56 31 49 41 39 41 41Corruption 38 37 37 52 38 48 29 45 51 27 43 38 47Crime, theft and disorder 33 25 48 51 32 47 41 28 47 28 33 32 46Tax administration 32 34 28 29 33 25 39 27 43 32 29 32 32Transportation 31 27 31 52 30 40 45 21 53 32 23 29 53Business licensing and permits 28 30 25 26 27 39 28 28 28 27 29 29 20Electricity 28 26 25 43 28 26 30 26 53 16 29 26 44Customs and trade regulations 24 26 13 31 23 31 24 23 34 22 22 23 35Macroeconomic instability 19 19 19 12 18 29 23 16 26 21 15 18 28Telecommunications 16 15 13 31 15 25 17 15 24 10 18 16 19Regulation on pricing and mark-ups 15 19 4 0 17 0 27 6 0 17 10 15 20Functioning of the courts 13 10 15 25 13 14 12 13 37 8 10 11 35Political instability 10 7 13 18 9 17 13 8 16 3 13 10 11Access to land 7 7 6 15 8 3 8 7 17 3 8 7 12Labor regulations 4 2 7 10 4 8 4 5 16 3 2 3 24Regulation on hours of operation 3 4 2 0 3 5 7 1 0 4 0 3 12Zoning restrictions 3 4 0 0 4 0 7 0 0 4 0 3 0Inadequately educated workforce 3 3 3 6 3 8 2 4 9 3 1 3 6

Source: ICA Survey

Size Ownership Location Industry ExportObstacle

Such negative perceptions vary across sector of activity as well as firm characteristics. Both manufacturing and services complain about tax rates and transportation; however, the manufacturing sector also complains about energy, and the service sector complains about access to finance. At the same time, a larger share of small firms (compared to medium and large firms) perceived tax rate, access to finance, and practices of competitors in the informal sector as severe constraints. Similarly, more firms outside Nairobi indicated the same constraints as binding

1314 All results are weighted.

51

compared to those in Nairobi. Finally, the share of nonexporters complaining about tax rate and access to finance is higher than that of exporters.

Table 3-15 Ranking and Rating of Business Constraints in Kenya (% of Firms)

Indicator Ranking Indicator RatingTax Rates 54 Tax Rates 58Access to Finance (availability and cost) 33 Access to Finance (availability and cost) 42Corruption 31 Practices of Competitors in Informal Sector 41Practices of Competitors in Informal Sector 26 Corruption 38Crime, Theft and Disorder 25 Crime, Theft and Disorder 33Transportation 24 Tax administration 32Electricity 23 Transportation 31Business licensing and Permits 20 Business licensing and Permits 28Tax administration 15 Electricity 28Macroeconomic instability 11 Customs and Trade Regulations 24Telecommunications 10 Macroeconomic instability 19Customs and Trade Regulations 7 Telecommunications 16Access to land 4 Regulations on pricing and mark-ups 15Regulations on pricing and mark-ups 4 Functioning of the courts 13Political instability 3 Political instability 10Inadequately educated workforce 2 Access to Land 7Functioning of the courts 2 Labor Regulations 4Labor Regulations 2 Hours of Operation 3

Zoning 3Inadequately educated workforce 3

Another way to look at these constraints is to identify which constraints are more problemaitic for high-performing firms. To examine this issue, we divided firms into two groups: firms above the 75th percentile of labor productivity and of employment. Figure 3-14 shows that, for both categories of firms, infrastructure, tax rates, and competition from informal firms remain the biggest problems, confirming what was indicated in the previous figures.

52

Figure 3-14 Top-Ranked Constraints by Labor Growth and Labor Productivity

A similar survey conducted in 2003 enables us to analyze the evolution of such perception over the last four years. Table 3-16 shows that, in the manufacturing sector,14 tax rates have been among the top five constraints since 2003; however, today they appear to have become the top constraint. Infrastructure (electricity and transport) have become more binding constraints in recent years. They moved from the mid-lower part of the rating in 2003 to being the second and third constraints in 2007. In contrast, telecommunications has eased as a constraint and today is among the least problematic. Crime, corruption, and practices by informal firms all have improved their ratings, although still remaining among the most pressing problems. Cost of finance was the second-most-important problem for manufacturing firms in 2003, but since then has improved, at least for the manufacturing sector (Table 3-16).

1415 The survey in 2003 covered only manufacturing firms; hence, comparisons across time refer only to this sector.

53

Table 3-16 Kenyan Firms in Manufacturing Sector Reporting Major or Very Severe Constraints,2007 and 2003 (%)

Indicator Rating 2007 Dynamics Rating 2003Tax Rates 1 57 h 4 68Transportation 2 53 h 12 37Electricity 3 53 h 9 48Corruption 4 51 h 2 74Practices of Competitors in Informal Sector 5 49 - 5 65Crime, Theft and Disorder 6 47 h 3 70Tax Administration 7 43 - 7 51Access to Finance (availability and cost) 8 36 h 1 75Customs and Trade Regulations 9 34 h 11 40Business Licensing and Permits 10 28 h 16 15Macroeconomic Instability 11 26 h 8 51Telecommunications 12 24 h 10 44Access to Land 13 17 h 14 25Labor Regulations 14 16 h 15 23Political Instability 15 16 h 6 52Inadequately Educated Workforce 16 9 h 13 28

Source: ICA Survey

Over the last four years, the perceived constraints identified by Kenyan firms have somewhat changed. In 2003, manufacturing firms were concerned primarily about finance (mainly cost), corruption, crime, and taxes. In 2007, we recorded an improvement in corruption, whereas tax rates reached the top position of perceived constraints. Similarly, in 2007 we recorded major improvement in macro instability and telecom, once among the major constraints to Kenyan firms but today no longer an issue (Table 3-16).

In a comparison across countries (Table 3-17), we can see that, for all major constraints identified earlier, Kenya performs worse than the best-performing comparator countries (South Africa, China and India). The only exception is electricity, in which Kenya performs better than China and India.15 The comparison with the other comparator countries––Senegal, Tanzania, and Uganda––is more mixed, with some bottlenecks perceived to be more of a problem in Kenya and others more problematic in the other countries.

1516 Throughout the report, cross-country data refers to different years. More specifically: India—2005, China—2003, South Africa—2003, Senegal—2003, Tanzania—2005, and Uganda—2005.

54

Table 3-17 Firms Reporting Major or Very Severe Constraints: International Comparison

Constraint Kenya China (2002)

India (2005)

South Africa (2003)

Senegal (2003)

Tanzania (2006)

Uganda (2006)

Tax Rates 58 37 35 19 51 36 60Access to Finance (Availability and Cost) 42 29 22 23 72 42 56Practices of Competitors in Informal Sector 41 24 8 16 49 29 38Corruption 38 27 28 16 40 20 22Crime, Theft and Disorder 33 20 9 29 15 18 14Tax Administration 32 27 27 11 48 20 23Transportation 31 19 7 10 36 15 25Business Licensing and Permits 28 21 9 3 7 17 13Electricity 28 30 36 9 31 89 85Customs and Trade Regulations 24 19 15 17 37 14 8Macroeconomic Instability 19 30 8 33 26 21 22Telecommunications 16 24 4 3 3 6 8Functioning of the Courts 13 n/a 5 9 13 6 3Access to Land 7 15 10 4 30 17 18Labor Regulations 4 21 14 33 16 5 1Inadequately Educated Workforce 3 31 15 35 18 17 9

Source: ICA Survey

These perceived constraints have a significant impact on indirect costs. In Table 3-18, we report the estimated impact of some of these constraints on the indirect costs of firms in Kenya. The survey data shows that firms in Kenya have to bear indirect costs that amount to approximately 20 percent of their total sales. Of these, electricity (production lost because of power outages) is the main component (7.1 percent of total sales). Crime and bribes are also relevant.

Table 3-18 Indirect Costs: All Formal Sectors, Kenya, 2007 (%)Total

All F

orm

al

Firm

s

Smal

l

Med

ium

Larg

e

Dom

estic

Fore

ign

Non

exp

orte

r

Expo

rter

Out

side

N

airo

bi

Nai

robi

Man

ufac

turin

g

Ret

ail

Oth

ers

Electricity 7.1 8.1 5.1 5.1 7.3 4.5 7.1 6.9 5.5 8.0 5.9 5.5 8.6Bribes 3.6 3.9 3.2 2.0 3.7 2.4 3.7 1.8 2.6 4.2 2.4 3.1 4.3Production lost while in transit 2.6 2.0 2.2 3.2 2.9 1.6 0.3 3.1 1.1 3.2 2.0 1.6 7.5Theft, robbery or arson 3.9 3.9 4.6 2.4 4.0 2.2 3.9 3.6 2.7 4.8 2.2 3.1 5.4Security 2.9 2.9 3.2 2.7 2.9 3.0 3.0 2.0 2.3 3.3 1.7 2.6 3.6Total 20.1 20.8 18.4 15.4 20.8 13.7 18.0 17.5 14.1 23.6 14.2 15.9 29.5

Indirect Costs as % of Sales

Firm Size Ownership Export Location Industry

It is interesting to note that these indirect costs affect different types of firms differently (Table3-18). First, the survey data show that the manufacturing sector is less burdened by these bottlenecks, whereas the retail sector and rest of the economy bear a higher share of costs (14 percent and 29 percent, respectively). Second, in general, electricity is more of a problem for small, domestic firms (8 percent and 7 percent of total sales, respectively) and firms based in Nairobi (8 percent of total sales). Third, bribes affect domestic, SMEs, and nonexporters to a significant extent, whereas crime affects mostly domestic SME firms, and firms located in Nairobi. Finally, transport is more of a problem for exporters and firms located in Nairobi.

55

Over the last four years, even such indirect costs have improved. In the manufacturing sector they have decreased from almost 18 percent to 14 percent in 2007; however, not all aspects of indirect costs have improved. While the cost of crime and security has improved, electricity and corruption costs have remained the same and transportation has slightly increased from 1.4 percent to 2 percent (Figure 3-15).

Figure 3-15 Indirect Costs in 2003 and 2007—Kenya Manufacturing Sector

From an international point of view, Table 3-19 shows that Kenya achieves the highest level of indirect costs among all comparator countries. Only Tanzanian firms bear approximately the same level of indirect costs, while India, China, and South Africa, respectively, bear indirect costs that are half or one-quarter of those of Kenyan firms. These results appear to confirm the perception that electricity, transport, and crime are significant problems for Kenyan firms.

Table 3-19 Indirect Costs, All Formal Firms: International Comparison (%)

Indirect Costs as % of Sales Kenya China 2002

India 2005

South Africa 2003

Senegal 2003

Tanzania 2006

Uganda 2006

Electricity 7.1 2.0 7.8 0.9 5.1 10.7 10.2Bribes 3.6 1.9 2.1 0.3 0.4 3.4 3.7Production lost while in transit 2.6 1.2 0.8 0.8 - 1.6 1.2Theft, robbery or arson 3.9 0.3 0.2 0.6 1.0 1.1 1.0Security 2.9 0.8 1.3 0.9 1.5 2.3 1.4Total Indirect Costs 20.1 6.2 12.3 3.6 n/a 19.0 17.6Source: ICA Survey

56

3.2 Tax Rates

All around the world, businesses tend to complain about tax levels. Nevertheless, in Kenya, complaints about tax rates top all other constraints. This perception has improved over the last four years, falling from 68 percent of Kenyan firms perceiving it as a major problem in 2003. Nevertheless, Kenya and Uganda are the two countries in the list of comparators with the highest share of firms complaining about tax rates. Kenya’s average is also higher than the average of low-income countries and Sub-Sahara African countries, excluding South Africa (Figure 3-16).

Figure 3-16 Firms Reporting Tax Rate As Major or Very Severe Problem (%)

0

10

20

30

40

50

60

70

Kenya China India SouthAfrica

Senegal Tanzania Uganda LowIncome

Sub-Sahara

Countries

Perc

enta

ge o

f fir

ms

Source: ICA Survey

In 2007, almost 60 percent of Kenyan managers cited the financial burden of taxation as the most serious obstacle to their operations and growth. This perception is more pronounced among small and large firms, domestically owned firms, or those located outside Nairobi. Statistically, there is no significant difference in such perception between exporting and non exporting firms.16

Kenya has reduced the corporate tax rates in recent years17 by making it more comparable to its neighbors in East Africa. Nevertheless, as noted in the World Bank’s Doing Business 2008, objective indicators of fiscal pressure suggest that the tax burden in Kenya is higher than in most comparator countries. As a matter of fact, Kenyan firms are required to pay half (50.9 percent) of their corporate income in taxes. This amount is lower than China’s and India’s but much higher than the other African comparator countries’ amounts. For instance, South African or Uganda firms face a tax burden of just 37.1 percent and 32.3 percent, respectively (Figure 3-17).

1617 The difference in perception across firm characteristics is estimated with a probit model with robust standard error. The result of the empirical relationship is presented in Appendix A1.1718 Deloitte & Touche

57

Figure 3-17 Total Amount of Taxes As Percentage of Profit: International Comparison

0

10

20

30

40

50

60

70

80

China Ghana India Kenya Senegal SouthAfrica

Tanzania Uganda

More specifically, the high tax burden faced by Kenyan firms is due mainly to the profit tax rate (32.5 percent), which is the highest of all comparator countries, including China and India. The profit tax in China and India is less than 20 percent and in South Africa is less than 25 percent. Kenya has profit tax rates over 7 percentage points higher than in the major comparator countries. On the other hand, labor taxes and contributions in Kenya are lower than in most comparator countries (Figure 3-18).

58

Figure 3-18 Breakdown of Taxes: International Comparison

0

10

20

30

40

50

60

70


Tanzania Uganda

Profit tax (%) Labor tax and contributions (%)Source: Doing Business 2007

One potential impact of a high tax regime is the presence of a larger informal sector. As we show in Chapter 6, one of the main reasons for informality is the negative perception associated with the tax burden. According to the 2006 Kenya economic survey, the informal sector constitutes 72 percent of the working population. The sector has grown by 37.2 percent over the past 4 years to 6.5 million workers.18

3.3 Corruption

Although the ranking has improved over the last four years, Kenyan firms still place corruption among the most important constraints to their businesses. Almost one-third of firms ranked corruption among the top three constraints; 38 percent rated it as a major or severe problem. Nearly 70 percent of firms that reported corruption as a binding constraint ranked it as a top constraint.

There appears to be no significant difference in such perception among firms with respect to their size, export orientation, ownership, and legal status; however, complaints about corruption are more pronounced among firms that are located in Nairobi.

From an international point of view, Kenya and Senegal are the two countries in which the perception of corruption is the highest, with almost 40 percent of firms complaining about this problem. Countries such as South Africa, China, and India appear to enjoy a much lower level of

1819 Business Daily Africa; http://www.bdafrica.com/index.php?option=com_content&task=view&id=4527&Itemid=5822

59

http://www.bdafrica.com/index.php?option=com_content&task=view&id=4527&Itemid=5822

corruption, with 16 percent, 27 percent, and 28 percent of firms, respectively, complaining about it (Figure 3-19).

Figure 3-19 Firms Perceiving Corruption As a Severe or Major Constraint: International Comparison (%)

0

5

10

15

20

25

30

35

40

45

Kenya China India South Africa Senegal Tanzania Uganda

The results of the present Enterprise Survey are confirmed by other data sources. Both the Transparency International Corruption Perception Index and the World Bank’s Governance Indicators show an improvement in Kenya’s ranking over the last few years (Figure 3-20). Nevertheless, Kenya’s corruption rating remains the worst among all comparator countries (Figure 3-21). In 2007, Kenya’s rank in the Corruption Perception Index was 150th of 180 countries surveyed.

60

Figure 3-20 Kenya—Evolution of Transparency International Corruption Rating

1.75

1.8

1.85

1.9

1.95

2

2.05

2.1

2.15

2.2

2.25

2003 2004 2005 2006 2007

Figure 3-21 Transparency International Corruption Rating International Comparison 2007

0

1

2

3

4

5

6

Kenya Ghana Uganda SouthAfrica

China India Senegal Tanzania

Corruption takes many different forms, from making payments for utility hookups to informal payments in public procurement. In general, three-fourths of firms in Kenya reported having to make informal payments to “get things done” with rules and regulations. This costs Kenyan firms approximately 4 percent of annual sales. By international standards, this is a considerable amount. Firms in comparator countries face a lower level of informal payments, with firms in China, India, and South Africa paying less than half that amount.

The Enterprise Survey data allows us to identify the many aspects of a business that creates opportunity for illegal payments. Firms were asked if and how much they were supposed to pay as informal payments in public procurement. The answers provided by Kenyan managers show a staggering difference between Kenya and the other comparator countries. Kenyan firms are

61

required to pay approximately 12 percent of the value of a public contract as informal payments. This is higher than all comparator countries (Figure 3-22).

Figure 3-22 Bribes in Public Procurement (Percent of Contract Value)

0

2

4

6

8

10

12

14

China 2002 India 2005 Kenya 2007 Senegal 2003 South Africa2003

Tanzania 2006 Uganda 2006

Bribes to tax inspectors are also common in Kenya. Corruption in the revenue authority is common knowledge in Kenya. Most recently, the commissioner general stated that next year, tax revenues will increase in part thanks to his crackdown on “leakages.”19 Within the Enterprise Survey data one-third of the surveyed firms reported being the subject to informal payment request from tax inspectors visiting them. This is high by international standards. With the exception of India, Kenya fares worse than all other comparator countries (Figure 3-23).

1920 Report on “Plug on tax ‘leaks’ helps Kenya to clean up” by the Financial Times (downloaded November 27, 2007 at http://www.ft.com/cms/s/0/516d3b92-9884-11dc-8ca7-0000779fd2ac.html?nclick_check=1)

62

Figure 3-23 Bribe Requests from Tax Inspectors—Cross-Country Comparison (Percent of Firms)

Source: ICA Survey

It is noteworthy that frequency of inspections seems to be correlated with bribe requests from tax inspectors. The more firms are visited, the more likely they are requested informal payments. Similarly, firms that admit paying the tax inspectors also declare approximately 3 to 8 percent fewer sales for tax purposes.

Licensing represents yet another opportunity for informal payments to take place. Kenyan firms are required to not only obtain licenses when they start operation, but also to renew licenses every year. Virtually all firms in Kenya are required to renew licenses and permits periodically; however, while around one-third of them need to renew licenses with the central government, almost all of the firms in Kenya need to periodically renew licenses with the local government (Figure 3-24). When dealing with licenses, Kenyan firms are requested informal payments approximately one-quarter of the time.

63

Figure 3-24 Percentage of Firms Requesting Licenses from Local and Central Government

0

10

20

30

40

50

60

70

80

90

100

Local government Central government

One license Two licenses Three licensesSource: ICA Survey

Informal payments might also occur when utility hookups are requested. In fact, one-quarter of Kenyan firms requesting utility hookups declared they have been asked for informal payments. Overall, such payments appear to be more frequent with construction permits and water hookups and least common with electricity connections and import licenses (Figure 3-25).

Figure 3-25 Bribes, Licenses, and Utilities: Share of Firms Requesting Informal Payments When Applying for Licenses and Utilities

0

5

10

15

20

25

30

35

40

Constructionpermit

Importlicense

Operatinglicense

Singlebusiness

permit

Expatriatew ork permit

Telephoneconnection

Electricalconnection

Waterconnection

Source: ICA Survey

64

One particular aspect of corruption that seems to be unique to Kenya is the common practice of the police to request payments from trucks in transit. The Enterprise Survey questionnaire asked managers to indicate to what extent this practice was common. Although not widespread, this phenomenon is significant, with 21 percent of firms reporting having to pay such payments. The average amount paid is approximately 2.5 percent of sales and is born more by the service sector than by the manufacturing industry.

Finally, another often-forgotten aspect of corruption relates to the functioning of the courts. If we look at the general perceptions, it appears that only 13 percent of firms consider the functioning of the court a problem. Hence, it might be an issue that doesn’t seem to need to be addressed; however, if we estimate the same perception for firms that had a dispute over payment and used the courts to solve it, the share of managers concerned about the functioning of the courts rises to 33 percent, at par with crime and tax administration. Furthermore, the discontent about the functioning of the courts is quite widespread. While the majority of firms admit that courts decisions are generally enforced, fewer than one-quarter of firms consider the Kenyan courts fair, impartial, and not corrupted, and an even fewer number consider them fast (Figure 3-26).

Figure 3-26 Courts Malfunctioning: Percent of Firms Considering the Court System Efficient

0%

10%

20%

30%

40%

50%

60%

Fair, impartial anduncorrupted

Quick Affordable Able to enforce itsdecision

Source: ICA Survey

The Doing Business indicators confirm our data by showing that the number of court procedures in Kenya are among the highest and that the cost of court proceedings are also quite significant with respect to most of our comparator countries (Figure 3-27). Related to the functioning of the courts is the procedure to close a business. Even in this case, the Doing Business indicators show that the cost and the timing of such procedures are among the highest. In Kenya, it takes over four years to close a business (second only to India, where courts are notoriously slow), and the cost of such procedures amount to approximately 22 percent of the value of the estate (Figure 3-28).

65

Figure 3-27 Court Procedures and Cost—International Comparison

0

5

10

15

20

25

30

35

40

45

50


Tanzania Uganda

Procedures (number) Cost (% of claim)Source: Doing Business

Figure 3-28: Time and Cost to Close a Business—International Comparison

0

2

4

6

8

10

12


Tanzania Uganda0

5

10

15

20

25

30

35

Time (years)

Cost (% of estate)

Source: Doing Business

66

3.4 Electricity

Findings from earlier firm-level surveys have highlighted the importance of a reliable power supply. For different reasons—strong economic growth in some places, economic collapse in others, war, poor planning, population booms, high oil prices, and drought—sub-Saharan nations face crippling electricity shortages.20 And yet Kenyan firms do not indicate power as a major constraint, although 85 percent of them report experiencing power outages. This apparent contradiction is explained by the fact that two out of three firms own a generator. Hence, while only 28 percent complain about electricity, 31 percent don’t because they have their own power supply. From a policy standpoint, however, we need to consider those with a generator as firms complaining about electricity. If we do that, then electricity rises as one of the top problems facing Kenyan firms in 2007.

