66
TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja Kacker and Ying Li PRMED Input for the CEM 2005 41668 Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized Public Disclosure Authorized

TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

  • Upload
    others

  • View
    4

  • Download
    0

Embed Size (px)

Citation preview

Page 1: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

TANZANIA: GROWTH , EXPORTS AND EMPLOYMENTIN THE MANUFACTURING SECTOR

Vandana Chandra, Pooja Kacker and Ying Li

PRMED

Input for the CEM 2005

41668

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Page 2: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

2

TANZANIA’S MANUFACTURING SECTOR:

GROWTH, EXPORTS AND EMPLOYMENT

INTRODUCTION

Remarkably, over the past forty years, the share of Tanzania’s manufacturing sector inits economy has remained virtually unchanged at around 8 percent of GDP (Table 1). Formost of the first 30 years, government’s experiment with state control and planningcapped the share of the private sector in the economy at 12 percent. However, since thelate 1980s, even after slow return to a market-based system, manufacturing’s role in theeconomy is unchanged: the vision of structural change that could modernize this ruraleconomy through diversification and industrialization remains distant. From a relativelysmall base, growth in manufacturing has picked up from about 2 percent per annumduring 1990-1998 to 6 percent in 1999-03 (Table 2). In the last two years (2002-03), itrose to 8 percent p.a.. However, as growth in the other sectors has also accelerated,manufacturing’s relative contribution to GDP, economic diversification and modern jobcreation is small. The overwhelming dominance of primary production, especially in therural sector continue to challenge policy makers to think harder about what more needs tobe done to unleash Tanzania’s manufacturing potential to move closer to the vision of amodern economy.

A vibrant manufacturing sector is critical for the Tanzanian economy for several reasons.First, as international experience has shown, manufacturing has been an important sourceof growth in several small countries, even in sub-Saharan Africa (see Boz ---- onLesotho, Mauritius etc.), leading the way to a modern economy envisioned in the latestPRSP. Typically, a growing manufacturing sector triggers the development of ancillaryactivities and better-paying jobs vital for a modern economy. Second, a largemanufacturing sector is necessary for export diversification to reduce Tanzania’sdependence on external shocks caused by weather cycles and terms of trade fluctuations.The latter routinely affect its export earnings tied to a basket comprised primarily ofprimary commodities such as gold, and traditional agricultural products (coffee, cashews,tea etc..). Besides adding uncertainty to foreign exchange earnings, the adverse effects ofterms of trade shocks also spillover to household incomes, especially of the rural poor.Weather shocks, such as droughts, frequently impact agricultural production thataccounts for over 50 percent of GDP and the majority of livelihoods in thispredominantly rural economy. As manufacturing is considerably less vulnerable to suchexternal shocks, it is valuable as a potential source of growth, exports and jobs that cangenerate higher and more sustainable incomes. Why Tanzania’s manufacturing sector hasnot yet developed into such a source of growth and what can be done to accelerate thisprocess is the subject of this paper.

Page 3