With the recent growth of the Kenyan economy, electricity consumption has been growing at a steady pace, reaching a growth rate of 5 percent in 2006.21 Henceforth, it is not surprising to note that over the last four years electricity has become more of a problem than it used to be. In fact, while in 2003 only 48 percent of manufacturing firms complained about this problem, in 2007 53 percent of manufacturing firms did, ranking this as the third-most-important bottleneck.

Additional survey evidence shows how serious the problem of power is. Close to 80 percent of firms in Kenya experience losses resulting from power interruptions. This is the highest value of all comparator countries, along with Uganda (Figure 3-29). In China, only 40 percent of firms report losses resulting from power outages, and in South Africa even less (13 percent).

Figure 3-29 Share of Firms That Experienced Losses from Electrical Outages

0

10

20

30

40

50

60

70

80

90

China India Kenya Senegal South Africa Tanzania Uganda

Source: ICA Survey

2021 “Toiling in the Dark: Africa’s Power Crisis,” The New York Times, July 29, 2007 (http://www.nytimes.com/2007/07/29/world/africa/29power.html?_r=2&oref=slogin&pagewanted=print&oref=slogin)2122 Africa Economic Outlook, African Development Bank/Organization for Economic Co-operation and Development

67

Furthermore, Kenyan firms experienced approximately eight power outages per month last year, each lasting approximately 4.5 hours. Hence, firms in Kenya lost on average the equivalent of approximately four days of production a month because of power outages. Given the dimension of such a problem, two out of three firms in Kenya own or share a generator and use it for 16 percent of their electricity needs (Table 3-20). Among all comparator countries, Kenya is the country with the highest share of firms owning a generator (Table 3-21).

Table 3-20 Frequency and Duration of Power Outages and Power Generator Ownership in Kenya

Indicator Total Manuf Retail Residual Small Medium Large Dom ForOut

Nairobi NairobiNon

exporter ExporterShare of firms that experienced outage 84 92 81 84 84 84 89 85 80 77 89 83 95

Share of firms with own generator

66 66 n/a n/a 34 67 91 63 79 69 65 56 84

Percentage of electricity from own or shared generator

16 16 n/a n/a 18 15 16 16 16 14 17 16 16

Average duration of outages per month (in hours)

33 31 27 39 35 29 29 33 36 36 32 33 37

Table 3-21 Power Outages and Usage of Electrical Generators International Comparison

Indicator Kenya India Senegal South Africa Tanzania Uganda ChinaShare of firms with own generator 66 52 62 9 42 25 18Percentage of electricity from own or shared generator 16 22 7 0 16 8 2Average duration of outages per month (in hours) 33 27 24 2 88 106 n/a

Source: ICA Survey

Owning a generator is, however, costly. Not only it is more expensive to generate electricity, but the capital investment of a generator accounts for approximately 3 to 5 percent of the total value of machinery and equipment. This explains why generators are mostly owned by medium and large firms.

The impact of unreliable power supply on production costs in not limited to the generation of electricity. As we have seen earlier, Kenyan firms suffered 7 percent of losses in sales because of power disruption. Small domestic firms are more affected by such disruptions. Nairobi-based firms report higher costs.

From an international perspective, the losses suffered by Kenyan firms are among the highest, as well as being the highest component of all indirect costs considered. Chinese and South African firms enjoy a much lower level of such losses (Figure 3-30).

68

Figure 3-30 Sales Lost Due to Power Outages (Percent)—International Comparison

0

2

4

6

8

10

12

China India Kenya Senegal SouthAfrica Tanzania Uganda

Obtaining a power connection is still a lengthy process in Kenya. One-quarter of the sampled firms requested a connection in the last 2 years and it took them 40 days to do so. This is double the time it takes firms in China and India, and over six times longer than it takes in South Africa (Figure 3-31). It is interesting to note that firms that were requested informal payments to set up an electric connection also report a longer period of time to receive it—79 days. The waiting time is even longer outside of Nairobi.

Figure 3-31 Days to Obtain an Electricity Connection—International Comparison

0

5

10

15

20

25

30

35

40

45

50


69

3.5 Transportation and Customs

The ability of a country to connect firms, suppliers, and consumers to global supply chains efficiently is essential to their competitiveness. Using seven measures of performance, a recent assessment of the logistics gap across countries22 has ranked Kenya 76th out of 150 economies, well behind South Africa (ranked 24th), China (30th), and India (39th). In the Enterprise Survey respondents identified transportation, together with electricity, as the two leading infrastructure constraints to doing business in Kenya. While 31 percent of firms rated it as a major bottleneck, one-quarter of respondents ranked it as one of the top three constraints. Unsurprisingly, firms outside Nairobi perceived this a major problem more often than firms located in the capital city. Particularly worried about the status of the transportation system is the manufacturing sector. Over 60 percent of manufacturing firms transport their own goods and in 2007, more than half of all of them reported transportation as a major problem, up from only 37 percent in 2003.

Supply chain problems often result in firms holding large inventories, which represent an additional cost for firms. Figure 3-32 shows that firms in Kenya hold on average 17 days of production in stocks of their most important input. While this is not high compared to other countries, inventories held by manufacturing enterprises in Kenya are much higher, with an average of 47 days. This is among the highest of all comparator countries (Figure 3-32).

Figure 3-32 Inventory Holdings—International Comparison

0

10

20

30

40

50


No

of d

ays

pf p

rodu

ctio

n

all

manufacturing

Source: ICA Survey

The strong discontent of Kenyan firms is echoed by the high direct and indirect costs they have to bear because of the quality of the transportation infrastructure. Inland transport costs in Kenya are much higher than in China and India—where they are just a fraction of what they are in Kenya—and among the highest of all comparator countries. Even worse, shipping a 40-foot

2223 Connecting to People: Trade Logistics in the Global Economy, 2007, World Bank.

70

container costs Kenyan firms much more than firms in all other comparator countries, except Uganda (Figure 3-33).

Figure 3-33 Inland Transportation Costs ($)

0

500

1000

1500

2000

2500

Kenya China India Senegal Tanzania Uganda SA

export

import

40" container cost ($)

0

500

1000

1500

2000

2500

3000

3500


export

import

Source: Connecting to People

Unfortunately, when we look at indirect costs Kenya does not perform any better. Kenyan companies lose 2.6 percent of their sales to spoilage and theft during transportation. This is the highest value of all comparator countries. China, India, and South Africa lose half that amount (Figure 3-34).23

2324 This value refers to exports.

71

Figure 3-34 Production Lost While in Transit—International Comparison

0

1

2

3

Kenya China India South Africa Tanzania Uganda

% s

ales

lost

The indirect costs presented previously affect different types of firms differently. Losses during transportation affect more large firms, exporters, and those located in Nairobi (Figure 3-35).

Figure 3-35 Transportation Losses by Firm Characteristics

0

1

2

3

4

Small Medium Large Domestic Foreign Nonexporter

Exporter OutsideNairobi

Nairobi Manufact. Retail

% s

ales

lost

Half of the transportation losses reported by Kenyan firms are due to transport delays, while the other half are due to theft during transportation. Crime, hence, is a problem also when transporting goods. As we have seen earlier, firms in Kenya admit to paying police when transporting goods. More specifically, 21 percent of them do so (25 percent of manufacturing firms). On average, they pay the equivalent of 1 percent of sales. These payments appear to protect firms from delays. Firms that admit they give informal payments in fact report a shorter time to reach Mombasa from Nairobi (three hours less).24

2425 The number of observations are too few to be able to estimate this value for the other cities.

72

Another aspect of transportation, not captured by the previously mentioned indirect costs, is the time it takes to transport goods within Kenya. Surveyed firms were asked to estimate how long it takes to ship goods from their city to Nairobi, Mombassa, Nakuru, and Kisumu. Firms reported that it takes approximately 14 hours from Nairobi to Mombasa, 4 hours between Nairobi and Nakuru, and another 4 hours from Nauru to Kisumu. Furthermore, over 10 percent of the traveling time is lost between Nairobi and Mombasa because of weighbridges, roadblocks, or any other control posts. These delays are little less if trucks are directed to Nakuru from Nairobi, 8 percent, and much less if directed to Kisumu from Nakuru, 2 percent.

The level of concerns voiced by Kenyan firms regarding transport is not unrelated to other factors such as the efficiency of customs administration. Clearing customs in Kenya is expensive. Custom costs—represented by custom clearance and ports handling—are approximately $500 ($550 for exports), almost double the cost in South Africa and India, and three times as much as in China (Figure 3-36).

Figure 3-36 Customs Costs ($)

0

200

400

600

800

1000

1200


export

import

Source: Doing Business

3.6 Crime

Crime remains a common occurrence in Kenya. The Kenya Human Rights Network said that in the first six months of 2007 approximately 300 murders took place in Kenya.25 For some observers the surge in criminal activities is linked to elections. It’s true that Kenya has always suffered cyclical violence each election year since 1992, when it became a multiparty democracy.

Regardless of its cause, crime remains a major problem for Kenya’s private sector. Although perception of crime has improved over the last few years—in 2003, almost 70 percent of manufacturing firms complained about it—today approximately one-third of Kenyan firms rate

2526 Reported by Reuters at http://uk.reuters.com/article/latestCrisis/idUKL1192304320070711 (downloaded on November 15, 2007) and http://www.time.com/time/world/article/0,8599,1680997,00.htm l (November 27, 2007).

73

http://www.time.com/time/world/article/0,8599,1680997,00.html

http://uk.reuters.com/article/latestCrisis/idUKL1192304320070711

crime as a major constraint. This is, nevertheless, even higher than South Africa, notoriously known for such a problem, and much higher than all comparator countries (Figure 3-37). China and India enjoy a much lower level of crime, with 20 percent and 9 percent of firms complaining about it, respectively.

Figure 3-37 Percentage of Firms Reporting Crime As a Major or Very Severe Obstacle to Business —International Comparison

Concerns about crime vary across firms. Compared to small firms, medium and large firms are significantly more likely to report crime as a severe constraint. Surprisingly, firms in Nairobi are less likely to perceive the severity of crime than those located in other cities. Finally, manufacturing firms also complain more about crime that firms in the other sectors of activity.26

When we look at more objective indicators of crime, the picture remains the same. Although objective measures of crime are hard to obtain, the Enterprise Survey asked a number of questions related to crime. The data shows that crime can add significantly to the costs of doing business in Kenya, both directly through theft and indirectly through security measures and protection payments extended to organized crime to prevent violence.

As indicated in Figure 3-38, Kenyan firms reported having suffered losses from crime averaging almost 4 percent of annual sales. Such costs seem significantly higher than costs experienced by firms in all other comparator countries, including South Africa, where losses were less than 1 percent. Similarly, in China and India objective cost data shows a much lower incidence of crime on sales (1/10 of the cost suffered by Kenyan firms).

2627 See Appendix A1 for a full set of regression results.

74

Figure 3-38 Crime Costs—International Comparison (Percent of Sales)

0

1

2

3

4

Kenya China 2002 India 2005 S. Africa2003

Senegal 2003 Tanzania2006

Uganda 2006

Source: ICA Survey

Similar results are obtained if we look at the indirect costs resulting from the installation of security measures within the business premises. In Kenya, about three-quarters of firms incurred extra costs to pay for security services (equipment personnel or professional security services) or to extend protection payments to organized crime. This is the highest of all comparator countries but South Africa, where 80 percent of firms had security systems installed. As in the case of direct costs, Kenya spends more on crime prevention measures than all other comparator countries (Figure 3-39). Kenyan firms spend an average of 2.9 percent of sales each year on security services, with China and India spending just 0.8 and 1.2 percent. Only Tanzanian firms come close to Kenyan firms, with 2.3 percent of sales spent on security. In addition to the cost of security, firms in Kenya also make informal payments to organized crime to prevent thefts and arson. Here once more, Kenya leads with almost 1 percent of sales paid, much higher than all other comparator countries.

75

Figure 3-39 Cost of Security Services in Percent of Sales—International Comparison

0

1

2

3

Kenya China 2002 India 2005 South Africa2003

Senegal2003

Tanzania2006

Uganda2006

If we sum together the cost of criminal acts (theft and arson) and the costs of security services (both legal and illegal) we can see than Kenyan firms do indeed face a much higher crime cost than all other comparator countries. The total impact on production costs is approximately 9 percent of sales (8 percent if we exclude protection payments to organized crime), which is considerably higher than in all comparator countries, including South Africa (where it reaches only 1.5 percent) (Figure 3-40).

Figure 3-40 Total Costs of Crime—International Comparison (Percent)

0

2

4

6

8

10

Kenya China India Senegal Uganda Tanzania* South Africa

Protection payments

Security systems

Crime

* Data on organized crime not available

76

Therefore, the security situation in Kenya has a significant impact on business decisions. In fact, 60 percent of firms responded that their hours of operation are affected by such constraints, mostly manufacturing and retails. Approximately 45 percent reported that their transport of goods is impacted—mainly in the manufacturing sector—and 40 percent mentioned that their investment decisions are affected by crime (10 percent to a significant extent) (Figure 3-41). All this evidence leads us to conclude that crime has a significant cost implication for Kenyan firms, and hence it remains a major constraint.

Figure 3-41 Impact of Security on Business Decisions (Percent of Firms)

0%

10%

20%

30%

40%

50%

60%

70%

Hrs of operation Transport Investments

To a significant extent To some extentSource: ICA Survey

3.7 Tax Administration

Despite recent reforms that have reduced the number of tax payments, tax administration remains a major burden for firms in Kenya. Approximately one-third of firms rated it as a major bottleneck, while 15 percent of them ranked it in the top three problems. Manufacturing firms complain much more, with approximately 43 percent of them rating it as a major bottleneck in 2007. This is lower than those complaining in 2003 (53 percent) thanks to the reduction in the number of payments and the decrease in the average number of visits received by tax officials. Nevertheless, manufacturing firms still complain a lot about tax administration. Across firms, small and large establishments also complain more, as well as firms outside Nairobi.

Three out of four firms in Kenya report to have been visited by tax officials in 2007. On average, they are visited once a month. There is, however, a wide variation between firms. Small firms receive more visits than medium and large firms. They report receiving more than one visit a month, while medium-sized establishments are visited once a month, and large companies once ever quarter. This seems justified by the fact that small firms in fact declare some 65 percent of sales for tax purposes, while medium and large firms are more tax obedient (with 80 percent and 85 percent declared, respectively). In 2003, the opposite was true: large firms were visited more often than small establishments. The number of visits, however, is not related to the amount of

77

sales declared for tax purposes, while the payment of bribes is significantly correlated with higher tax evasion.27

International comparisons appear to confirm Kenyan firms’ perception. All our comparator countries but China experience a much lower degree of visits by tax administration officials (Figure 3-42).

Figure 3-42 Time Spent in Dealing with Tax Officials—International Comparison (number of times visited last year)

0

2

4

6

8

10

12

14

16

18

20

China* India South Africa* Uganda Tanzania Kenya * DaysSource: ICA Survey

The tax filing system in Kenya is cumbersome. According to the Doing Business indicators, Kenyan firms spend about 430 hours to prepare forms, file, and pay taxes (Figure 3-43). At the same time, establishments need to make 41 tax payments a year to the revenue authority, while in South Africa firms need to make only 11 payments and spend 25 percent less time.

2728 See more in the section on corruption.

78

Figure 3-43 Paying Taxes—Cross-Country Comparisons (number of payments and time to fill out forms)

0

10

20

30

40

50

60

70

China India Kenya Senegal SouthAfrica

Tanzania Uganda

No o

f pay

men

ts

0

100

200

300

400

500

600

700

800

900

1000

Hou

rs

Payments (number)Time (hours)

On the other hand, VAT refunds are relatively efficient in Kenya. The majority of firms report receiving the refund within 60 days of submitting the application; however, for the rest of the firms it takes much longer, with 30 percent of the firms reporting to have to wait up to one year, and the remaining 10 percent more than one year.

3.8 Business Licensing and Permits

Although only 20 percent of managers interviewed place licenses among the top three constraints, this remains an aspect of the business environment where the government must continue its reform efforts. In fact, additional survey data shows that licensing remains a significant problem for firms in Kenya. First of all, as shown above, corruption in obtaining licenses is common in Kenya, with 25 percent of firms admitting having to pay illegal payments in such instances. Hence, it is possible that respondents discounted licensing as problems because they have already identified it as such in the corruption question. Secondly, from an international perspective, more firms in Kenya complain about business licensing today (28 percent) than in all other comparator countries, with small and large firms as well as foreign firms complaining more (Figure 3-44). Finally, over the last four years Kenyan firms have perceived this constraint as more binding, moving from 15 percent of firms complaining about it in 2003 to 28 percent in 2007.28

2839 Recall that since in 2003 the sample was comprised only of manufacturing firms, every panel comparison refers to manufacturing firms.

79

Figure 3-44 Percentage of Firms Complaining About Business Licensing and Permits—International Comparison

The Kenyan government has recognized the importance of this aspect, and a number of reforms directed at reducing the number of licenses have been approved in 2006. The reform program has eliminated 110 business and cut both the time and cost of getting building permits. The still-ongoing program will eventually eliminate or simplify at least 900 more of the country’s 1,300 licenses. This might appear to stand in contrast with the firm’s perception on the severity of the licensing constraint. Nevertheless, this is not the case, since the Enterprise Survey data was collected at the same time these reforms were implemented, hence the current data is unable to reflect the impact of such reforms.

Nevertheless, the Kenyan government has made substantial achievements in licensing, as proven by the fact that in 2007 Kenya appeared among the top 10 reformers in the world, according to the Doing Business indicators. As an example, since 2005 Kenya has been able to improve both the time (by 30 percent) and the cost to deal with construction licenses (by 20 percent). Today, Kenya is the best performer among our comparator countries in terms of number of procedures, time, and cost to deal with licenses. (Figure 3-45).

80

Figure 3-45 Duration (in Days), Cost (Percent of Income per Capita) and Number of Procedures Required to Get Business Licenses—International Comparison

0

200

400

600

800

1000

1200

1400

1600

China India Kenya Senegal South Africa Tanzania Uganda0

5

10

15

20

25

30

35

40

Duration (days)

Cost (% of income per capita)

Procedures (number)

2365

This positive impact of such reforms is also reflected in the amount of time managers need to spend in dealing with requirements imposed by government regulations. While in 2003, managers had to dedicate approximately 14 percent of their time to this, in 2007 they spent approximately half of that. By an international comparison, Kenya is much better than most of our comparator countries (Figure 3-46).

Figure 3-46 Manager’s Time Spent Dealing with Regulations

0

5

10

15

20

25

30

Kenya Kenya China India Senegal SouthAfrica Tanzania Uganda

2003

2007

Reforms notwithstanding, there are, however, other areas of the business environment related to business licensing where Kenya does not perform as well as comparator countries. For instance,

81

starting a business in Kenya remains a lengthy process. Kenya is second only to Senegal in the number of days required to start a new activity, with China, India, and South Africa requiring approximately one-third less time. The problem is not in the number of procedures necessary to follow, where Kenya performs relatively well. The bottleneck appears to be in the actual processing time. Similarly, the cost of these procedures is also relatively high when compared to both China and South Africa (Figure 3-47).

Figure 3-47 Duration (in Days) and Cost (Percent of Gross National Income per Capita) to Start a Business—International Comparison

0

10

20

30

40

50

60

70

China India Kenya Senegal South Africa Tanzania Uganda0

20

40

60

80

100

120

Duration (days) Cost (% GNI per capita)

In addition to obtaining a license, renewing them is another area where there is room for improvement. The majority of firms interviewed by the Enterprise Survey declared that they need to renew licenses (up to three) periodically from the local government. Although on average it takes about 18 days to get a domestic business license and 43 days to obtain an expatriate license, the great majority of firms can obtain a license within a month. Nevertheless, the processing time to renew licenses is much longer for a significant part of the firms in Kenya. In fact, while the majority of firms (60 percent) obtain a renewal in up to one month, for the remaining 40 percent it might take as long as 1 year. This pattern is the same for both local and central government licensing agencies ().

82

Figure 3-48 Average Time to Renew a Business License—Kenya (Percentage of firms)

0%

10%20%30%40%50%

60%70%80%90%

100%

Obtain a single business permit renew license

Up to 30 days Up to 6 months Up to 1 year

Finally, some licenses in Kenya are more expensive than in other countries. More specifically, the cost to prepare trade documents is the highest of all comparator countries (Figure 3-49).

Figure 3-49 Trade Documents Preparation Costs ($)

0

200

400

600

800


exportimport

83

More in general, approximately one-quarter of the firms in Kenya use facilitators to deal with licenses. Medium and large firms use them more than small firms. On average, it costs small firms around half of a percentage point of sales to use facilitators, with medium and large firms paying less than small firms (Figure 3-50).

Figure 3-50 Usage and Cost of Facilitators When Dealing with Licenses

0%

5%

10%

15%

20%

25%

30%

35%

40%

Small Medium large

Shar

e of

firm

s us

ing

faci

litat

ors

0

0.1

0.2

0.3

0.4

0.5

0.6

Cost

(% s

ales

)

share using facilitators cost (% sales)

84

4 Access to Finance4.1 Access to Finance from an International Perspective

Enterprise development requires the availability of external financing to bridge the gap between internally generated resources and the financing needs of dynamic firms. We begin this chapter with an examination of firm perceptions with respect to the cost and availability of external finance. We then turn to objective measures of the availability and cost of external financing. While we will try to provide some context for the supply side of external financing, much of this chapter will draw on firm-reported measures of access to finance.

The main findings are that medium and large firms in Kenya face a favorable external financing regime. Relative to a number of more dynamic economies such as South Africa and India, firms in Kenya have good access to bank finance; however, the same cannot be said of micro and small firms: these firms have considerably lower access to financing. Ongoing reforms in the sector to overcome some of the structural barriers facing these types of firms and in the sector as a whole must be carried out to consolidate the gains made over the last five years.

We begin by summarizing firm perceptions of access to and cost of finance. About 36 percent of small, medium, and large enterprises (SMLEs) and 76 percent of microenterprises in the manufacturing sector in Kenya rated access to and cost of finance as a major or severe obstacle (Figure 4-51).29 In addition, a higher proportion of formal nonmanufacturing firms report finance as a major or severe impediment to firm operation and growth: 48 percent of retail firms and 41 percent of services firms.

While cross-country comparisons of perceptions data are difficult to make, it is instructive to examine perceptions of the external financing regime in Kenya relative to a set of comparators.30

Figure 4-51 shows that relative to middle-income South Africa and the fast-growing economies of China and India, a higher proportion of manufacturing firms in Kenya report access to and cost of financing as a major or severe impediment to operation. Only about 10 percent of SMLEs in the manufacturing sector rated access to finance as a serious concern in South Africa. The corresponding estimates are 15 and 23 percent for China and India, respectively. The comparison with the other African comparators, however, is more favorable and consistent with a number of other studies on the Kenyan banking sector: 41 percent of firms in Tanzania, 55 percent in Senegal, and 60 percent of firms in Uganda report being constrained by poor access to finance. A potential concern with a comparison of national averages is that differences in the composition of manufacturing sectors across the comparator countries drive the differences in perceptions that we observe. We obtain the same ranking of countries, however, when we restrict the comparison to only firms in the garments and textiles subsectors.2930 The question on perceptions asks if the “access to and cost of financing” is a major or severe impediment to firm operation and growth. In the rest of the paper we use the phrase “access to finance” and “access to and cost of finance” interchangeably.3031 Differences in access to and the cost of finance are driven by a variety of factors, including economic growth prospects, the nature of bank ownership and regulation, and other demand and supply side factors that determine the quality of the marginal firm that has access to finance.

85

Figure 4-51: Perceptions About Access to Finance Are More Positive in Namibia Than in Many of the Comparator Countries, Especially Those in Southern African Customs Union

SMLEs

0% 20% 40% 60% 80%

Kenya

South Africa

Tanzania

Senegal

Uganda

India

China

% of firms reporting finance is serious obstacle

Microenterprises

0% 20% 40% 60% 80%

Kenya

SouthAfrica

Tanzania

Uganda

% of firms reporting finance is serious obstacle

Source: Investment Climate Surveys

Note: Cross-country comparisons are only for manufacturing enterprises

A similar picture emerges when comparing microenterprises in Figure 4-51. Only 27 percent of microfirms in South Africa report that their growth is impeded by the lack of access to finance, compared to nearly thrice the fraction in Kenya.31 Access to external financing for microfirms appears more favorable in the less-developed private sector of Tanzania, where only 46 percent of firms report access to finance as a major or severe impediment. A similar proportion of microfirms in Uganda, however, report being constrained by the lack of availability and cost of finance.

How much have perceptions changed over the last three years in Kenya and to what extent are changes related to shifts in the external finance regime? Using data from the 2003 and 2006 surveys, we restrict the analysis to a set of firms that were surveyed in each year. This addresses concerns that differences in perceptions across the two survey periods are driven by differences in the composition of samples. A total of 75 percent of these firms reported being constrained by access to finance in 2003. In 2007, only 36 percent of the same set of firms report access to finance as a major or severe impediment. The decline in the proportion of firms constrained by access to finance is large (a 40 percent decline) and suggests a significant improvement in the external financing regime.

This decline in firm perceptions is consistent with changes in the performance of the banking sector over the last five years (CEM, World Bank 2007). First, improvements in fiscal

3132 Microfirms are defined as firms with five or fewer employees.

86

management over the last five years induced a shift out of government securities and into credit to the private sector. Consequently, increased competition in the banking sector has led to a large expansion in credit to the private sector. The subsequent decline in the nominal cost of borrowing underlines both the effects of improved fiscal management and competition in the banking sector. An examination of Figure 4-52 confirms a decline in the weighted average cost of short- and long-term finance by almost 5 percentage points across the two survey years. The decline in the nominal cost of borrowing is reinforced by rising inflation between 2003 and 2007, suggesting an even larger decline in the real cost of borrowing.

Figure 4-52 Nominal Cost of Borrowing Has Fallen Over the Last Three Years

Source: Central Bank of Kenya

While firm perceptions are informative indicators of problems in the financing regime, they do not provide a sufficient description of the external financing opportunities available to firms. For that we turn to objective measures of the type and range of financing options.32

A total of 73 percent of formal manufacturing firms in Kenya (SMLEs from here on) have access to at least one credit product (overdraft, line of credit, or a loan). This is considerably higher than in both Tanzania (33 percent) and Uganda (23 percent) but the same as in South Africa (75 percent).

3233 These measures are derived from firm-reported data and as such do not address a variety of aspects of the market for external financing. For instance, our data allows us to observe whether a firm does not have a loan, an overdraft, or other debt instrument, but is uninformative as to why a firm is unable to secure financing. While one possible interpretation is that there are problems on the supply side, it is quite likely that the firm has sufficient internal resources, did not present a sound financing proposition to external financiers, keeps bad records, or does not have sufficient collateral—all demand-side constraints that preclude a conclusion that banks are unwilling to lend. Later in this chapter, we report the results of an inquiry into why firms do not use external financing options in an attempt to further elaborate on possible bottlenecks in the lending regime.

87

Predictably, only a small fraction of microenterprises report having any type of bank credit. A total of 22 percent of microenterprises had a loan, credit line, or overdraft. This is slightly higher than in Tanzania (19 percent) and considerably higher than Uganda (12 percent). Comparable information is not available for other comparator countries. So while the same proportion of firms in Kenya and Uganda report being impeded by access to and the cost of finance, a microfirm in Kenya is twice as likely as a microfirm in Uganda to have access to bank debt. The lower reports of perceptions in Tanzania potentially reflect lower demand for bank debt relative to its East African neighbors.

In addition to the basic data on whether firms have loans, firms are also asked about how they finance working capital and long-term investment requirements.33 On average, manufacturing firms in Kenya finance 51 percent of working capital and 59 percent of new investments with retained earnings. For working capital, this is considerably lower than in all the other African comparators. Perhaps surprisingly, firms in Kenya finance a lower proportion of working capital requirements out of retained earnings than firms in South Africa, and about the same proportion as firms in India (Figure 4-53).

Although firms are not heavily dependent upon retained earnings for financing working capital, the use of bank debt is moderate. The average SMLE financed about 14 percent of working capital with bank financing; this is slightly lower than the share of bank finance of working capital needs in South Africa. While the share of bank debt of working capital needs in Kenya is higher than the other East African countries (Tanzania 8 percent and Uganda 6 percent), it is only about half the level of the proportion financed by banks in India.

What makes up the gap in working capital financing is the use of trade credit. A total of 31 percent of the working capital needs of Kenyan firms are financed by trade credit. Trade credit is the leading source of working capital external to the firm. The share of working capital financed by trade credit in Kenya is higher than in any of the other comparator countries. Trade credit contributes only 12 percent and 9 percent of working capital requirements in South Africa and India, respectively. Even in the other East African neighboring countries, Tanzania and Uganda, where bank debt is less accessible, trade credit accounts for only 20 percent of working capital requirements. The use of trade credit in Senegal is considerably lower: only 5 percent of working capital needs are financed by trade credit.

The foregoing suggests that relationships between firms in Kenya support a considerably high level of working capital financing that is cheaper than bank debt. While the nature of these dense firm-firm relationships might be difficult to replicate between banks and firms, this represents a profitable niche that can be served by an expanding and improving banking sector.

3334 Retained earnings are income that the firm has left after paying wages, for intermediate inputs, and other costs. Retained earning can either be used to finance investment in the firm for expansion or could be distributed to owners through dividend payments.

88

Looking at longer-term financing, firms in Kenya finance a lower share of their investment with retained earnings than the other African comparator countries do. The average firm in Kenya finances about 59 percent of its new investment with retained earnings. This is comparable to the corresponding share in South Africa, and lower than Senegal (67 percent), Tanzania (80 percent) and Uganda (80). In comparison, the typical firm in India finances just over 50 percent of new investment with retained earnings.

Unlike working capital requirements, firms in Kenya use bank financing more intensively than firms in the comparator countries to finance new investments. More than 30 percent of new investments in Kenya are financed by bank debt, compared to only 5 percent in Tanzania and Uganda, and 15 to 20 percent in Senegal, South Africa and India (Figure 4-54).

Figure 4-53 Manufacturing Firms in Kenya Use Trade Credit to Finance a Greater Share of Working Capital Than All Other Comparator Countries

Source: Investment Climate SurveysNote: Cross-country comparisons are only for manufacturing enterprises. The data for China was incompatible with Kenya data.

89

It is instructive to explore some explanations for the more extensive use of bank financing in Kenya. Below we elaborate on two plausible demand-side explanatory variables: the cost of debt and the quality of information for assessing debt applications.34 The first of these explanations is likely the most important consideration of firms.

An important determinant for the demand for bank debt is the real cost of borrowing. The composition of financing for working capital or investment is a function of the costs of various sources of finance. Firms will use external financing intensively only if internal resources are inadequate and the cost of debt is low. While the cost of a shilling borrowed includes a variety of costs other than interest, we have data only on the interest cost of borrowing. As recent survey evidence has shown, however, noninterest charges and fees constitute a considerable share of the cost of borrowing. These charges include negotiation, commitment, legal, valuation, processing, and insurance fees. Data on each of these potential loan fees was not collected as part of this survey, but much of the scanty evidence suggests that interest costs account for the principle share in the cost of finance. As Figure 4-55 demonstrates, the real interest rate facing borrowers plays an important role in explaining the intensive use of debt financing in Kenya. Median real interest rates are the second lowest among comparator countries, with only Chinese firms facing lower real costs of borrowing.35

3435 These are by no means the only or even the most important explanatory variables. We focus on these because the data to investigate their plausibility is available.3536 We need to keep in mind that low borrowing costs in China and India (and, to a much lesser extent, Kenya) reflect structural features of the financial systems, particularly government ownership of major banks.

Figure 4-54 Manufacturing Firms in Kenya Are More Dependent on Bank Debt—Relative to Comparators—to Finance New Investment

Source: Investment Climate Assessment surveysNote: Cross-country comparisons are only for manufacturing enterprises.

90

Despite low real interest costs of borrowing, another determinant of financing is the quality of information that firms produce and maintain. Information that can be reliably used to evaluate financing propositions makes banks more willing to lend. An examination of the prevalence of externally audited accounts is a good measure of the quality of information produced by firms. A total of 88 percent of manufacturing firms in Kenya produce audited financial statements. This is considerably higher than the two East African neighbors of Tanzania (58 percent) and Uganda (44 percent), but lower than South Africa (97 percent). Only 70 percent of firms in Senegal and 84 percent in India produce reliable information by this metric. The overall quality of the information produced by firms in Kenya contributes to the favorable finance regime represented by the factors presented previously.

91

Figure 4-55 The Median Annual Real Cost of Borrowing in Kenya Is Low and Loan Duration Terms Are Reasonable


92

Another important determinant of the demand for financing is the amount of collateral required to secure a loan. Almost 90 percent of firms with loans were required to post collateral. This is among the highest of all comparator countries (Figure 4-56). The average value of collateral requirements, however, is 110 percent of loan value. The average collateral-loan ratio is considerably low compared to other countries and is lower than the average reported in 2003 (175 percent against 125 percent36 in 2007). The frequently used form of collateral is machinery and equipment. A total of 60 percent of borrowing firms posted machinery and equipment as collateral. Land and buildings are the next-most-frequently used form of collateral and was posted by more than 50 percent of secured debtors. Accounts receivable and inventories are also often accepted as collateral, with over 45 percent of firms posting them. The use of personal assets, however, represents only 28 percent of collateral posted. Unlike its East African neighbors, where modal collateral requirements are lumpy assets such as land and buildings, the use of movable assets and receivables is an important indicator of financial market development and sophistication in Kenya.

Figure 4-56 Collateral Requirements

Finally, we examine one aspect of the terms under which credit is extended: the duration of the loan. Ideally, a firm would like to match the cash flows from its investments with loan repayment obligations. For investments with long payback periods, firms would generally prefer longer loan durations. While we are unable to show the desired length of loan terms from the firm’s perspective, we can show the actual duration for the loans in the sample. Figure 4-55 illustrates that the median duration of loans in Kenya is in the middle of the pack of comparator countries.37

The typical loan in Kenya has a term of three years compared to four years in Senegal, and five years in India and South Africa. Compared to its East African neighbors, loan duration terms in Kenya appear considerably more favorable.38

3637 Manufacturing sector only.3738 It is possible that responses to this question refer to ex post durations rather than ex ante commitments by banks.3839 Once again, a cross-country comparison can produce confounds that are driven by the quality of the typical firm that has a loan. For example, if comparing only the top 10 percent of firms in country A compared to 70 percent of firms in country B; the typical borrower in country A is a higher quality firm than the corresponding typical firm in country B. The comparisons

93

Despite nearly 40 percent of firms reporting major or severe impediments to growth as a result of poor access to finance, the cross-country comparisons paint a much more favorable picture of the finance regime for firms in Kenya. It is underpinned by low real-interest costs of borrowing and the good quality of information produced by firms. Compared to its East African neighbors, Kenyan firms enjoy a superior advantage in access to finance. While Kenyan firms use as much if not more bank debt as in India and South Africa, they face shorter loan durations.

4.2 Effect of Size on Access to Credit.

Having shown that aggregate measures of access to credit in Kenya are relatively good, we examine in more detail variation in access to credit within Kenya. Essentially, we ask how firm characteristics affect the likelihood of access to finance. In this section, we also examine the relative prospects of formal nonmanufacturing firms in the retail and other services sectors with respect to the financing regime.

Firm size is an important determinant of access to credit. Access to credit is significantly more difficult for microenterprises than for small enterprises and considerably more difficult for small enterprises than for medium and large enterprises (Figure 4-57). This pattern is consistent with a well-established stylized fact across a variety of developing and even some developed economies. Consequently, microenterprises are more likely to report that access to finance is a serious obstacle. Even within the sample of formal firms, firms with fewer than 20 employees are twice as likely as firms with more than 20 employees to report finance as a major or severe constraint. While perception measures do not adequately capture dimensions of a firm’s ability to borrow, we find similar size disparities in more objective measures of access to credit. Microenterprises are less likely to have a bank account. Only 41 percent of microenterprises have a bank account, compared to 90 percent of small and 99 percent of medium and large firms. This disparity is likely a major source of differences in access to credit. A bank account represents a banking institution’s primary channel of information about a potential borrower. These results are consistent with the just-completed financial access report for Kenya, which finds that informality and individual bank-account holding are negatively related (www.fsdkenya.org/finaccess). The recent calls by the Central Bank governor for the establishment of a low-fee checking account resonates with some of the changes highlighted by the result above.39 While loan application rates are similar between micro- and small formal firms, the proportion of small firms with a loan is double the proportion of microfirms that have a loan. The differences in access to an overdraft facility are even more glaring. Among microenterprises, only 3 percent have access to an overdraft facility compared to 66 percent among medium and large enterprises.

above do not control for the quality of the typical borrower.3940 See a recent speech by the governor of the Central Bank of Kenya on bank charges and fees on August 28, 2007: http://www.bis.org/review/r071107e.pdf.

94

http://www.fsdkenya.org/finaccess

The impact of firm size is evident even when we restrict the analysis to the formal sample. Within the SMLE subsample, small firms have less access than medium and large firms. For example, only 27 percent of small firms report having any credit products, compared with 55 percent of medium and large firms. This difference is statistically significant even after controlling for other factors that might affect access (see Technical Appendix). Small firms are about 17 percentage points more likely to report that access to finance is a serious obstacle compared to large firms, ceteris paribus. In terms of objective measures, the gap widens between the smallest formal firms and large firms: small firms are about 46 percentage points less likely to have overdraft facilities and about 42 percentage points less likely to have a loan or overdraft than large firms, holding all other factors constant. Relative to large firms, medium firms are only about 13 percentage points less likely to have overdraft facilities and 16 percentage points less likely to have a loan or overdraft. Small firms are also about 10 percentage points less likely to apply for a loan; however, after controlling for other factors that might affect the likelihood of submitting a loan application, this effect is no longer statistically significant.

Do differences in access also translate to differences in the price of debt or the term of the loans received by firms of different size? Table 4-22 shows the median price and duration for all the formal firms in the sample. As the table shows, medium firms pay about 200 basis points more than large firms. Controlling in regression analysis for a variety of firm characteristics we obtain a statistically significant-size premium in the median price of debt. Small and medium firms pay about 100 basis points more than large firms, all other factors equal; however, the size advantage

Figure 4-57 Both Subjective and Objective Indicators Suggest That Access to Credit Is a More Serious Obstacle for Micro- And Small Enterprises Than for Medium-Sized and Large Enterprises

Source: Investment Climate Assessment surveysNote: Includes both manufacturing and nonmanufacturing enterprises.

95

does not translate to other aspects of the debt contract. Median loan durations are identical across all size categories and collateral requirements are slightly lower for the smallest formal firms.

Table 4-22 Median Interest Rates and Loan Duration by Firm Size

Size Categories Annual Interest Rate

Loan Duration, Months

Collateral Requirements( percent loan value)

Small (5–19 employees) 14.0 36.0 110.0Medium (20–99 employees)

14.0 36.0 120.0

Large (100 + employees) 12.0 36.0 118.5Total 13.9 36.0 120.0

4.3 Characteristics of Loan Products

Of the 657 firms in the formal sample, firms report 64 lines of credit and 208 loans. The majority of credit products observed is due to the manufacturing sector. There are only 12 lines of credit and 14 loans reported in the microenterprise sample of 124 firms. Almost all loans and credit lines to the formal sector (SMLEs) are issued by private commercial banks, with state-owned banks accounting for only an 8 percent share of all loans. Microfinance institutions dominate lending to microfirms, although private commercial banks account for 31 percent of loans to this sector (Table 4-23).

Table 4-23 Credit Line/Loan Providers Microenterprises SML EnterprisesType of Financial Institution Obs. Percent Obs. PercentPrivate commercial banks 8 31 percent 223 82 percentState-owned banks and/or government agency

5 19 percent 23 8 percent

Nonbank/microfinance institutions 12 46 percent 25 9 percentOther 1 4 percent 1 1 percentTotal 5 100 percent 272 100 percent

Source: Investment Climate Assessment surveysNote: SML enterprises include both manufacturing and nonmanufacturing enterprises; obs. = observations

About half of the loans in our sample were obtained in 2007, with the earliest loan in 1974. The size of loans varies from 20 thousand to 20 billion Kenya shillings, with the median of 5 million (Table 4-24).40 As a fraction of the estimated current value of the firm’s fixed assets the average loan-fixed assets ratio is about 36.5 percent (median 12.5 percent), which indicates a relatively moderate degree of leverage. The average and median nominal interest rates are about 13 percent. The average and median loan maturity is about three years.

Table 4-24 Loan CharacteristicsVariable N Min Median Max

4041 At 2006 exchange rates, this corresponds to a median loan size of about $70,000.

96

Year of approval 270.00 1974 2006 2007Total duration in months 271.00 2 36 240Amount at the time of approval (millions)

271.00 20,000 5,000,000 20 billion

Average annual interest rate ( percent)

272.00 5.00 13.95 100.00

Collateral as a percentage of loan amount

245.00 20.00 120.00 400.00

Source: Investment Climate Assessment surveys Note: Includes both manufacturing and nonmanufacturing enterprises.

4.4 Loan Applications and Rejections

Measures of loan application and rejection rates provide important information about impediments to access to finance. We find that a relatively small proportion of firms applied for a loan in 2007. A total of 26 percent of microenterprises applied for a loan in 2007. Among formal SMLEs, 26 percent of small firms and more than one-third of medium and large firms applied for a loan in 2007. Rejection rates are surprisingly low for microenterprises: only 13 percent of loan applications were rejected. The corresponding rejection rate for small enterprises is 21 percent while only 12 percent of large firms have loan applications rejected. The reasons for these rejections provide important insights into the types of policy interventions likely to improve access. While the sample sizes used for this analysis are too small to be conclusive, it is worth exploring reported reasons. Inadequate collateral is the most frequently cited reason for rejection of loans among small formal firms. For medium and large firms, incompleteness of loan applications accounts for nearly half of all loan rejections (Table 4-25).

Given the low rejection rates, it is surprising that application rates are not higher. One plausible explanation is that self-selection into applications produces a high-quality pool of loan applicants.

Table 4-25 Reasons for Loan Rejections Reason SML Enterprise

Small Medium-large

Collateral or cosigners unacceptable 59 percent 19 percentInsufficient profitability 6 percent 6 percentProblems with credit history or report 18 percent 6 percentIncompleteness of loan application 6 percent 44 percentConcerns about level of debt already incurred 0 percent 19 percentOther objections 11 percent 6 percentTotal 100 percent 100 percentSample size 17 16

Source: Investment Climate Assessment surveysNote: Includes both manufacturing and nonmanufacturing enterprises.

Understanding why firms do not apply for loans is an important starting point for identifying the bottlenecks that are potentially rectifiable by policy interventions. We present the reasons reported by firms that did not apply for loans in Table 4-26. Microenterprises are less likely to report “no need for loan” as a reason for nonapplication. Only 10 percent of microenterprises say they don’t need loans, compared to 38 percent of small and 60 percent of medium and large firms. This corroborates the evidence presented previously that access to credit, particularly for

97

micro- and small firms, is much worse compared to medium and large firms. Microenterprises are also more likely to be priced out of the market because of collateral requirements—43 percent of microfirms compared to 12 and 7 percent of small and medium and large firms, respectively, report that collateral requirements discouraged loan applications. Given the preponderance of fixed assets as collateral and the size and scope of microfirms, it is not surprising that collateral requirements are an impediment to accessing finance for microfirms. Regression results confirm the importance of physical fixed assets in access to finance. Ownership of land is associated with a 19 percentage-point increase in securing an overdraft or loan. In addition, firms that own land are nearly 13 percentage points less likely to have a loan application rejected (see Technical Appendix). Together with collateral requirements, the application process itself is considered a major problem by both micro- and small firms, even though small firms complain more about the interest rates. More specifically, a little more than one-quarter of small formal firms and one-sixth of medium and large firms do not apply because of unfavorable interest rates. Less than 5 percent of firms across the entire size distribution report size of loan and maturity to be deterrents to accessing external finance. This suggests the absence of any rationing of credit. Finally, 10 percent of small formal firms find the application process complicated.

Table 4-26 Reasons for Not Applying for a Loan or Line of Credit Microenterpri

seSML Enterprise

Reason Small Medium and large

No need for a loan 10 percent 38 percent 60 percentApplication procedures are complicated 24 percent 11 percent 6 percentInterest rates are not favorable 13 percent 26 percent 17 percentCollateral requirement are unattainable 43 percent 12 percent 7 percentSize of loan and maturity are insufficient 5 percent 3 percent 2 percentDid not think it would be approved 2 percent 5 percent 1 percentOther 2 percent 5 percent 7 percentTotal 100 percent 100 percent 100 percentSample size 92 231 216Source: Investment Climate Assessment surveysNote: Includes both manufacturing and nonmanufacturing enterprises.

4.5 Conclusion

This chapter has provided a description of firm perceptions of the cost and availability of financing and objective measures and correlates of access to finance. The overarching conclusion is that on average, Kenyan firms face a favorable lending regime with a high proportion of SMLEs with access to bank debt and only about one-third of firms reporting access to finance as an impediment to operation and growth. The favorable financing regime is characterized by low reported real costs of debt, a moderate duration of loans, and a high proportion of firms with good quality information. Relative to a number of richer or more dynamic economies, medium and large firms in Kenya use external financing as intensively as firms in India and South Africa.

There are important differences in access and cost of finance across firm size categories, however. Micro- and small firms face the most difficult financing prospects. As fragile,

98

relatively opaque (in terms of information) firms, it is not surprising that access is considerably lower, even though a large fraction of microfirms are financed by specialized microfinance institutions. Small formal firms also face a size disadvantage with respect to access and cost of financing.

In addition, only 41 percent of microfirms own a bank account. Since banks collect critical information about a firm’s ability to repay debt from its accounts, impediments to opening bank accounts represent a serious barrier to access. Recent calls to reduce fees and charges on checking and current accounts will go a long way in improving access to microenterprises.

Finally, the absence of unsecured lending represents another impediment to access to finance among small and microenterprises. A total of 90 percent of the loans in the data are secured by primarily by machinery, land, and buildings. While collateral loan ratios are lower for small firms, ownership of bankable security continues to deprive small firms of access to bank debt. The development of relationship-intensive lending and a simplification of loan application processes are crucial to the expansion of access to credit.

99

5 Labor Markets and Human Capital

A well-functioning labor market is vital to the success of the government’s policies to establish a globally competitive economy and provide jobs to a growing population. This chapter uses firm-level data provided by the personnel managers of surveyed firms together with individual-level employee survey data to describe the labor market in the manufacturing, retail, and services sectors. The chapter starts with a broad description of firm perceptions on a variety of labor market constraints and firm responses to these constraints, including training. The chapter then examines wage-setting behavior using firm- and worker-level data.

The data used for this chapter comes from the Enterprise Survey of 657 firms from the manufacturing (396), retail (150), and services sector (111). Because of data availability, wage-setting behavior is examined only for manufacturing firms. The individual-level data comes from 1,160 workers matched to the sampled firms in the manufacturing sector.

The average firm in the sample employs 187 workers in the manufacturing sector, 41 in the retail sector, and 65 in the services sector. Median firm size is considerably lower, with a median of 60 workers in manufacturing, 12 in retail, and 20 in the services sector. As is typical for this level of development, employment is dominated by a handful of very large firms. In the manufacturing sector, firms in the food and textiles subsectors are considerably larger than other firms with average employment of more than 200 workers. In the services sector, construction firms employ an average of 100 workers. Consistent with theories of exporting, exporters in the manufacturing sector are nearly 3.5 times as large as nonexporters (Clerides et. al. 1998).

The workforce of the typical firm in the manufacturing sector is 23 percent female, 10.5 percent part-time and 85 percent of firms report that their typical worker has more than six years of schooling. The use of part-time employment is quite extensive, with 19 percent of manufacturing, 20 percent of services, and 15 percent of retail employment accounted for by part-time workers.

The average worker in the employee sample is 33 years old, has 8.9 years of working experience, has been with the current firm for 6.2 years, and has completed 13.6 years of schooling. The gender composition of the employee sample is slightly higher than the average estimates derived from the firm sample: 33 percent of the sampled workers are female.

Labor market constraints are very low on firms’ lists of the impediments to productivity-enhancing growth. Two constraints are pertinent to this chapter: the extent to which an inadequately educated workforce and labor regulations constrain the growth and operations of enterprises. For both constraints, an overwhelming majority of firms in Kenya do not perceive either to be a major or very severe impediment to growth. Less than 20 percent of all manufacturing firms report either constraint to be a major or very severe impediment. The same holds true for the retail and services sector: less than 5 percent of retail firms and other services

100

firms report being inhibited by either constraint. For firms in Kenya, labor regulations are much more important constraints than inadequate skills.

5.1 Worker Skills

Labor market constraints are at the bottom of firms’ lists of the impediments to productivity-enhancing growth. In particular, the shortage of skilled workers is the least important constraint to the operation and growth of manufacturing firms in Kenya. About 8 percent of manufacturing firms report the shortage of skills as a major or severe impediment to growth. Figure 5-58 shows the proportion of firms that report being constrained by a poorly educated workforce in Kenya and across a set of comparator countries. It is striking that the sample of firms surveyed in Kenya in 2006 have the lowest proportion reporting major constraints tied with Uganda. Twice the proportion of firms report inadequate skills in neighboring Tanzania. Predictably, middle-income countries register a higher proportion of firms unhappy with the quality of the workforce.

We include the unweighted proportion of manufacturing firms surveyed in 2003 as a guide to trends in firm perceptions of formal education over the last three years. The data shows a very large downward trend in concerns about skills in Kenya. In particular, the proportion of firms concerned has dropped from 28 percent to a little over 8 percent; a 70 percent decline in the proportion of firms concerned about formal worker training.41

4142 Similar results are obtained if we use the panel portion of our sample.

Figure 5-58 Manufacturing Firms Are at the Bottom of the Pack Concerning Perception of Skills in the Labor Force

Source: Investment Climate Assessment surveysNote: Cross-country comparisons are only for manufacturing enterprises. The figure shows percentage of firms that report that skills shortage is a

major or severe constraint to firm operation in all the countries shown

101

Is it possible that the 2006 estimates conceal important variation across firm characteristics? We examine this possibility by looking at firm perceptions of formal worker training across sectors, firm size, export status, and ownership categories (Table 5-27). Firms in the manufacturing sector are twice as likely as retail firms and more than six times as likely as service firms to report skills shortages as a constraint to performance. Larger, foreign-owned, and exporting firms in the manufacturing sector are more likely to report that lack of skills is a major or severe impediment than other firms are.

Table 5-27 Percent of Firms Reporting Skills Shortage as a Major or Severe ConstraintFirm Category Manufacturing Retail Services

Small (5–19 employees) 4.1 4.0 2.3Medium (20–99 employees) 11.1 0 0Large (100 + employees) 9.2 33.3 0

Nonexporter 6.8 3.3 1.80Exporter 9.6 0 0

Domestic 6.8 2.7 1.3Foreign 13.2 10.0 10

Weighted Average 8.9 3.5 1.4Source: Kenya 2006 Investment Climate Survey

We advance three tentative explanations of why firms do not report any immediate concerns with the formal education of the workforce. First, it is possible that firms have made the necessary input-mix adjustments that are compatible with a low-skills workforce. Second, it is possible that the quality of formal training has risen sufficiently to match firm needs. This would suggest that firm-based training is an adequate substitute for poor formal education. Finally, it is possible that other binding constraints in the business environment dominate the importance of schooling. In other words, in an environment of low growth, driven by poor infrastructure services, we would not expect skills constraints to top firms’ lists of concerns.

We use data on the average education level of workers to examine the extent to which concerns about skills varies with skill intensity of operation. For the next two explanations, we can examine the extent to which training and employment growth affect perceptions of worker schooling.

Firms were asked to report the education level of the typical worker in the firm. We use this data to examine if perceptions of skill shortages are related to average skill intensity of the firm. Table 5-2 shows the percentage of firms reporting major or severe constraints because of skills shortages. As average education levels increase, the proportion of firms with concerns about the availability of skills declines from 10 percent for low education firms to 6.5 percent for high education firms. These differences in perceptions, however, are not statistically significant, suggesting the availability of skills is not an important bottleneck.

102

Table 5-28 Do Reports of Skill Constraints Vary by Worker EducationAverage Education Level Percentage of Firms Reporting Major or

Severe Impediment As a Result of Skill Shortages

0–3 years 10.004–6 years 8.007–12 years 8.8313 years + 6.49Note: The estimates shown above are restricted to the manufacturing sector.

We explore the extent to which concerns about low levels of skills correspond to adequate responses by the firms by looking at training. Table 5-29 shows the proportion of firms that report being constrained by inadequate worker schooling by training and above-median growth.

Table 5-29 Do Reports of Skill Constraints Vary By Training or Employment Growth?Yes No

Does firm provide training? 9.87(2.43)

7.38(1.68)

Firm employment growth above median

9.41(2.06)

7.22(1.86)

Note: Standard errors in parentheses. The estimates shown above are restricted to the manufacturing sector. Median employment growth between 2003 and 2006 is 7.7 percent.

Table 5-29 demonstrates that complaints about inadequately schooled workers are not associated with whether the firm provides training or whether the firm had an annual employment growth above 7.7 percent. This is consistent with the fact that inadequate worker education is low on firms’ list of constraints.

Another way of discriminating between some of the previous explanations is by examining the number of years of schooling of a typical worker in the typical firm in the manufacturing sector in an international perspective.42 It is important to point out that simple comparisons of years of schooling completed could under- or overestimate differences in learning achievement given cross-country differences in the quality of a year of education. As Table 5-30 shows, the typical worker in the modal firm in Kenya has between 7 and 12 years of schooling. A total of 68 percent of firms report that their typical worker has between 7 and 12 years of schooling. While this is lower than the corresponding estimates in middle-income South Africa, it is higher than the other African comparators. A higher proportion of firms report typical education levels of more than 12 years: 17 percent of Kenyan firms report average education levels of more than 12 years of schooling—higher than in any of the comparators except Uganda.

4243 Education data was not collected in the retail and services sectors.

103

Table 5-30 Percent of Firms Saying that the Average Worker in the Firm Has Completed Different Levels of Schooling

0–6 years 7–12 years >12 yearsUganda 36 45 18Tanzania 35 57 8Kenya 15 68 17

South Africa 10 78 12Source: Investment Climate Assessment surveysNote: Cross-country comparisons are only for manufacturing enterprises. Comparable data are unavailable for China and India.

An important avenue of human capital deepening is through firm-based training; however, the ability of firms to impart the requisite skills will depend on a variety of factors that include the extent of firm-level demand for skills development, the availability of external training by specialized firms, and financial and space constraints at the firm level. We examine the extent to which firms support skills development through on-the-job training. We abstract from implicit learning-by-doing (worker experience) and focus instead on formal on-the-job training programs.

Figure 5-59 Percent of Firms Providing Training/Percent Workers Trained

Source: Investment Climate Assessment surveysNote: The figure shows percentage of firms that provide firm-based training and the percentage of skilled and unskilled worker that are trained. Only data for manufacturing firms are available.

104

In Kenya, about 41 percent of firms provide training to their workers. Of the firms that provide training, nearly two-thirds of skilled workers and about 50 percent of unskilled workers received training (Figure 5-59).

To understand how training can be extended to more workers in Kenya, it is useful to identify the correlates and determinants of firm-based training. Figure 5-59 shows the proportion of firms with on-the-job training and the percentage of workers trained across a range of firm characteristics. There is a striking firm-size training provision relation: large firms are twice as likely to provide training as medium firms. The largest firms are three times as likely to provide training as small firms. The proportion of workers trained, both skilled and unskilled, does not follow the same pattern: in fact, the pattern for skilled workers appears to be reversed, with the smallest firms training a higher proportion of skilled workers.

As with firm size, exporting firms and, to a much less extent, foreign-owned firms, are considerably more likely to provide training as nonexporters and domestically owned firms, respectively. In addition, foreign-owned firms train a higher proportion of unskilled workers than domestically owned firms. Not all the relationships shown in Figure 5-59 are robust, controlling for other associations. For instance, it is likely that the exporting and foreign ownership differences in training arise from firm-size differences in each of these two categories. To address this, we investigate the correlates of firm-provided training using regression analysis. We carry out a firm- and individual-level analysis. The details of this exercise are contained in the Technical Appendix. We summarize the results of this exercise in the following text.

After controlling for a variety of factors, only the firm-size gradient is significant: a large firm is about 36 percentage points more likely to provide training than a small firm. Likewise, a medium firm is about 19 percentage points more likely to provide training than a microfirm. We also find that firms that are active in human immunodeficiency virus (HIV) prevention or testing of their workers are 13 percentage points more likely to provide training than other firms. Similar results have been observed with respect to training and HIV prevention in other countries in SSA (Ramachandran and others 2005). This variable is assumed to measure the degree to which firms are sensitive to turnover of skilled workers and the skill intensity of production.

An examination of training at the individual-worker level suggests that formal schooling is an important complement of firm- and individual-financed training. An extra year of formal schooling is associated with a 3 to 5 percentage-point increase in the likelihood of receiving training (both firm- and self-financed). Further, we find evidence of a negative gender gap in firm-provided training: female workers are nearly 5 percentage points less likely to receive firm-based training and union workers are more likely to receive firm-based training (see TechnicalAppendix).

It is instructive to evaluate the extent of training provision in an international perspective. Manufacturing firms in Kenya lag behind the comparator countries with respect to on-the-job training (). Slightly more than two out of five firms provide training in Kenya, compared to over 70 percent of firms in China and over 60 percent in South Africa. Only India and Uganda have a slightly lower share of firms that provide training.

105

Conditional on providing training, firms in Kenya compare more favorably, falling in the middle of the distribution of all comparator countries with respect to the proportion of the workforce that is trained. Interestingly, there is a moderate negative correlation between the percent of firms offering training programs and the proportion of workers trained for those firms with training programs (-0.3). For example, China ranks first with respect to the percent of firms with training programs but ranks in the bottom half for percent of workers trained. It is important to point out that the data used in this table cannot determine the quality of the training provided.

Table 5-31 Firm-Based Training: Prevalence and Percent of Workers TrainedCountry Percent Firms Offering

Training Percent Production Workers

Trained Percent Nonproduction

Workers TrainedIndia 2005 16 7 6Uganda 2006 32 61 28Kenya 2006 41 66 50Tanzania 2006 42 69 31Kenya 2003 48 - -South Africa 2003

64 45 47

China 2003 72 48 25Source: Investment Climate Assessment surveysNote: Cross-country comparisons are only for manufacturing enterprises.

A comparison of the 2003 and 2007 samples suggests that the provision of firm-based training has not changed over time. In 2007, 45.6 percent of the firms provided training, while in 2003 44.2 percent did. The difference is not statistically significant, suggesting that skills development does not appear to have changed much among formal, established firms.

5.2 Labor Regulation

Labor regulations govern the terms under which firms hire, utilize, and fire workers. These terms include remuneration guidelines, leave and overtime policies, and separation policies. We investigate the extent to which this regulatory regime is an impediment to firm operation in Kenya. Unlike concerns about the quality of the workforce, labor regulations are a moderate impediment to firm operation and growth. A total of 16.3 percent of firms in manufacturing find labor regulations to be a severe or major constraint to growth and operation (Figure 5-60). The corresponding proportion in the retail and services sectors is less than 3 percent.

As Figure 5-60 shows, labor regulations constitute a modest obstacle to the operation of firms in Kenya. In fact, this is in the bottom 5 constraints of the 17 bottlenecks presented (see Chapter 3). From an international perspective, Kenya registers an intermediate proportion of firms that are constrained by labor regulations. Crucially, Kenya is behind both of its East African neighbors on this matter: only 7 percent of firms in Tanzania and 1 percent in Uganda complain about labor regulations. The graph shows that labor regulations are considerably more constraining in South Africa and Mauritius than they are in East Africa.

106

The trend in firms’ concerns about labor regulations has been moderately downward. Nearly 22 percent of the panel sample in 2003 found labor regulations to be a major or severe constraint, compared to 15.9 percent in 2006. This difference is not statistically significant and allows us to conclude that for formal, established firms, the relative ranking of labor-regulations-related impediments has not changed over time.

Firms were asked to report an elasticity of employment with respect to two aspects of labor regulations: hiring and firing workers. Firms were asked if they would hire or fire more workers if the regulations governing both aspects were removed. In all, 2 percent of firms reported that labor regulations had affected hiring decisions, nearly 4 percent of firms report that regulations had affected their firing decisions, and 5 percent of firms had both firing and hiring constrained by labor regulations. In all, only 11 percent of firms report being constrained by labor regulations. The corresponding proportion of firms in the retail and services sectors is less than 2 percent. In general, the regulatory regime governing the hiring, remuneration, and firing of workers in Kenya appears reasonable to firms in all three sectors.

This is consistent with other evidence. The Doing Business report collects detailed information on how labor regulations affect hiring, firing, and rigidity of employment. Based upon these regulations, the report calculates objective measures that assess how strict labor regulation is in the country. Kenya is ranked 68th out of 165 countries surveyed in 2006. This ranking is considerably higher than all the middle-income comparators and much lower than Uganda. Note that, relative to the investment climate data, Tanzania performs particularly poorly by this measure (Figure 5-61).

Figure 5-60 Manufacturing Firms in Kenya Are in the Middle of the Pack Regarding Labor Regulations

Source: Investment Climate Assessment surveysNote: Cross-country comparisons are only for manufacturing enterprises. The figure shows percentage of firms that report that labor regulation is a major or severe constraint to firm operation in all the countries shown.

107

5.3 Wages

Assuming uniform worker productivity across countries, the level of wages paid to workers would determine the competitiveness of the manufacturing sector in Kenya. The level of wages and its growth trajectory is particularly important given that Kenya has been engaged in a 10-year strategy of attracting foreign direct investment. Given the advantages of a low-regulatory burden and relatively well-educated workforce (with respect to the region), it is important that wage levels remain competitive to support an attractive low-cost production environment. Rising wages that are not commensurate with productivity gains are likely to result in the flight of foreign direct investment to more favorable destinations and greater competitive pressure from imports.

5.3.1 Cross-Country Comparisons

This section compares median wages paid to various worker categories with wages in comparator countries. Once again, it is important to point out that these comparisons do not account for differences in human capital or the sectoral composition of manufacturing in the comparator countries. Figure 5-62 shows the median monthly wage in dollars paid to production workers.

Figure 5-61 Labor Regulation Does Not Appear To Be Particularly Burdensome in Kenya

Source: World Bank (2006a)

108

The median monthly wage for a full-time permanent production worker in Kenya is $116.43

Wages in Kenya are generally higher than in other East African comparators. For example, median monthly compensation in Kenya is about 30 percent higher than in Tanzania and Uganda. Median production wages in Kenya are a small fraction of median pay in South Africa.

A comparison with the economies that dominate global manufacturing is telling. Kenyan wages are higher than wages in China and India. The typical Indian production worker earns about 60 percent of the Kenyan worker’s wage, while a corresponding Chinese worker earns about 80 percent of the Kenyan production worker’s median monthly earnings. Given that the aggregate numbers above conceal differences in sample composition, we restrict the analysis to the food and garments sectors for each of our comparators. Figure 5-63 presents the results of such an analysis. Again, the ordering of median monthly wages is unchanged. Focusing on the garments sector, the median monthly wage paid in Kenya is nearly twice the wage paid in Tanzania. Median wages in Uganda, China, and India are about 70 percent of wages in Kenya.

4344 Cross-country comparisons of wages using median wages for full-time permanent production workers can be different from comparisons using average labor costs from the firm’s financial statements for several reasons. One notable difference between the two measures is that labor costs from the firms’ financial statements include wages for nonproduction workers, managers, and professionals. Other things, including the ratio of production to nonproduction workers, ratios of skilled to unskilled production workers, differences in average (relative to median) education levels, differences in ratios of full-time and part-time workers, differences in ratios of permanent and temporary workers, and many other factors, also can affect results.

Figure 5-62 Median Monthly Wages for Production Workers Are Higher in Kenya Than They Are in China and India and the Other East African Economies

Source: Investment Climate Assessment surveysNote: Cross-country comparisons are only for manufacturing enterprises. The figure shows median monthly wages in constant 2005 US$.

Deflators and exchange rates are from World Bank (2007).

109

In addition, we include a comparison between median wages (in 2005$) in the 2003 manufacturing sample. The previous estimates show that median wages in 2006 are virtually unchanged from the 2002 levels. An attempt to examine the trend in wages using firms that are surveyed in both rounds produces only 28 firms with data across the two time periods. While the sample does confirm an upward trend in production worker compensation, the sample is too small to produce reliable estimates.

5.3.2 Comparisons Across Firms in Kenya

Understanding the wage-setting mechanisms operating in the labor market is vital to the design of policies to improve the performance of the labor market. In this direction, we examine the variation of wages across firm size, unionization rates, and firm activity.

A wide range of wage-setting mechanisms has been identified in the literature. The predominant explanations include efficiency wage motivations, the role of collective bargaining, search frictions, and fairness norms. In order to identify which of these mechanisms best explains the wage patterns, the analysis looks at various factors to control for differences in monitoring costs, collective bargaining arrangements, and selective matching of high-quality workers and firms. The appendix presents detailed econometric results that test some of these mechanisms in a regression framework in which competing wage-setting mechanisms are represented by one or more control variables.

Table 5-32 shows tentative evidence of the efficiency-wage mechanism: very large firms pay median wages for production workers that are about 20 percent higher than the wages of small

Figure 5-63 Median Monthly Wages in the Food and Garments Sectors

Source: Investment Climate Assessment surveys

Note: Cross-country comparisons are only for manufacturing enterprises in the food and garments sectors. The figure shows median monthly wages in constant 2005 US$. Deflators and exchange rates are from World Bank (2007).

110

firms. The econometric analysis presented in the appendix confirms this estimate. While firm-level data shows weak results, evidence from worker-level regressions provides stronger evidence of a link after controlling for individual worker characteristics: a worker earns more in a larger firm.

Table 5-32 Median Monthly Wages by Occupation in 2005 U.S. DollarsFirm Category Production Workers Nonproduction Workers<20 93 15020–99 116 205>99 116 231

Nonexporter 104 173Exporter 116 231

Domestic 104 185Foreign 116 220

Employment growth below median

116 231

Employment growth above median

110 173

Total 116 202Note: All wages are converted to 2005 dollars using the exchange rate from the World Development Indicators.

Foreign-owned firms pay about 20 percent more for nonproduction workers than domestically owned firms. Exporters pay about 30 percent more for nonproduction workers compared to nonexporters. These differences in remuneration policy for nonproduction workers are statistically significant (see Technical Appendix). Apparent differences in pay for production workers by ownership and export status are not statistically significant.

Another interesting finding is that firms with access to external credit do not appear to pay higher wages than firms without access to external credit after controlling for other factors (see Technical Appendix). Furthermore, we find that firms that use an external auditor do not pay more or less than nonexternally audited firms. We find no evidence for rent sharing: firms with higher profits do not pay more (both at the firm and at the individual level).

Firms that provide training to workers pay higher wages to both production and nonproduction workers. This is consistent with human capital theory: both the worker and the firm share gains in productivity resulting from training.

There is little evidence to support the idea that collective bargaining has a large impact on wages rates. Firms with higher unionization rates do not appear to pay higher wages to production workers than firms that do not and union members do not appear to receive higher wages than other workers do after controlling for other variables that might affect wages.

It is important to note that unionization rates are not very high in Kenya. Among the comparator countries, only Tanzania, China, and the middle-income economies have higher unionization rates (Figure 5-64). A total of 31.3 percent of workers in Kenya’s manufacturing sector are members of a union. Unionization rates are even lower in the retail sector (7.4 percent) and

111

services sector (7.3 percent). Unionization rates are higher in larger firms. About 41 percent of workers in large manufacturing firms are unionized, compared to less than 17 percent in small firms.

Worker characteristics have a strong effect on wages. An extra year of schooling increases earnings by about 7 to 9 percent—on the high end of the distribution of returns to schooling found in other developing countries. Returns to an extra year of schooling average only 4 percent in Uganda. We also document high returns to worker experience. As has been established previously, returns to an extra year in the labor market are positive at the beginning of a worker’s career and negative toward the end of the career. An additional year of experience increases wages by about 3 to 4 percent at the beginning of the career. While we find a negative coefficient on our measure of gender, the estimate is not significant. There is no evidence of gender discrimination holding constant worker attributes. Surprisingly, we find that workers who are union members earn nearly 25 percent less than nonunionized workers. The size of this effect declines as we include worker characteristics, suggesting worker-firm matching as a possible explanation for this finding. Workers who obtained their job through the network earn significantly less than workers hired through more formal channels. Finally, we find that firm size is a still significant factor in determining worker earnings.

5.4 Absenteeism

Workers miss an average of a half-day per month because of own illness and a further half-day per month because of illness in the family. Figure 5-8 shows the comparison in worker

Figure 5-64 Unionization Rates in Kenya Are Moderate

Source: Investment Climate Assessment surveysNote: Cross-country comparisons are only for manufacturing enterprises. The figure shows the percentage of workers in the manufacturing sector that are unionized.

112

absenteeism across Uganda, Tanzania, and South Africa. The typical worker misses fewer days in Kenya compared to the rest of East Africa, but more than in South Africa. Using the estimates for South Africa as a reasonable standard, a firm in Kenya loses about eight days a year because of worker absenteeism. This is equivalent to just under 3 percent of working time in a calendar year.44

One-fifth of firms asked in the survey reported that worker absenteeism had increased because of illness. Across sectors, there was minimal variation in the prevalence of illness-related worker absence with 22 percent of firms in the manufacturing sector, 15 percent in retail, and 19 percent in the services sector. When asked specifically about HIV-related worker absence, an even smaller proportion of firms reported experiencing an uptick in worker absence. Overall, 5.5 percent of firms reported an increase in worker absence with the manufacturing sector leading with nearly 7 percent of firms compared to 3.7 percent and 3.2 percent in the services and retail sectors. The pattern of reporting high absenteeism appears to rise moderately, peaking at a firm size of about 100 (300) workers for the overall disease burden (HIV/AIDS) and then leveling off thereafter.

4445 This assumes a working calendar of just fewer than 250 days.

Figure 5-65: Absenteeism Is Lower Than in the Rest of East Africa but Higher Than in South Africa

Source: Investment Climate Assessment surveysNote: Cross-country comparisons are only for manufacturing enterprises. The figure shows the average number of days that a worker reports having been away from work because of own illness or illness in the family over the last 30 days.

113

Figure 5-66: The Disease Environment and Changes in Worker Absenteeism[[please add callout]]

0.2

.4.6

.81

Like

lihoo

d

1 2 3 4 5 6 7 8 9Log employment

Worker Illness

0.2

.4.6

.81

Like

lihoo

d

1 2 3 4 5 6 7 8 9Log employment

Worker HIV+

Likelihood of Firm Reporting High Absenteeism

Note: The figure shows the likelihood that firms report high absenteeism as a function of log firm size. The figure is constructed using a nonparametric fan locally weighted regression. The vertical line corresponds to a firm size of 100 employees.

114

6 Microenterprises in Kenya

The Kenya ICA survey included a separate survey of microenterprises, those with fewer than five employees. It is estimated that more than 80 percent of manufacturing employment in Kenya is generated by firms in this sector. Understanding the characteristics of this sector, their impediments to growth and their reasons for choosing to remain informal is essential for the government to design appropriate policies that will encourage firms to become formal, stimulate industrial growth, and reduce the current dualism in the industrial sector.

6.1 Registration Characteristics

A total of 124 microfirms were surveyed in Kenya. A total of 52 percent of these firms were located in Nairobi, 16 percent around each of the other cities: Mombasa, Nakuru, and Kisumu. The majority of firms surveyed were in manufacturing (74.2 percent); others were in the construction, retail, and service sectors. Most firms (75 percent) were sole proprietorships, others were partnerships.

Based on the information collected, the microenterprises surveyed can be subdivided into those that have any formal registration and those that do not. Firms are classified as “registered” if they have at least one of the following:

Registered name with the Office of the Registrar or other government institutions responsible for approving company names,

Registered with the Office of the Registrar, the local courts, or other government institutions responsible for commercial registration,

An operating or trade license or otherwise registered for a general business license with any municipal agency, and

Obtained a tax identification number from the tax administration or other agency responsible for tax registration.

In our microenterprise survey, we see that of the 124 firms that were surveyed in this group, 58 percent have a municipal license, less than 40 percent have a commercial license, only 28 percent have their company name registered, and only 27 percent are registered for tax purposes. Those that are registered for tax purposes have most other registrations also. In all subsequent analysis, these are defined as “formal” microenterprises. Firms select themselves to register and formalize operations: we examine the characteristics of this group versus those that choose informality (e.g., not registered for tax purposes), to identify key factors that govern these choices. As discussed in the microenterprise literature, firms may choose to formalize to gain greater access to the formal financial system, to avail themselves of public infrastructure facilities and other government services. They do so also to avoid the burden of tax evasion and noncompliance. On the other hand, firms that choose to remain informal are likely to do so when the costs of being formal and adhering to all the regulatory and tax laws are greater than the benefits provided by formality. They may also do so if they simply do not have the knowledge

115

required for this purpose. The observed choice of firms is a based on a combination of previously mentioned factors. These are examined in the following section.

6.2 Investment Climate for Microenterprises

Firms were asked to rank various areas of the investment climate to determine which constraints present the largest obstacles to enterprise operations. These rankings are presented in Figure 6-67, disaggregated by formal versus informal firms.

Figure 6-67 Business Constraints: Percent Ranking Problem To Be Major or Severe

Benefits of Formality: Access to Finance and Access to Land [[add heading level?]]

From Figure 6-67, we see that more than 80 percent of informal microenterprises rank access to finance to be a major constraint, compared to 55 percent of formal micros. Similarly, access to land is ranked as a major constraint by 33 percent of informal firms against less than 15 percent of formal micros. Do these differences in rankings reflect differential access to the formal financial sector by registered firms? Microenterprises rely significantly more on internal funds and retained earnings to finance their working capital and investment than firms in the formal sector. On average, microfirms finance about 78 percent of working capital and 85 percent of new investment with retained earnings, compared to 61 and 68 percent for formal businesses.45 Unsurprisingly, bank financing accounts for only 3.8 and 5.1 percent of microfirms’ working capital and new investment needs compared to 10.9 and 23.3 percent for establishments in the formal sector (Figure 6-68). This is likely explained by the fact that trade credit-producing relationships are long-lasting relationships and

4546 Note that the SMLE sample in this section includes the retail and other services sectors.

116

microfirms are generally very young or face a high probability of closure that potentially discourages the provision of supplier credit.

Figure 6-68 The Shares of Bank Debt and Trade Credit Are Increasing Functions of Firm Size; Use of Internal Resources Is a Declining Function of Firm Size

117


Figure 6-69 Microenterprise Financial Characteristics

As Figure 6-69 demonstrates, good account keeping is essential for obtaining access to the banking sector, particularly for borrowing. Approximately 40 percent of formal microfirms report audited accounts, compared to a negligible number of informal micros. Most formal micros have access to deposit accounts, while less than 30 percent of informal micros have this access. The distinction is even greater for borrowing—more than 40 percent of formal micros have loans or a line of credit, compared to only around 10 percent of informal microfirms. This suggests a potentially large and positive effect of formalization on access to credit; however, some caution is required in interpreting this association. On the one hand, microfirms that engage in some formal activities might generate better information on which banks can lend. Alternatively, semiformality likely signals a measure of unobserved firm quality or the desire to formalize. Informal firms, on the other hand, prefer not to engage in any levels of formality and consequently are not bankable.We examine this issue by estimating probit regressions (Table 6-33). The regressions below examine the probability of having a deposit account, and the probability of having a current loan. Explanatory variables include entrepreneur education, previous experience, land ownership, and registration status. We see that education is very significant in determining linkages—university-educated entrepreneurs are more likely than others to have bank accounts and loans, ceteris paribus. Registration status matters, after controlling for entrepreneur skills. Formal firms are much more likely to have a bank account and loans, indicating either some requirement within the banking sector, or some self-selection out of the banking sector by informal firms.

118

Table 6-33 Probit Results: Likelihood of Having Access to Banking Sector Loan Deposit Account

Intercept -1.31*** 1.39***(0.27) (0.538)

Years of experience of manager 0.01 0.01(0.03) (0.024)

Tax registration 1.76*** 0.98***(0.30) (0.290)

University degree 1.20*** 1.12**(0.47) (0.564)

Own land 0.66 -0.41(0.43) (0.402)

No. of observations 124 124Restricted likelihood -52.86 -70.7145

Access to land is also ranked as a bigger problem by informal firms, compared to formal microenterprises. Most micro firms, both informal and formal, do not own land. Only 13 percent of microfirms own land, compared to 40 percent of enterprises in the formal sector. Contrary to what we would expect, however, four times as many informal firms (36 percent) own land than formal firms (9 percent). Nevertheless, they do not have access to the banking system. Hence, the issue of registration is more important than being able to provide collateral. In addition, as shown in the finance chapter, the major constraint to access to finance for microfirms is the complexity of the application procedure.

Costs of Formality: Taxes, Burden of Inspections, and Business Licensing

When examining the ranking of constraints across firm types, we see that besides problems pertaining to electricity and transport (which impact firms both in formal manufacturing and the informal sector, and are discussed in detail in the business climate chapter), formally registered microfirms find the burden of tax administration and regulatory requirements to be a major constraint.

One of the main benefits of informality is the ability to avoid taxation. The ICA data confirms this presumption. In Kenya, informal firms declare only 20 percent of their sales, compared to over 80 percent declared by formal microfirms. What is unique is that the spread is highest in Kenya, indicating that this tax obstacle drives the choice of informality (Figure 6-70). This explains why informal firms do not report the tax burden as the top reason for not choosing formality. For them, the main reason is the minimum capital requirement and the costs of registration.

119

Figure 6-70 Percentage of Income Reported for Tax Purposes: Microenterprises

On the other hand, there is no significant difference between formal and informal microfirms on the reasons for choosing informality. Although the data shows small variations across firms, the financial burden of registration and taxation plus the minimum capital requirements are the main reasons why firms do not choose formality (Figure 6-71). Not surprisingly, in a probit regression analysis, only the administrative burden of complying with all tax laws appears as a significantly more important burden for formal microfirms than informal microfirms.

Figure 6-71 Perceived Reasons for Choosing Informality

120

Similarly, microfirms are more easily subjected to harassment by tax officials. Although microfirms are visited as frequently as formal firms—close to 80 percent of microfirms report having been visited by tax officials last year—the difference between visits to formal and informal micros is striking. Firms that have chosen informality are visited once every three to four days by tax officials,46 compared to formal microfirms, which are visited once every three to four months (Figure 6-72). This is much higher than any comparator country. In many instances, microfirms report corresponding expectations of bribes—36 percent of those that report inspector visits say that a bribe was expected at the time of the visit. Again, there is a big difference between formal and informal firms. Formal microfirms reported being requested for bribes 15 percent of the time, while informal firms almost 3 times as often (44 percent). More generally, informal firms are more subject to bribes. They pay 1 percent of sales more in illegal payments to get things done compared to formal microfirms.

Figure 6-72 Median Number of Visits/Required Meetings with Tax Officials per Year

4647 The mean values are even higher for Kenya: 141 for informal firms, 33 for formal firms. Firm interviewers reported that informal microfirms were typically visited by inspectors to get free access to firm services, such as a free meal, etc.

121

7 Investment, Innovation, and Exports

7.1 Introduction

One of the main conclusions in the World Bank’s Africa Development Indicators 2007 is that more investment is needed to sustain long-term development in SSA (World Bank 2008). Basically, without more investment, innovation, and technological upgrading, Kenyan businesses will struggle to compete and grow. Exporting is a potentially very important channel through which firm performance can improve. The domestic market for high-value-added manufacturing goods is relatively small (albeit larger than for many other African economies), and so growth will, at least in part, have to be achieved through exporting. In this chapter, we analyze the patterns of innovation, investment, and exporting, paying special attention to the constraints and bottlenecks that hamper good performance. We also look at changes in investment and exporting over time, by linking the data for 2007 to a similar enterprise survey in Kenya fielded in 2003.

Using sampling weights to account for the stratified nature of the sample, we estimate that 68 percent of the firms across all sectors made some investment in 2007. In manufacturing, 57 percent made some investment, and the average rate of gross investment is 0.10. Compared to other African countries in the past (see Bigsten and Söderbom 2006, for a survey), these results indicate that investment is fairly high. The data also indicate that many firms are actively taking steps toward modernization, in the form of introducing new production processes or acquiring technological innovations. When we combine the Kenyan dataset for 2007 with that for 2003, however, we find that the proportion of investing firms was significantly lower in 2007 than in 2003, suggesting a slowdown in investment. We do not see a similar trend for investment rates, however, and so the aggregate implications may not be overly negative. We find some evidence in the data that poor access to credit and economic instability, as well as other weaknesses in the investment climate, hamper investment.

We find that 35 percent of the firms in the manufacturing sector were exporters in 2007, and that the share of exports in total output volume is low. Thus, exporting remains at most a side activity for most firms. We find that the probability of exporting is much lower in 2007, compared with 2003. We also find that the transaction costs associated with exporting are fairly high, which is probably one reason why exporting remains quite low. Exporting appears less sensitive to variations in the quality of the investment climate than investment; however, we do find a statistically significant negative relationship between the average days to clear customs for exporters in a particular location and the propensity to export.

7.2 Who Invests, and How Much? Who Innovates?

One of the strengths of the present dataset is that it contains unusually detailed information on the nature of investments and innovations undertaken by Kenyan enterprises. Table 7-34 shows that, over the last three years, 66 percent of the firms had introduced a “new or significantly

122

improved process” of production, while 58 percent had undertaken technological innovations (note that these questions were not asked of nonmanufacturers).

Table 7-34 Summary Statistics: Full Sample and by Sector

(1) All Sectors (full

sample)(2)

Manufacturing(3) Retail and

IT

Weighted Mean Weighted MeanWeighted

MeanAny exporting? 0.35 0.35ISO certification? 0.12 0.2 0.13Introduced new process? 0.66 0.66Introduced new products? 0.71 0.71Bought fixed asset? 0.68 0.57 0.62Became part of global production network? 0.08 0.08Acquired technological information? 0.66 0.66Acquired technological innovation? 0.58 0.58Investment/capital 0.15 0.1 0.11log employees 2.45 3.68 2.09Employees 42.45 120.24 27.22Firm age (years) 17.68 21.61 8.81Any foreign ownership? 0.06 0.16 0.06LocationNairobi 0.58 0.78 0.5Mombasa 0.22 0.11 0.29Kisumu 0.1 0.05 0.11Nakuru 0.1 0.06 0.1Manufacturing subsectorFood 0.03 0.24Textile 0.02 0.19Other manufacturing 0.02 0.19Retail and IT 0.47

Note: Some questions were not asked of all types of firms, hence the variationin the number of observations.

Figure 7-73 shows that the most common “leading way” of technological innovation is through investment in new equipment, and so there is a close link between innovation and investment in physical capital. Indeed, the proportion of firms undertaking any investment in 2007 is 0.68, thus similar to the proportion of firms innovating. It is noteworthy that the proportion of firms that invest is much lower in the manufacturing sector than in retail and information technology (IT), and especially hotel and transport. The average gross investment to capital ratio is 0.15, suggesting that the fixed capital stock has grown by some 5 to 10 percent after depreciation between 2006 and 2007.47

4748 Pinning down the actual depreciation of fixed capital is always a difficult task. Most economists would probably agree the real depreciation rate for plant and equipment in manufacturing is between 5 and 10 percent. It should be noted, however, that the nonmanufacturing sectors have a large influence on this result, and measuring investment rates for these sectors is potentially problematic. For example, there are a lot of missing values in the capital stock among nonmanufacturing firms, and so investment

123

Figure 7-73 Sources of Technological Innovation

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45

Licensing or turnkey operations fromint'l sources

Other

Transferred from parent company

Licensing or turnkey operations fromdomestic sources

Developed or adapted within firmlocally

Developed with equipment ormachinery supplier

Developed in cooperation with clientfirms

By hiring key staff

Embodied in new equipment

On the whole, these numbers suggest higher rates of investment and innovation than what has been documented for manufacturing firms in other African countries (see Bigsten and Söderbom 2006 for a survey). It is clear that investment is more common, and higher, in the nonmanufacturing sectors than for manufacturing (Table 7-34). It is possible that the link between investment and innovation is weaker for nonmanufacturing firms than for manufacturing firms (the questions on process and technology innovations were not asked of nonmanufacturers, so we cannot establish this factually with these data). Nevertheless, the relatively high average investment rate in the nonmanufacturing sector is an interesting new fact coming out of these data.

With our data, comparing and contrasting the characteristics of firms that invest and innovate to those that do not is straightforward. Table 7-35 shows mean values of a range of firm-level characteristics, distinguishing between investors and noninvestors (column 1), and between firms that recently introduced new technology and those that did not. It is clear that exporting, having ISO certification, and being part of a global production network are much more common activities among investing firms. Of course, the direction of causality is ambiguous here, but these numbers nevertheless indicate a strong positive association between operating in the international market and investing. We see exactly the same pattern when we look at introduction of new technology. Unsurprisingly, the more progressive class of firms (with more

to capital ratios for nonmanufacturing firms are computed based on small subsamples. For the manufacturing sector, the average investment rate is 0.10, which is still quite high.

124

investment, exports, and international operations) tends to contain larger enterprises than the comparison group. More surprisingly, however, there is no evidence that firms with foreign ownership invest or innovate more than domestically owned firms.

Table 7-35 Who Invests? Who Innovates?1. Bought Fixed asset? 2. Introduced New Technology?(a)

No Yes No YesProportion exporters 0.31 0.39 0.32 0.45Proportion with ISO certif. 0.05 0.15 0.08 0.28Proportion part of global production network 0.05 0.10 0.03 0.12log employees 2.27 2.54 3.49 4.35Employees 25.62 50.48 65.81 183.95Firm age (years) 12.85 19.99 23.90 22.45Proportion foreign ownership 0.06 0.06 0.18 0.17Nairobi 0.51 0.62 0.76 0.82Mombasa 0.21 0.22 0.10 0.12Kisumu 0.12 0.09 0.02 0.05Nakuru 0.16 0.07 0.11 0.01Food 0.03 0.03 0.24 0.28Textile 0.04 0.01 0.23 0.17Other manufacturing 0.03 0.02 0.18 0.20Retail and IT 0.55 0.44 0.00 0.00

Observations(b) 323 458 171 225Note: Sampling weights were used for these calculations.(a) This question was not asked of firms in retailing, IT, hotel, transport, or other.(b) Casewise missing values ignored.

Investment and Innovation: Results from Regression Analysis

To shed further light on the characteristics of firms that invest and innovate, we now consider the results from various regressions. We have three basic goals with the analysis. First, we want to contrast the characteristics of firms that do invest and innovate with those that do not. Second, we want to investigate if observed differences in investment across locations and sectors correlate with measures of the quality of the investment climate across locations and sectors. Third, we want to test if, conditional on basic observables, the propensity to invest has changed since the last time a major survey was fielded in Kenya, i.e., in 2003. To achieve these goals, we rely on empirical specifications in reduced form, in which the explanatory variables include firm size (measured as the natural logarithm of total employment), firm age, and dummy variables for location, sector, and foreign ownership.48

Column 1 in Table 7-36 shows the results from a probit regression in which the dependent variable is equal to one if the firm did any investment in fixed capital in 2007, and

4849 Of course, these specifications should not be interpreted as representing a “decision rule” for investment in a structural sense. They may, however, be interpreted as reduced-form specifications. Suppose, for example, the decision rule is such that investment can be written as a function of the demand for the firm's products (e.g., Adda and Cooper 2003, pp. 192–3). If demand varies across sectors (for example), then a regression of investment on sector dummies can be given a reduced form interpretation.

125

zero otherwise.49 The marginal effect associated with firm size is positive and significant at the 1 percent level, confirming that large firms are more likely to invest than small firms. Firm age has a very small and statistically insignificant effect on the likelihood of investment. Thus, young firms appear just as likely to invest as old firms, conditional on the other variables in the model. We also find the coefficient for foreign ownership small and insignificant. Looking across locations, it is clear that the propensity to invest is highest in Nairobi and lowest in Nakuru, conditional on the other explanatory variables.

Similar regression results show that there is a positive and significant correlation between exporting and investment (and innovation) conditional on age, foreign ownership, location, and sector. As soon as a firm size is added to this set of explanatory variables, however, this correlation falls drastically and becomes insignificant. These findings can be interpreted as follows: if we compare firms of the same age, and with the same ownership, sector, and location, firms that export tend to also invest (or innovate). But if we compare firms of the same size (and the same age and with the same ownership, sector, and location), there is no evidence that exporting firms tend to invest more than nonexporting firms. Of course, size in itself may not be exogenous: firms that have a high propensity to invest, innovate, export, or any combination of the three, may be large precisely because of this, and so exactly what these results imply regarding the causal mechanisms between exporting and innovation is not entirely clear.

Table 7-36 Investment and Innovation Regressions

(1) Any Investment (probit)

(2) Technological Innovation (probit)

(3) New Production Processes (probit)

(4) Investment to Capital Ratio (tobit)

Partial Effect Partial Effect Partial Effect Partial Effect

log employment 0.046** 0.154** 0.111** 0.005Firm age 0.000 -0.005** -0.003* -0.001Any foreign ownership -0.012 -0.009 0.065 0.011Textiles -0.215** -0.074 -0.043 -0.121**Metal and machinery 0.010 -0.058 -0.033 -0.059Wood -0.112 -0.064 -0.028 -0.072Other -0.108 -0.012 -0.012 -0.077Retail and IT 0.078 -0.029Hotel and transport 0.217** 0.155*Mombasa -0.126* -0.058 -0.105 -0.132**Nakuru -0.294** -0.486** -0.334** -0.227**Kisumu -0.135* 0.201* 0.187* -0.118*log likelihood -487.1 -226.6 -227.7 -191.6Observations 781 396 396 519

We next attempt to investigate if firm characteristics can account for the observed variation in investment across firm characteristics. Column 2 in Table 7-36 shows results from a probit

4950 To facilitate interpretation we report for continuous variables (size and age) marginal effects, and for each dummy variable the discrete change in the estimated probability of exports as the dummy variable changes from 0 to 1.

126

where we model the likelihood of acquiring any technological innovations over the last three years, while in column 3 we model the likelihood of introducing “any new or significantly improved process,”50 Large and young firms are much more likely to acquire technological innovations than small and old firms. We obtain qualitatively very similar results for the introduction of new production processes. That new firms have higher innovation rates than old firms is not surprising, yet we know that most new firms are small, which, according to the results, may hamper innovation. There might therefore be a case for facilitating innovations among new, small firms. There is no evidence that foreign ownership matters in the context of innovation. Both measures of innovation are lower in Nakuru than elsewhere, conditional on the other explanatory variables in these models.

Column 4 in Table 7-36 shows results from a tobit regression in which the investment to capital ratio is the dependent variable. No evidence is found here that investment rates vary systematically with size, age, or ownership, but there are clear signs of sometimes large differences across locations (investment is highest in Nairobi, lowest in Nakuru) and sectors (investment is highest in the residual sector, followed by food processing, and lowest in textiles).

Investment and the Investment Climate

It is clear from the analysis above that investment and innovation vary substantially across sectors and regions. We now investigate if this can be attributed to variation in the investment climate across sectors and locations. From a methodological point of view, analyzing the effects of the investment climate is difficult, partly because reliable measures of the investment climate are hard to obtain, and partly because there may be little genuine variation in the investment climate across firms in, say, the same town. These and other issues that arise when analyzing the role of the investment climate are discussed in detail by Escribano and Guasch (2005).

To obtain measures of the investment climate quality we use the data on perceived constraints to the operations of the establishment (see chapter 3), as well as objective data on the frequency of power outages or theft, for example (more on this below). Following the suggestion of Escribano and Guasch (2005), we use these data to construct aggregate measures of the investment climate, which are constant across firms in the same location, sector, and size group. This ensures, for example, that two small garment enterprises located next to each other face the same investment climate, which would seem quite realistic. Also, as noted by Escribano and Guasch (2005), smoothing the data in this way should have the additional advantage of mitigating endogeneity bias. Of course, this procedure rules out the use of location and sector dummies in the regressions, as these would be collinear (by construction) with the investment climate variables. Our approach below is to replace these dummies by our measures of the investment climate to see if the results are consistent with the idea that a poor investment climate worsens firm performance.51

5051 These questions were asked only of manufacturing firms.5152 From a strictly statistical point of view, such a procedure does not guarantee the results can be interpreted as causal, since investment and exporting may vary across locations and sectors for reasons other than the investment climate. On the other hand, documenting significant location and sector effects is of limited policy interest, if we do not know why these arise.

127

We begin by using data on perceived constraints to the operations of the establishment, constructing dummy variables that are equal to one if a particular constraint (there are 17 such constraints [see chapter 3]) is rated as a “very severe obstacle,” and zero otherwise. We then regress each dummy on a set of dummy variables for sector, location, and size, and compute predictions based on these regressions. These predictions are interpretable as estimates of the proportion of firms in the population belonging to a particular sector-size-location cell for which a given constraint is “very severe.” We take these estimates as our measures of the investment climate quality.

Next, we investigate if these investment climate measures correlate with observed differences in investment across location, size, and sector. Using the probit results shown in Table 7-36, column 1, we compute predicted probabilities of investment across regions, size groups, and sectors, holding the remaining explanatory variables constant. We then check if these predictions—which thus vary only across regions, size groups, and sectors—appear correlated with the investment climate data. Figure 7-74 illustrates the relationship between investment and access to finance. In panel A of the graph, the vertical axis shows the predicted probability of positive investment (based on the probit results in Table 7-36, column 1). There is a very clear negative relationship between location-sector-size investment and access to finance: sector-size-location categories that have high proportions of firms indicating financing is a very severe obstacle have much lower predicted probabilities of positive investment relative to sector-size-location categories for which these proportions are low. Panel B of Figure 7-74 shows how predicted investment rates across locations, sectors, and size groups correlate with the finance access variable. The correlation is again negative, but statistically weaker than for the probability of positive investment, shown in panel A. In Figure 7-75, we show the relationship between perceived macroeconomic instability and the predicted proportion of positive investment (panel A) and expected investment rate (panel B). There is quite a strong negative correlation in both cases.

128

Figure 7-74 Investment and Lack of Finance

a) Any investment (probit predictions) b) Investment to capital ratio (tobit predictions)

.2.4

.6.8

Pr(

Inve

stm

ent>

0)

0 .2 .4 .6 .8Finance access 'very severe obstacle' (proportion)

0.0

5.1

.15

.2E

(I/K

)

0 .2 .4 .6 .8Finance access 'very severe obstacle' (proportion)

Figure 7-75 Investment and Macroeconomic Instability a) Any investment (probit predictions) b) Investment to capital ratio (tobit predictions)

.2.4

.6.8

Pr(

Inve

stm

ent>

0)

0 .2 .4 .6Macro instability 'very severe obstacle' (proportion)

0.0

5.1

.15

.2E

(I/K

)

0 .2 .4 .6Macro instability 'very severe obstacle' (proportion)

129

Table 7-37 shows the results based on objective constraints to the operations of the firm. Losses in domestic transit resulting from breakage, spoilage, or theft have the right signs (negative) and are often statistically significant. Thus, firms operating in places or sectors plagued by high such losses tend to invest and innovate less than other firms. In contrast, power outages, insufficient water supply, and corruption appear not to be hampering investment or innovations.

Table 7-37 Objective Investment Climate Constraints and Investment

Proportion of Firms in {location, size group, sector} for Whom […] Is True (year 2006)

(1) Any Investment (probit)

(2) New Production Processes (probit)

(3) Technological Innovation (probit)

(4) Investment to Capital Ratio (tobit)

Partial Effect

Partial Effect

Partial Effect

Coeff.

Experienced power outages 0.186 1.173 a 1.988 a 0.146

Experienced insufficient water supply 0.383a -0.064 0.108 0.072

Experienced some loss in domestic transit because of breakage or spoilage

-0.184 -0.304 a -0.362 a -0.002

Experienced some loss in domestic transit because of theft

-0.396 a -0.805 a -1.127 a -0.076

Made informal payments to police when transporting goods

0.863 a 0.343 0.411 0.366 a

Made protection payments to organized crime

0.262 -0.690 -1.490 a 0.001

Experienced theft, robbery, vandalism, or arson

0.371 -0.497 -0.606 0.291 a

Note: The numbers in the table are based on regressions where the dependent variable is modeled as a function of ln employment, firm age, foreign ownership, and an investment climate constraint variable. Cells with an “a” index indicate effects that are significant at the 5 percent level of significance or better.

Changes over Time

Finally, we pool the Kenyan data for 2007 with the data from 2003, and consider changes in investment over time. Our main objective here is to see if, controlling for the basic observable variables, there is any indication that the propensity to invest, or the investment rate, has changed systematically since 2003.52 We thus adopt the same reduced form specifications that were used in Table 7-36, but, for comparability, drop sectors that were not covered in both periods. Results are shown in Table 7-38. Column 1 shows marginal effects from a probit regression modeling the probability to invest. The estimated partial effect on the time dummy for year 2007 imply that the likelihood of a positive investment is some 11 percentage points lower in 2007 compared with 2003.

5253 We do not have comparable data on innovations and the introduction of new process for the two periods, and so have to restrict ourselves to modeling the propensity to invest and the investment rate.

130

This effect is significantly different from zero at the five percent level. The corresponding effect is much weaker when modeling the investment rate (column 2): the tobit coefficient on the year dummy is negative but very small and far from significant. The fact that the proportion of positive investments has fallen while the average investment rate appears constant indicates that the distribution of investment across firms has changed between the two periods. Thus, while in 2007 there were more zeros compared to 2003, there were also more large investments, keeping the mean approximately constant.

Table 7-38 Investment Regressions Based on Pooled 2002 and 2006 Data

(1) Probit: Investment>0

(2) Tobit: Investment/Capital

Partial Effect Coefficientlog employment 0.055 a 0.011Firm age -0.002 -0.002 a

Any foreign ownership 0.115 0.045Mombasa -0.107 -0.071Kisumu -0.055 -0.084Nakuru -0.256 a -0.174 a

Textiles -0.207 a -0.111 a

Metal and machinery -0.045 -0.059Wood -0.082 -0.066Year dummy 2006 -0.105 a -0.004

log likelihood -352.4 -177.8Observations 558 544

a=statistically significant

7.3 Who Exports, and How Much? What Are the Main Bottlenecks?

We now turn to exporting. Returning briefly to Table 7-34 we see that 35 percent of the manufacturing firms in the sample do some exporting, which is similar to results reported for other African manufacturing datasets in the past (Bigsten et al. 2004). 53 A detailed comparison of the characteristics of exporters and nonexporters is shown in Table 7-39.

5354 Nonmanufacturing firms are excluded from the analysis of exports because of lack of data.

131

Table 7-39 Who Exports? (Manufacturing Only)

1. Any Exporting? 2. Exporters Only:Any Exporting Outside SSA?

No Yes No YesHas ISO certif. 0.11 0.33 0.35 0.32Introduced new process? 0.61 0.75 0.70 0.77Introduced new products? 0.66 0.79 0.71 0.82Bought fixed asset? 0.54 0.62 0.61 0.63Became part of global production network? 0.04 0.13 0.19 0.11Acquired technological information? 0.59 0.78 0.77 0.78Acquired technological innovation? 0.53 0.66 0.64 0.67Investment/capital 0.12 0.07 0.07 0.07log employees 3.10 4.74 4.78 4.73Employees 58.48 233.26 313.99 204.62Firm age (years) 18.46 27.36 25.66 28.03Any foreign ownership? 0.11 0.24 0.19 0.26Nairobi 0.79 0.76 0.72 0.77Mombasa 0.08 0.16 0.19 0.15Kisumu 0.07 0.03 0.01 0.03Nakuru 0.06 0.06 0.08 0.05Food 0.30 0.14 0.17 0.13Textile 0.20 0.17 0.27 0.14Other manufacturing 0.15 0.25 0.26 0.25

Observations(a) 307 146 102 43Note: Sampling weights were used for these calculations.(a) Casewise missing values ignored.

Column 1, which does not distinguish between different destinations of exporting, confirms the finding documented in the previous section that exporting firms are more prone to invest and innovate. It is clear that exporters are much larger, in terms of employment, and older than nonexporters. There is also evidence of foreign ownership being more common among exporters; however, as we shall see below, this is mainly because firms with foreign ownership tend to be relatively large. Focusing on exporters only, there are few obvious differences across regional exporters and exporters outside SSA. This is somewhat surprising, as one might expect exporting outside SSA to be more demanding, in terms of technological capacity and efficiency, than regional exporting. It is true that exporters outside SSA are somewhat larger and younger than regional exporters, on average, but these differences are not statistically significant.

The dataset is very rich in terms of its coverage of bottlenecks and transaction costs incurred by exporters. Table 7-40 shows detailed summary statistics on exporting and transaction costs associated with exporting (for these calculations, nonexporters have been deleted from the data). On average, exporting firms export about 29 percent of their total production, and so there is little specialization in exporting. On average, 84 percent

132

of exported goods are shipped to other African countries, and so few exporting concentrate on exporting outside Africa. Output lost in transit because of breakage or theft is 2 percent of the total output value, on average. The cost of clearing customs is as high as 6.4 percent of the consignment value. Combining the two thus results in a transaction cost equal to about 8 percent of the output value, which is significant. It takes nearly a weak to clear customs, on average.

Table 7-40 Characteristics of Exporters (Manufacturing Only)Weighted Mean Sample Median Sample Standard

Deviation

Share of output exported 28.9 20 27.00

Share of exports to Africa 83.9 100 34.24

Time from startup until entry into exporting (years)

2.55 1 3.25

Percentage of export output value lost in transit because of breakage or theft

2.04 0 5.11

Average number of days for clearing customs

5.78 3 5.69

Cost of clearing customs, as percentage of consignment value

6.36 3 7.27

Note: These questions were not asked of the microfirms, or of the nonexporters. The number of observations is 145, except for “time from startup until entry into exporting,” for which we have 51 observations, as we only include firms that were started in 1990 or later. Sampling weights were used for the computation of the means, but not for the medians or standard deviations.

We now look into the factors that correlate with exporting. Table 7-41 shows probit marginal effects based on the sample of manufacturing firms. Clearly, firm size plays an important role here, being highly significant, both from a statistical and economic point of view. The estimated marginal effect of log employment is 0.19, which is interpretable as a semielasticity: that is, an increase of employment by 1 percent is associated with an increase in the estimated likelihood of exporting by 0.19 percentage points. This may not seem a large effect; however, the sample variation in firm size is large. To give one specific example, the predicted probability of exporting for a firm with 10 employees is 0.05, while for a firm with 100 employees it is 0.40. This shows that the size effect is economically important. Söderbom and Teal (2000) argue that the positive association between firm size and exporting arises because firms face significant fixed costs to entering the exports market because of bureaucratic procedures, the establishment of new marketing channels, and the need for a certain minimal size to meet export orders.

133

Table 7-41 Modeling the Decision to Export (Manufacturing Only)

(1) Any Exporting (probit)Partial Effect

log employment 0.192**Firm age 0.003Any foreign ownership 0.047Textiles 0.187*Metal and machinery 0.336**Wood 0.197Other 0.257**Mombasa 0.062Nakuru 0.187*Kisumu 0.005

log likelihood -192.9Observations 453

Now consider the role of firm age. One of the issues central to the policy debate on how to stimulate exports is whether breaking into exports markets takes time, perhaps because firms need to learn about marketing strategies, distribution channels, etc., or whether firms can export soon after inception. Our analysis indicates that the effect of firm age in this context is very small. Table 7-41 shows that the marginal effect of age on the likelihood of exporting is 0.0026, indicating that an additional year of operation raises the likelihood of exporting by about 0.3 percent. For a new firm (age = 1) with 50 employees, the predicted probability of exporting is equal to 0.20, while for an otherwise identical firm that has existed for 10 years, this probability is 0.23, evaluated at sample means of all other explanatory variables. The age effect is statistically significant at the 10 percent level but not at the 5 percent level. There is thus some weak evidence that the likelihood of exporting increases with firm age, and so while young firms are in a somewhat worse position than older firms when it comes to entering the exporting market, this effect is not dramatic. Somewhat surprisingly, the coefficient on foreign ownership is totally insignificant. Thus, the notion that foreign ownership plays a role for exporting, perhaps because firms with foreign ownership have better access to foreign markets and foreign technology, is not supported by the data.

Exports and the Investment Climate

We now ask how the investment climate variables described above correlate with exporting. As can be seen in Table 7-42, using objective investment-climate variables leads to somewhat stronger results. While most marginal effects are insignificant, the two that are significant both have the right sign. There is thus a negative and significant relationship between the likelihood of exporting and the average days to clear customs.

134

The point estimate of -0.03 means that cutting the clearing time in customs by one day would increase the probability of exporting by some three percentage points. Cutting the waiting time by five days would increase the likelihood of exporting by about 15 percentage points, thus increasing the proportion of exporters from about one-third to approximately half of all firms.

Table 7-42 Objective Investment Climate Constraints and Exports

(1) Exports (probit)

Partial EffectPower outages -0.121

Insufficient water supply -0.282

Any loss in domestic transit because of breakage or spoilage

-0.214

Any loss in domestic transit because of theft

0.351

Informal payments to police when transporting goods

-0.006

Protection payments to organized crime

-0.397

Firm experienced theft, robbery, vandalism, or arson

0.487

Uses own Web site in communications with clients and suppliers

0.585 a

Average days to clear customs for exporters

-0.026 a

Average cost to clear customs for exporters ( percent of consignment value)

-0.011

Percent of exported consignment value lost in transit because of breakage, spoilage, or theft

0.013

a=statistically significant

7.4 Conclusions

It is well known that performance varies dramatically across firms in Africa. Some firms perform very well, recording high profits, high productivity, and fast growth; others fall behind. To enable more firms to become successful, key goals for African industrial policy should be to facilitate the adoption of new technology and the penetration of new markets. Countries that cannot break out of a situation in which most firms supply the domestic market with low value-added products are unlikely to see a significant expansion of jobs in the sector, or to have manufacturing play a major role in reducing poverty.

135

In this chapter we have reported key findings on investment, innovation, and exporting in Kenya. We have found that investment is relatively high, and that exporting firms tend to innovate and expand at higher rates than nonexporters. We have shown that exporting, investment, and innovation are much more common among large firms than small ones. In the last 10 to 15 years the number of small, typically informal, firms has grown enormously in Kenya. While this certainly has generated jobs, this class of firms is not a seedbed for exports, investment, and innovation. We have found empirical evidence indicating that investment is higher in areas where the investment climate is relatively good.

136

8 RecommendationsThis report identified the main constraints to private sector development in Kenya on the basis of a survey conducted in 2007 of approximately 650 firms in 4 locations in the country. Perception and objective indicators confirm that tax rates, finance, and corruption remain the three most important impediments. Electricity and transport are also identified by Kenyan managers as the main infrastructure constraints, as well as security and licensing. In order to address these bottlenecks, we propose the following recommendations.

Policy Matrix

Problem Action Taken Recommendations

1. Taxes.

High taxes are the most reported bottleneck in Kenya. Objective indicators of fiscal pressure suggest that the tax burden in Kenya remains higher than in most comparator countries. Kenyan firms are still required to pay half of their corporate income in taxes, an overall amount that is much higher than in the other African comparator countries.

Kenya has recently reduced the tax rates faced by corporations. The most important reforms in corporate income taxes have focused mainly on lowering rates in efforts to combat global competition. Rates have been reduced from a peak of 45 percent in 1990 to around 30 percent today.

Conduct an in-depth study of the effective marginal rate of taxation to determine the extent of excessive taxation—taking into account also rebates and fiscal incentives—across different sectors.

2. Finance

Although we observed a decline in the proportion of firms constrained by access to finance since 2003—from 75 to 36 percent—access to credit is significantly more difficult for smaller firms. A total of 90

Regulations for private credit bureau operators are expected to be gazetted in June 2008, and the Central Bank is preparing to license the first operator(s)—supported by the bank’s Financial and Legal Sector technical assistance project. In parallel, IFC is working with the Kenya

Enhance credit information infrastructure. With the new regulations to be issued by the Central Bank of Kenya, support will be given to enabling private bureaus to operate in Kenya. Several international and local companies are reporting keen interest in applying for the license once regulations are issued. IFC is providing assistance to the Kenya Bankers Association on the code of

137

percent of microenterprises and 60 percent of small firms declare they need loans, compared to 40 percent of medium and large firms. Hence, firm size is an important determinant of access to credit. Among microenterprises, only 3 percent have access to an overdraft facility compared to 66 percent among medium and large enterprises. Similarly, only 27 percent of small firms report having any credit products, compared with 55 percent of medium and large firms. Together with collateral requirements, the application process itself is also considered a major problem by both micro- and small firms.

Bankers Association on ensuring effective participation of banks in the new credit bureau(s).

A reform program for the companies’ registers has been developed and is being implemented by the Registrar General, supported by the Bank’s Financial and Legal Sector technical assistance project.

Focus on land registration and transfer issues has intensified following the 2007 elections.

conduct and strategy for bank participation in the private bureaus.

Upgrade corporate registries, collateral registries, and public record systems. The scope of financial information infrastructure should include efficient access to corporate information, registries of secured lending charges and court records, etc.

Computerize property registration process and simplify taxes and fees. Efficient land registries and the ability to easily perfect and transfer land titles are an important vehicle to provide property owners with access to collateralized financing. Backlog and paper-based records necessitate that all history of transactions relevant to the property must be checked every time.

Promotion of new products is being undertaken by the private sector—as in the revolution in m-banking created by Safaricom’s M-Pesa and Equity Bank’s vast increase in client outreach. New product development has also been supported by DFID/World Bank Financial Deepening Trust in such areas as weather insurance, warehouse receipts, and payments system innovation.

Such capacity building is being supported by the Bank’s MSME project, which promotes

Promote the application of innovative products and technology to expand access to finance. Capacity building for banks and microfinance institutions in the use of different lending technologies, secured lending, leasing, mortgage finance, and, in the longer run, the promotion of new products such as warehouse receipts or weather insurance are likely to have a large impact on financial depths.

In order to promote improved access by small businesses to the products and services of commercial banks, facilitate the provision of capacity building for small businesses to better understand the requirements of banks (how to

138

lending to SMEs and business development services.

FSD Kenya is supporting the Central Bank in developing improved disclosure of bank interest rates and fees to all consumers in order to encourage stronger competition in financial markets to ensure that the benefits from increasing productivity and efficiency in the banking sector give rise to benefits in pricing among consumers. Very preliminary indications suggest that price-based competition is leading to downward pressure on prices.

approach banks for business loans and how to use bank services) and prepare them for a relationship with a commercial bank. The training would be organized in collaboration with local training business-development service providers and training institutes, and should be sponsored by local banks.

Increase transparency as regards interest rates and noninterest charges and fees (such as negotiation, commitment, legal, evaluation, processing, and insurance) on checking and current accounts.

Establish a clear time table for the creation of credit bureau

Facilitate capacity building for banks in order to develop and market new products

3. Corruption

Although its ranking has improved over the last 4 years, Kenyan firms still place corruption among the most important constraints to their businesses. Nearly 70 percent of firms that reported corruption as a binding constraint ranked it as a top constraint. Corruption takes many different forms, from making payments for utility hookups to informal payments in public procurement. In general, three-fourths of firms in Kenya reported having to make informal

The GoK now posts on the ministries’ Web sites

1) Conduct an in depth study of corruption in the country

2) Give prosecutorial power to the Anti Corruption Authority and publicize better the successful anti corruption cases

3) Tax administration: continue reforms aimed at

Minimizing human contact between taxpayer and officials and make the process more transparent by relying heavily on information technology to file tax returns;

Establishing independent internal and external audits;

139

payments to “get things done” with rules and regulations. This costs Kenyan firms approximately 4 percent of annual sales. The Enterprise Survey data allows us to identify the many aspects of a business that creates opportunity for illegal payments. For instance, Kenyan firms are required to pay approximately 12 percent of the value of a public contract as informal payments. One-third of the surveyed firms reported being the subject to an informal payment request from tax inspectors visiting them. This is high by international standards. Licensing represents yet another opportunity for informal payments to take place. When dealing with licenses, Kenyan firms are requested informal payments approximately one-fourth of the time. Furthermore, one particular aspect of corruption that seems to be unique to Kenya is the common practice of the police to request payments from trucks in transit. Finally, the share of managers concerned about the functioning of the courts—out of those that actually used them—rises to 33 percent, at par with crime and tax administration.

all information on contracts, including names of contractors, decisions of the Procurement Appeals Board, bidders and tender outcomes, and contractors’ performance. Contracts above 5 million are posted on the Web site hosted at Treasury. Plans at an advanced stage for local hosting at PPOA.

The GoK is proposing to blacklist companies found to have been involved in cases of corruption in accordance with the new procurement law. No requests for blacklisting have been received so far from any of the procuring entities.

GoK is taking steps to accelerate implementation of a more coordinated and prioritized e-government Initiative, with public access to procurement as one of the highest near-term prioriies.

and

Introducing organizational changes of the Revenue Authority: incentives for high performers, sanctions for corrupt behavior, career development, and competitive salaries;

4) Public procurement: continue reforms aimed at

Reviewing procurement rules with the goal of simplifying tender documents; reducing the minimum value of a contract for single source; and introducing anticorruption laws, performance standards, and sanctions;

Improving transparency in public-private interactions through e-procurement, publication of tender documents and tenders received, and public participation in negotiations;

Introducing a vetting system (conducted by international firm, possibly with involvement of civil society) to prequalify companies interested in bidding for government contracts to address conflict of interests and fraudulent companies;

Establishing an independent tender evaluation and auditing and monitoring of unit rates; and

Supporting a greater level of integrity and professionalism

140

GoK is taking steps to establish mobile visiting courts in sparsely populated areas. Visiting courts at Mpekotoni, Archers Post, Wamba, Loitokitok, Dadaab, Kakuma, and Marimanti upgraded to fully pledged courts.

GoK is taking steps to incorporate alternative dispute resolution mechanisms and provision of legal aid schemes. The Rules Committee considering experiences learned from a study tour, with a view toward coming up with a pilot project in the Milimani Commercial Court.

GoK is taking steps to launch comprehensive wireless-based public information hubs in districts and constituencies, with public access to government a high priority.

Restructuring and privatization of Telkon Kenya is ongoing.

among multinationals and domestic companies through professional associations, codes of conduct, monitoring and benchmarking, and integrity pacts.

5) Police

Have observers join the trucks to monitor the request for bribes. Use recording systems to monitor traveling time and illegal behavior.

Establish computerized checkpoints to make the process more transparent and quick with the least-possible interaction between truck drivers and police officials. Educating truck drivers about the automated system will also reduce the harassment faced by them.

Install an electronic weighing station.

Involve associations engaged in trucking operations in sensitizing truck drivers to comply with the rules and regulations.

Establish an independent police complaints commission entrusted with following up on the implementation of the reform program.

Reduce the discretionary power of police

Conduct effective educational campaigns of traffic rules to reduce ability

141

of police to extort bribes

6) Utility

Complete the liberalization of fixed line telephony.

Privatize some forms of service delivery (utility hookups).

Use citizen report cards to assess the performance and quality of services and monitor progress. Publish progress reports periodically based on surveys of customer and timely audits.

4. Electricity

Close to 80 percent of firms in Kenya experience losses because of power interruptions. This is the highest value of all comparator countries. Consequently, almost 70 percent of firms have generators, which are costly to obtain and operate. Power disruption costs Kenyan firms approximately 7 percent of sales. In a cross-country comparison, these losses are among the highest.

Since June 2006, KPLC has been managed by an international management services contractor.

KPLC made a profit in both FY 2005/6 and FY 2006/7. During FY2007/8, KPLC’s performance has continued to improve—e.g., network losses reduced. At present, government is providing a nontargeted subsidy to electricity consumers of K Sh .60 per kWh.

The conversion of the Electricity Regulatory Board to the Energy Regulatory Commission on July 7, 2007, was an important step in the right direction. By taking this action, the government has moved the power sector one step closer to being overseen by an

Increase public investment in energy generation, transmission, and distribution to increase connectivity;

Encourage increased private financing and investment in the energy sector—today, the private sector accounts for 12 percent of power supply;

Establish clear rules for private generators’ “open access” to transmission network, the concept of which was established in the Energy Policy; and

Ensure electricity pricing maintains the financial viability of power companies, while protecting the most vulnerable consumers.

Develop the legal framework for investments in energy

Consider using the least cost development plan to increase investments in energy

142

independent regulatory entity with the clear legal authority for performing the universal tasks of a regulatory entity: setting tariffs and quality of service standards, and licensing operators.

The establishment of the Rural Electrification Authority in 2007 has transferred the responsibility for rural electrification from KPLC to the Rural Electrification Authority. The authority will manage the Rural Electrification Fund, with an expected annual turnover of K Sh 4 billion (US$60 million) of government funds plus any donor funds made available to it.

5. Transport

Managers identified transportation, together with electricity, as the two leading infrastructure constraints to doing business in Kenya. The strong discontent of Kenyan firms is echoed by the high direct and indirect costs they have to bear because of quality of the transportation infrastructure. Even worse, shipping a 40-foot container costs Kenyan firms much more than firms in all other comparator countries, except Uganda. Unfortunately, when we look at indirect

Road

Kenya’s transport system is important not only for Kenya itself but also for its regional partners. Kenya requires about US$1,500 million to bring the primary road network back into a good condition, while the Road Maintenance Fuel Levy yields about $200 million per annum.

The Kenya Roads Board finalized the overall road-sector expenditure strategy and investment plan, but this now awaits governmental adoption.

The government passed the Kenya Roads

1) Roads

The Ministry of Finance should establish a system for ensuring proper investment planning and management. This would, among other things, involve[[please change “c” and “d” to “a” and “b”, etc., throughout]]

c. Issuing guidelines for a minimum level of preparation of projects before they are submitted for budget requests, including compatibility with the overall sector strategy and development plan, economic analysis, confirmation of having prepared detailed designs based on field investigations and the

143

costs Kenya does not perform any better. Kenyan companies lose 2.6 percent of their sales to spoilage and theft during transportation. This is the highest value of all comparator countries.

Act and has established three roads authorities, namely, the Kenya National Highways Authority, Kenya Rural Roads Authority, and the Kenya Urban Roads Authority, in order to efficiently manage the entire road network in Kenya.

The chair and members of the board of the three authorities have been appointed and the authorities will become operational as soon as the three CEOs and senior staff are appointed. The government has also adopted a detailed road sector policy and strategy paper that will form the basis for future programs and reforms in the sector.

Ministry of Roads has drafted a policy paper on the involvement of the private sector in the management of truck weigh stations and axle load control.

Aviation

The Kenya Airports Authority and the Kenya Civil Aviation Authority have been given financial autonomy and now retain the revenue generated from their operations, which had been previously remitted to the Treasury.

The responsibility for

required bidding documents, and readiness for implementation; and

d. Strengthening the institutional structure for implementing the guidelines. A special unit could be set up to screen projects submitted for budget funds. Such a unit would have a close working relationship with the MTEF and BSD units in the Ministry of Finance and would be the repository of a multiyear rolling investment program containing an inventory of appraised and priority-ranked projects for budgetary consideration in the future.

Ongoing reforms in the roads sector should be expedited. This would involve

d. Expediting the operationalization of the Kenya National Highways Authority, the Rural Roads Authority, and the Urban Roads Authority;

e. Strengthening the residual Ministry of Roads to perform its overall policy, planning, and coordination role;

f. Promoting the use of long-term output and performance-based contracting or concessioning for maintenance and management of the major

144

passenger, baggage, and mail security screening at the airports has been transferred from the police to Kenya Airports Authority, allowing for better monitoring, control, and training of security staff.

The regulations for safety and security have been harmonized and adopted by all four member countries of the East African Community.

road network by the private sector, starting with the Northern Corridor.

The government should improve governance in the road sector:

f. Strengthen the Engineers’ Registration Board and empower it further to discipline and sanction engineers and firms who perform poorly and violate its charter with regard to professional conduct and ethics. The same would be true for the Association of Consulting Engineers.

g. Assist the construction industry in establishing a professional body for construction contractors (national construction council, or a contractors’ registration board) and strengthen it to engage in self-regulation.

h. Develop a comprehensive construction industry development policy and establish a dedicated construction industry development board to implement the policy to enhance the performance of the construction industry.

i. Ensure regular updating of contractors’ qualifications and capacity; facilitate training in different aspects of construction

145

and supervision techniques; and reprimand poor performance.

j. Approve policy on private sector participation in the management of weigh stations and control of axle load regulations.

Improve public transportation system

Facilitate more private involvement in transport

2) Port and Maritime

Expedite conversion of Kenya Ports Authority to a landlord authority.

Concession the Mombasa container terminal(s), the dockyard and marine services, and the bulk oil terminals.

Streamline cargo clearance procedures and remove the police escort system for transit cargo by road (except for hazardous and military supplies).

Introduce risk-based targeting for cargo inspection and verification.

Implement a harmonized customs clearance system and one-stop border posts in accordance with COMESA protocols.

Review and ensure compatibility of local maritime regulations with the

146

International Maritime Organization treaties.

3) Aviation

Expedite safety and security enhancement at Jomo Kenyatta International Airport and strengthen the Kenya Civil Aviation Authority to obtain International Air Safety Association and United States Transportation Security Administration Category 1 clearance to operate direct flights to and from the United States.

4) Railways

Expedite putting in place the independent multi-sector regulatory body, in particular, for the railway sector;

Convert the residual Kenya Railways Corporation into an asset holding company which would also monitor and evaluate the performance of the concession.

6. Licensing and Regulatory Governance

Approximately 20 percent of managers interviewed place licenses among the top three constraints, and more firms complain about them than in all comparator countries. Reforms

Since 2005, Foreign Investment Advisory Service/IFC-World Bank, with support from development partners, have provided technical assistance to GoK (the Business Regulatory Reform Unit at Treasury and other parts of government) on licensing and regulatory reforms.

Follow up with the implementation of the licensing reforms.

Reduce the overall burden of licenses imposed on businesses, including a reduction in time and costs of obtaining a license to undertake business operations.

Continue establishment of an electronic register of licenses.

147

notwithstanding, Kenya does not perform as well as comparator countries in such areas as starting a new business, renewing licenses, and the costs of licenses.

The Kenyan government has recognized the importance of the licensing burden, and a number of reforms directed at reducing the number of licenses were approved in 2006 and 2007. The reform program has identified, for the first time, 1,325 active business licenses and eliminated 315, simplified 379, and cut both the time and cost of getting building permits. Notably, 23 out of a priority list of 26 problematic licenses identified by businesses have been eliminated or simplified. The still-ongoing program will eventually eliminate or simplify at least 900 more of the country’s 1,300 licenses.

In the next stages, the regulatory reform and capacity-building project will assist GoK in preparing and implementing a regulatory reform strategy. The strategy and its supporting implementation projects will continue to support the licensing reforms (including setting up an electronic registry of all valid licenses), streamline inspections procedures, introduce a system for vetting new licenses, address regulatory reforms at the local government level, and build the capacity of stakeholders to

Adopt a regulatory reform strategy to serve as a framework for licensing and other regulatory reforms, and to ensure their sustainability.

Reduce the burden imposed on businesses by on-site inspections.

Tackle licensing and regulatory reforms at the local government level.

Introduce a system for vetting proposed regulations to ensure that they do not place an undue burden on businesses.

Reduce the cost of trade documents.

Reduce minimum capital requirement to register a company.

Reduce the costs to start a business.

Reduce the time taken to start a business.

Reduce time for processing a VAT refund.

Reduce the number of payments for social security contributions and for VAT payments.

Establish online filing, as is already done in South Africa and Mauritius.

Harmonize the different tax identification numbers (PIN, VAT, etc.) into one universal number

Identify clear responsibilities

148

ensure the sustainability of the licensing reforms.

to continue licensing reforms

Improve information and transparency on regulatory reforms and outcomes

Reduce time for VAT refunds by allowing firms to use it as credit toward next payment

Reduce number of licenses by local authority and clarify the legal status of the ‘circular’

149

References

Adda, Jerome, and Russell Cooper. 2003. Dynamic Economics. Quantitative Methods and Applications. Cambridge, Massachusetts, and London, England: The MIT Press..

Bigsten, Arne, and Måns Söderbom 2006. “What Have We Learned from a Decade of Manufacturing Enterprise Surveys in Africa?” World Bank Research Observer 21 (2): 241–65.

Bigsten, Arne, Paul Collier, Stefan Dercon, Marcel Fafchamps, Bernard Gauthier, Jan Willem Gunning, Abena Oduro, Remco Oostendorp, Catherine Pattillo, Måns Söderbom, Francis Teal, and Albert Zeufack. 2004. “Do African Manufacturing Firm Learn from Exporting?” Journal of Development Studies 40 (3): 115–41.

Campos, J. Edgardo, and Sanjay Pradhan. 2007. “The Many Faces of Corruption: Tracking Vulnerabilities at the Sector Level.” The World Bank, April 2007.

Escribano, Alvaro, and J. Luis Guasch. 2005. “Assessing the Impact of the Investment Climate on Productivity Using Firm-Level Data: Methodology and the Cases of Guatemala, Honduras, and Nicaragua,” World Bank Policy Research Working Paper 3621. Washington, DC: The World Bank.

The Financial Sector Deepening (FSD) Kenya. (www.fsdkenya.org/finaccess)

Government of Kenya. Various Years. “Economic Survey.” Central Bureau of Statistics.

Government of Kenya. Various Years. “Statistical Abstract.” Central Bureau of Statistics.

Hausmann, Ricardo, Jason Hwang, and Dani Rodrik. 2005. “What You Export Matters.” Working Paper 11905. National Bureau of Economic Research. December 2005.

International Monetary Fund. “International Financial Statistics.” Online database.

International Monetary Fund. 2007. “Kenya—Staff Report for the 2006 Article IV Consultation, Second Review Under the Poverty Reduction and Growth Facility Arrangement, and Request for Extension and Rephasing of the Arrangement, Reduction in Access and Waiver of Performance Criteria.” IMF, African Department. March 13, 2007.

Klinger, Bailey, and Daniel Lederman. 2004. “Discovery and Development: An Empirical Exploration of ‘New’ Products.” World Bank Policy Research Working Paper 3450. November 2004.

150

http://www.fsdkenya.org/finaccess

MDRA. 2006. “A Report on Corruption in the Trucking Operation in India.” MDRA, September 2006.

Ndung’u, Njuguna. Survey on bank charges and lending rates in Kenya.

Söderbom, Måns, and Francis Teal. 2000. “Skills, Investment and Exports from Manufacturing Firms in Africa.” Journal of Development Studies 37 (2): 13–43.

Söderbom, Måns, and Francis Teal. 2003. “Are Manufacturing Exports the Key to Economic Success in Africa?” Journal of African Economies 12 (1): 1–29.

Speaking notes of Prof. Njuguna Ndung’u, governor of the Central Bank of Kenya, at the launch of the survey on bank charges and lending rates, Nairobi, August 28, 2007.

The World Bank. 2008. Africa Development Indicators 2007. Washington, DC: World Bank.

World Bank and United Nations Conference on Trade and Development (UNCTAD). “World Integrated Trade Solution (WITS).” COMTRADE database, maintained by the United Nations Statistical Division. Various Years

World Bank Group. Various Years. Doing Business. Washington, DC: World Bank.

World Bank, Various Years. “World Development Indicators Database.”

151

Technical Appendix

APPENDIX 1: ACCESS TO FINANCE

Access IndicatorsIn this section, we perform multivariate analysis of access indicators to determine the extent to which differences observed from simple tabulations discussed above remain in a regression framework. Many firm characteristics are correlated, and thus one characteristic could proxy for the effect of another characteristic in these simple tabulations. For example, foreign-owned firms tend to be larger, older, and in particular sectors, while younger firms tend to be smaller and tend to be run by indigenous entrepreneurs. Manufacturing firms are larger than retail firms and female owners are more likely to be African.

Appendix Table 1 reports the results of probit regressions with several dependent variables used to capture a variety of aspects of access to external finance. These include whether the firm has any credit products, loan applications and rejections, and perceptions of finance as an obstacle. The sample used to estimate these regressions includes all small, medium, and large formal enterprises. The coefficients reported are marginal probits evaluated at the mean of the independent variables.

Many of the results that we report in the bivariate analysis are no longer significant. The size effects are robust for the perceptions and access regressions, but not for loan application or rejection. We find that small firms are less likely than larger firms to have an overdraft or loan. Loan rejection rates, as reported in the text, are not significantly different across the size categories.

Firms with some foreign ownership are significantly less likely to use overdraft facilities. While we estimate that these firms are also less likely to have external financing, the point estimate is not significant. As reported in the main chapter, firms with African owners are more likely to report higher access obstacles. Our regressions, however, do not provide reliable evidence that there is any racial or ethnic discrimination in the credit markets: controlling for other firm characteristics, firms with African owners are as likely as other firms to have external financing. We do observe, however, a higher application rate among African-owned firms (rejection rates are no different, ceteris paribus).

Firms that own land are significantly more likely to access bank debt and less likely to have loan applications rejected. This highlights the role of collateral in the lending decision. Limited liability status does not appear to have an independent effect on access to credit, other factors remaining the same.

Two additional variables are included in regression analysis—whether the firm has an external auditor and the age of the firm. The external audit makes the firm’s financial statements more reliable and thus reduces the information asymmetry between the firm and financial institutions and should improve firm’s access to finance. We find that firms with an external auditor are not significantly more likely to apply for and obtain external

152

finance. The age of a firm captures a variety of plausible channels, leading to more bank debt for older firms; however, while the signs of the coefficients are in the direction we would expect, the precision of the estimates do not provide confidence of a positive effect of age.

Finally, we find evidence consistent with the bivariate results reported in the chapter that manufacturing firms are more likely to use external finance than the retail or other services sectors. Manufacturing firms are also more likely to apply for loans, but since rejection rates are uniform across all sectors, this is probably an important source of the difference in access.

Appendix Table 1 Correlates of Access to Finance(1) (2) (3) (4) (5) (6) (7)

Dependent Variable

Finance Major/Severe Constraint

Has Overdr

aft

Has Overdraft/L

oan

Applied for

Loan

Loan Rejecte

d

Any Workin

g Capital Bank

Financed

Any New

Investment

Bank Finance

dSome foreign ownership

0.081 -0.167 -0.130 -0.091 -0.131 -0.056 0.066

(0.069) (0.056)**

(0.073)+ (0.056)

(0.043)**

(0.055)

(0.064)

Has African owners

0.139 -0.042 0.047 0.169 0.031 0.043 -0.115

(0.047)** (0.054)

(0.058) (0.041)**

(0.058)

(0.046)

(0.048)*

Limited liability

-0.109 0.078 0.078 0.034 0.042 0.061 0.041

(0.052)* (0.061)

(0.058) (0.049)

(0.071)

(0.052)

(0.054)

Owns land -0.005 0.077 0.189 0.039 -0.129 0.162 0.074(0.050) (0.048

)(0.047)** (0.045

)(0.060

)*(0.046

)**(0.047)

Uses external auditor

-0.094 0.116 0.077 0.026 -0.122 0.041 0.085

(0.057) (0.061)+

(0.060) (0.054)

(0.076)

(0.055)

(0.057)

Firm exports

-0.055 0.050 0.086 0.068 -0.019 0.083 0.061

(0.060) (0.061)

(0.064) (0.055)

(0.071)

(0.055)

(0.058)

0–4 years old

0.121 0.086 0.114 -0.008 -0.045 0.001 -0.055

(0.071)+ (0.081)

(0.074) (0.067)

(0.068)

(0.069)

(0.070)

5–9 years old

0.182 0.132 0.214 0.092 -0.048 0.097 -0.016

(0.062)** (0.068)+

(0.061)** (0.059)

(0.067)

(0.061)

(0.062)

10–19 0.107 -0.040 0.080 0.047 -0.098 0.088 -0.002

153

years old(0.072) (0.068

)(0.072) (0.065

)(0.057

)+(0.068

)(0.067)

20+ years old

0.040 -0.070 -0.073 0.036 -0.027 -0.027 0.144

(0.079) (0.074)

(0.085) (0.074)

(0.083)

(0.069)

(0.069)*

20–99 employees

-0.101 0.358 0.267 0.039 0.003 0.095 -0.215

(0.051)* (0.056)**

(0.051)** (0.051)

(0.065)

(0.054)+

(0.055)**

100+ employees

-0.166 0.484 0.377 0.108 0.034 0.131 -0.203

(0.066)* (0.065)**

(0.054)** (0.070)

(0.102)

(0.073)+

(0.075)**

Manufacturing

0.203 0.409 0.358 0.226 0.073 0.080 0.174

(0.059)** (0.063)**

(0.065)** (0.055)**

(0.069)

(0.061)

(0.061)**

Retail 0.033 0.229 0.111 0.089 0.130 0.062 0.039(0.064) (0.093

)*(0.069) (0.070

)(0.127

)(0.067

)(0.062)

Observations

656 656 656 656 208 656 656

154

APPENDIX 2: LABOR MARKETS AND HUMAN CAPITAL

We present the results of multivariate regressions to qualify the bivariate associations that we discuss in the chapter. Regressions are presented at the firm and individual level for training and wages. For the training regressions, we estimate the likelihood that a firm provides on-the-job training to its workers. Typical controls include export status, ownership, firm size, and vintage and sector. Using worker data, we estimate the likelihood of receiving either firm- or self-financed training using worker attributes such as gender, schooling, experience, and the matched firm characteristics outlined previously.

We use the firm- and individual-level wage regressions to provide insights into the wage-setting mechanisms, as well as the returns to various worker attributes.

Appendix Table 2 Training Determinants: Firm Level[[please provide callouts for all tables]](1) (2) (3) (4) (5) (6)

20–99 employees 0.186 0.189 0.188 0.136 0.143 0.146(0.069

)**(0.069

)**(0.072

)**(0.072

)(0.073

)*(0.076

)100+ employees 0.355 0.350 0.359 0.331 0.330 0.347

(0.077)**

(0.077)**

(0.081)**

(0.081)**

(0.081)**

(0.084)**

Exports 0.082 0.060 0.052 0.068 0.046 0.039(0.061

)(0.061

)(0.062

)(0.065

)(0.066

)(0.067

)Foreign owned -0.072 -0.070 -0.074 -0.085 -0.087 -0.086

(0.069)

(0.069)

(0.069)

(0.069)

(0.070)

(0.070)

Firm age 0.003 0.002 0.001 0.003 0.002 0.002(0.005

)(0.005

)(0.005

)(0.005

)(0.005

)(0.005

)Firm age squared -0.000 -0.000 -0.000 -0.000 -0.000 -0.000

(0.000)

(0.000)

(0.000)

(0.000)

(0.000)

(0.000)

Typical worker has more than 6 years of school

-0.039 -0.073 -0.097 -0.009 -0.039 -0.056

(0.089)

(0.092)

(0.098)

(0.091)

(0.093)

(0.100)

Percent unionized 0.001 0.001 0.001 0.001 0.001 0.001(0.001

)(0.001

)(0.001

)(0.001

)(0.001

)(0.001

) Percent part-timeseasonal 0.002 0.002 0.002 0.002 0.002 0.002

(0.001)

(0.001)

(0.001)

(0.001)

(0.001)

(0.001)

Engaged in HIV prevention 0.129 0.127 0.119 0.116(0.055

)*(0.056

)*(0.057

)*(0.058

)*Capacity utilization -0.002 -0.003

(0.002)

(0.002)

Uses external audit 0.044 0.021(0.101

)(0.105

)

Sector controls No No No Yes Yes Yes

155

Observations 385 385 385 385 385 385F-test firm size matters 19.41 18.82 17.90 16.28 16.08 16.21prob>F 0.00 0.00 0.00 0.00 0.00 0.00Note: Robust standard errors in parentheses; * significant at 5 percent; ** significant at 1 percent. Reported coefficients are dprobit estimates from a probit regression with the likelihood of providing training as the dependent variable. Last two rows report the results of a formal test that size is not associated with training provision. As the p-values of this test indicate, we can reject this hypothesis as conventional levels of statistical significance.

156

Appendix Table 3 Training Determinants: Individual Level Dependent Variable: Worker Received any

Kind of TrainingDependent Variable: Worker Received Firm-

Financed Training(1) (2) (3) (4) (5) (6) (7) (8)

Years of schooling

0.047 0.040 0.042 0.056 0.007 0.004 0.001 0.035

(0.006)**

(0.006)**

(0.007)**

(0.014)**

(0.004)+ (0.004) (0.005) (0.013)**

Worker experience

-0.014 -0.017 0.006 0.024 -0.010 -0.013 -0.001 -0.057

(0.008)+ (0.009)+ (0.010) (0.016) (0.006) (0.007)+ (0.008) (0.019)**

Experience squared

0.000 0.001 -0.000 -0.001 0.000 0.000 0.000 0.002

(0.000) (0.000) (0.000) (0.001) (0.000)+ (0.000)+ (0.000) (0.001)**

Worker is female

-0.004 -0.038 0.017 0.009 -0.025 -0.049 0.003 -0.096

(0.035) (0.036) (0.048) (0.065) (0.028) (0.028)+ (0.038) (0.077)Union member

-0.035 -0.020 0.054 -0.161 -0.075 -0.070 -0.074 0.010

(0.033) (0.034) (0.043) (0.074)* (0.027)**

(0.028)* (0.033)* (0.082)

Worker full-time

0.125 0.083 0.128 0.361 0.147 0.133 0.163 0.270

(0.051)* (0.055) (0.064)* (0.086)**

(0.031)**

(0.034)**

(0.036)**

(0.111)*

Worker is single

0.037 0.008 0.157 -0.035 -0.076 -0.142

(0.044) (0.050) (0.069)* (0.034) (0.036)* (0.082)+Professional worker

0.037 -0.008 0.069 0.011

(0.081) (0.098) (0.073) (0.078)Skilled production worker

-0.023 0.012 -0.270 0.000 -0.007 -0.134

(0.063) (0.082) (0.125)* (0.051) (0.063) (0.141)Unskilled production workers

-0.152 -0.082 -0.513 -0.057 -0.042 -0.210

(0.064)* (0.087) (0.101)**

(0.053) (0.065) (0.154)

Nonproduction worker

0.041 -0.023 -0.389 0.065 -0.046 -0.255

(0.071) (0.089) (0.114)**

(0.062) (0.063) (0.136)+

20–99 employees

0.157 -0.107

(0.091)+ (0.071)100+ employees

0.343 0.139

(0.086)**

(0.074)+

Firm exporter

0.156 0.087

(0.040)**

(0.033)**

Foreign owned

-0.204 -0.230

(0.040)**

(0.027)**

Firm age -0.014 -0.007(0.004)*

*(0.003)*

Firm age squared

0.000 0.000

(0.000)**

(0.000)*

157

Typical worker more than 6 years of school

0.138 0.075

(0.059)* (0.044)+ Percent unionized

-0.000 0.002

(0.001) (0.001)**

Observations

969 969 781 553 969 969 781 351

Note: Robust standard errors in parentheses; * significant at 5 percent; ** significant at 1 percent. Reported coefficients are dprobit estimates from a probit regression with the likelihood of receiving training as the dependent variable. Specifications (4) and (8) include firm fixed effects.

158

Appendix Table 4 Determinants of Average Wages: Firm Level Dependent Variable: Log

Average Wage for Production Workers

Dependent Variable: Log Average Wage for

Nonproduction Workers(1) (2) (3) (4) (5) (6)

20–99 employees 0.078 0.075 0.066 0.034 0.009 -0.010(0.068) (0.071) (0.071) (0.102) (0.108) (0.108)

100+ employees 0.213 0.182 0.155 0.202 0.129 0.070(0.087)* (0.095)

+(0.096) (0.112)

+(0.122) (0.121)

Exports -0.022 -0.015 -0.020 0.136 0.123 0.105(0.059) (0.058) (0.057) (0.079)

+(0.078) (0.078)

Foreign owned 0.091 0.100 0.097 0.201 0.216 0.220(0.067) (0.066) (0.066) (0.096)* (0.093)* (0.093)*

Firm age 0.008 0.009 0.008 0.016 0.017 0.015(0.005)

+(0.005)

+(0.005) (0.007)* (0.007)* (0.007)

+Firm age squared -0.000 -0.000 -0.000 -0.000 -0.000 -0.000

(0.000) (0.000) (0.000) (0.000)+

(0.000)+

(0.000)

Typical worker has more than 6 years of school

0.005 0.072 0.092 -0.074 -0.100 -0.078

(0.084) (0.096) (0.099) (0.103) (0.114) (0.118)Skill ratio of production workers 0.204 0.148 0.131 -0.040 -0.033 -0.043

(0.101)* (0.102) (0.101) (0.119) (0.126) (0.125) Percent unionized -0.000 -0.000 -0.000 0.001 0.001 0.001

(0.001) (0.001) (0.001) (0.001) (0.001) (0.001)Provides training 0.001 0.001 0.001 0.001

(0.001)+

(0.001)+

(0.001)+

(0.001)+

Capacity utilization 0.004 0.004 0.001 0.001(0.002)*

*(0.002)*

*(0.002) (0.002)

Access to external credit -0.036 -0.052 0.065 0.034(0.056) (0.056) (0.071) (0.071)

Access to trade credit -0.035 -0.038 0.139 0.136(0.111) (0.111) (0.173) (0.170)

External audit -0.143 -0.143 0.077 0.081(0.105) (0.105) (0.151) (0.151)

Gross profit 0.000 0.000(0.000) (0.000)

Constant 4.363 4.150 4.172 4.969 4.654 4.654(0.134)*

*(0.214)*

*(0.212)*

*(0.162)*

*(0.301)*

*(0.297)*

*Observations 385 385 380 379 379 374R-squared 0.17 0.20 0.21 0.19 0.20 0.21F-test firm size matters 3.16 1.95 1.35 2.39 1.04 0.44prob>F 0.04 0.14 0.26 0.09 0.36 0.65Note: Robust standard errors in parentheses; * significant at 5 percent; ** significant at 1 percent. Reported coefficients are an ordinary least square regression with log of average worker earnings as the dependent variable. Last two rows report the results of a formal test that size is not associated with earnings. As the p-values of this test indicate, we can reject this hypothesis as conventional levels of statistical significance only for the basic specifications (1) and (4).

159

Appendix Table 5 Determinants of Worker Earnings Dependent Variable: Log Monthly Earnings

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

Years of schooling

0.084 0.084 0.084 0.085 0.092 0.081 0.074 0.071 0.072 0.066 0.066 0.074

(0.007)**

(0.008)**

(0.007)**

(0.007)**

(0.008)**

(0.007)**

(0.008)**

(0.008)**

(0.008)**

(0.009)**

(0.009)**

(0.008)**

Experience

0.057 0.050 0.057 0.054 0.036 0.024 0.030 0.031 0.029 0.044 0.044 0.043

(0.014)**

(0.011)**

(0.014)**

(0.016)**

(0.015)*

(0.015)

(0.017)+

(0.016)+

(0.017)+

(0.018)*

(0.018)*

(0.013)**

Experience squared

-0.001

-0.001

-0.001

-0.001

-0.000

-0.000

-0.001

-0.001

-0.001

-0.001

-0.001

-0.001

(0.001)+

(0.000)*

(0.001)+

(0.001)+

(0.000)

(0.000)

(0.001)

(0.000)

(0.000)

(0.001)+

(0.001)+

(0.000)*

Female -0.078

-0.042

-0.078

-0.076

-0.052

-0.039

-0.010

-0.003

0.000 -0.020

-0.019

-0.022

(0.048)

(0.038)

(0.048)

(0.048)

(0.043)

(0.041)

(0.048)

(0.049)

(0.049)

(0.058)

(0.057)

(0.035)

Single 0.107 0.068 0.105 0.106 0.103 0.130 0.107 0.092 0.101 0.119 0.119 0.078(0.08

6)(0.05

5)(0.08

7)(0.08

7)(0.07

8)(0.076)+

(0.079)

(0.082)

(0.082)

(0.092)

(0.092)

(0.054)

Union member

-0.338

-0.259

-0.339

-0.339

-0.243

-0.210

-0.148

-0.171

-0.153

-0.266

-0.266

-0.226

(0.058)**

(0.056)**

(0.058)**

(0.058)**

(0.053)**

(0.052)**

(0.060)*

(0.062)**

(0.063)*

(0.070)**

(0.069)**

(0.053)**

Worker full-time

0.287 0.125 0.281 0.282 0.095 0.073 0.075 0.062 0.056 0.276 0.277 0.068

(0.115)*

(0.090)

(0.118)*

(0.118)*

(0.088)

(0.086)

(0.088)

(0.087)

(0.084)

(0.116)*

(0.116)*

(0.091)

Any training

-0.005

-0.005

-0.004

-0.002

-0.007

-0.012

-0.015

-0.005

-0.005

-0.019

(0.004)

(0.004)

(0.005)

(0.004)

(0.004)+

(0.004)**

(0.004)**

(0.005)

(0.005)

(0.005)**

Tenure, years

0.005 0.006 0.010 0.016 0.017 0.023 0.019 0.019 0.001

(0.010)

(0.010)

(0.010)

(0.011)

(0.011)

(0.012)+

(0.012)

(0.012)

(0.009)

Obtained current job through network

-0.419

-0.436

-0.406

-0.427

-0.500

-0.501

-0.354

(0.059)**

(0.069)**

(0.070)**

(0.075)**

(0.080)**

(0.081)**

(0.060)**

Log weekly hours

1.441 1.397 1.390 1.503 1.553 0.126

(0.206)**

(0.199)**

(0.189)**

(0.218)**

(0.219)**

(0.154)

20–99 employees

-0.169

-0.169

-0.175

-0.102

-0.101

(0.104)

(0.113)

(0.113)

(0.116)

(0.116)

100+ employees

0.089 0.087 0.093 0.197 0.197

(0.106)

(0.114)

(0.118)

(0.124)

(0.124)

Exports 0.031 -0.031

-0.032

(0.062)

(0.068)

(0.067)

Foreign - 0.140 0.140

160

owned 0.013(0.07

3)(0.084)+

(0.084)+

Firm age

-0.013

-0.009

-0.009

(0.006)*

(0.006)

(0.006)

Firm age squared

0.000 0.000 0.000

(0.000)+

(0.000)

(0.000)

Gross profit

-0.000(0.00

0)Constant

7.747 8.709 7.751 7.738 2.338 2.996 3.136 2.731 2.779 8.518 8.517 8.646

(0.152)**

(0.270)**

(0.153)**

(0.153)**

(0.803)**

(0.773)**

(0.742)**

(0.822)**

(0.805)**

(0.284)**

(0.283)**

(0.610)**

Observations

965 965 965 965 965 965 786 786 786 769 769 965

R-squared

0.20 0.63 0.20 0.20 0.31 0.36 0.37 0.38 0.39 0.28 0.28 0.66

F-test firm size matters

12.62 10.22 11.13 9.29 9.72

prob>F 0.00 0.00 0.00 0.00 0.00Note: Robust standard errors in parentheses; * significant at 5 percent; ** significant at 1 percent. Reported coefficients are an ordinary least square regression with log of average worker earnings as the dependent variable. Last two rows report the results of a formal test that size is not associated with earnings (columns 7–11). As the p-values of this test indicate, we can reject this hypothesis as conventional levels of statistical significance. Specification (12) reports the results of an ordinary least square estimation controlling for firm fixed effects.

161

APPENDIX 3: ENTERPRISE SURVEY IN KENYA: SAMPLE SURVEY DESIGN

Productivity and Investment Climate Survey

Survey Coverage

The World Bank Enterprise Survey in Kenya targeted establishments located in Nairobi, Mombasa, Nakuru, and Kisumu in the following industries (according to International Standard Industrial Classification, revision 3.1): all manufacturing sectors (group D); construction (group F); retail and wholesale services (subgroups 52 and 51 of group G); hotels and restaurants (group H); transport, storage, and communications (group I); and computer and related activities (subgroup 72 of group K). For establishments with five or more full-time permanent paid employees, this universe was stratified according to the following categories of industry:

1. Manufacturing: Food and Beverages (group D, subgroup 15);2. Manufacturing: Garment (group D, subgroup 18);3. Manufacturing: Other Manufacturing (group D, excluding subgroups 15 and 18);4. Retail Trade: (group G, subgroup 52);5. Rest of the universe, including

Construction (group F); Wholesale trade (group G, subgroup 51); Hotels, bars and restaurants (group H); Transportation, storage, and communications (group I); and Computer-related activities (group K, subgroup 72).

The survey also sampled a selection of microestablishments (establishments with fewer than five full-time permanent paid employees) from the targeted universe, without stratification by industry.

Sampling Methodology

Establishments with Five or More Full-Time Paid Permanent Employees

A satisfactory list of establishments was sourced from the Kenya National Bureau of Statistics, the Kenya Association of Manufacturers, the Kenya National Chamber of Commerce, the Kenya Private Sector Alliance, and the Federation of Kenya Employers. These lists were merged together into a master list that was validated, updated where possible, and then used to establish the initial population size for each stratum. The final population size in all strata and locations was 6,562, with the vast majority of establishments operating in the rest of the universe and manufacturing strata (see Appendix Table 6).

In Kenya, the survey includes panel data collected from establishments surveyed in the 2003 PICS in Kenya. That survey included establishments in all three manufacturing strata distributed across the entire country. In order to collect the largest possible set of

162

panel data, an attempt was made to contact and survey every establishment in the panel, provided it was located in one of the four cities covered by this survey and operated in the universe under study.

The remainder of the sample (including the entire rest of universe and retail sample in each city) was selected at random from the master list by a computer program.

MicroestablishmentsIn this survey, the microestablishment stratum covers all establishments of the targeted categories of economic activity with fewer than five employees. For many reasons, including the small size of establishments, their expected high rate of turnovers, the high level of “informality” of establishments in many activities and, consequently, the difficulty in obtaining trustworthy information from official sources, EEC Canada selected an aerial sampling approach to estimate the population of establishments and select the sample in this stratum for all regions of the survey.

The following procedure was followed for sampling microestablishments:

Step 1: Districts and specific zones of each district with a high concentration of microestablishments were identified;

Step 2: A count of all microestablishments in these specific zones was conducted;

Step 3: The count by zone was converted into one list of sequential numbers for the whole survey region and a virtual list was created with establishment numbers;

Step 4: A computer program performed a random selection of establishment numbers from that virtual list; and

Step 5: Based on the ratio between the number selected in each specific zone and the total population in that zone, a skip rule was created and applied for, selecting the corresponding establishments in each zone.

Enumerators applied the skip rule defined for that zone, as well as how to select replacements in the event of refusal or other cause of nonparticipation.

Population and Sample SizeAppendix Table 6 Population Size by Stratum and Sampling Region

Nairobi Mombasa Nakuru Kisumu TotalManufacturing 751 108 49 37 945Food and beverages 189 27 16 16 248Garments 79 23 3 5 110Other manufacturing 483 58 30 16 587Retail 1,193 627 307 364 2,491Rest of the universe 1,863 534 308 421 3,126Total —formal 3,807 1,269 664 822 6,562Micro 153,423 64,138 1,317 1,328 220,206Total 157,230 65,407 1,981 2,150 226,768

163

Appendix Table 7 Target Sample Size by Stratum and Sampling RegionNairobi Mombasa Nakuru Kisumu Total

Manufacturing 242 52 35 34 363Food and beverages 78 15 10 18 121Garments 78 21 13 9 121Other manufacturing 86 16 12 7 121Retail 54 18 25 24 121Rest of the universe 60 18 22 21 121Total —formal 356 88 82 79 605Micro (four employees or fewer) 61 20 20 20 121Total 417 108 102 99 726

Appendix Table 8 Final Sample Size by Stratum and Sampling RegionNairobi Mombasa Nakuru Kisumu Total

Manufacturing 274 51 34 37 396Food and beverages 73 13 8 16 110Garments 62 13 2 5 82Other manufacturing 139 25 24 16 204Retail 59 16 26 25 126Rest of the universe 69 20 22 24 135Total—formal 402 87 82 86 657Micro (four employees or fewer) 64 20 20 20 124Total 466 107 102 106 781

Appendix Table 9 Effective Sample—Panel by Stratum and Sampling RegionNairobi Mombasa Nakuru Kisumu Total

Manufacturing 112 25 20 5 162Food and beverages 23 4 7 3 37Garments 11 3 14Other manufacturing 78 18 13 2 111Retail Rest of the universeTotal—formal 112 25 20 5 162Micro (four employees or fewer)Total 112 25 20 5 162

164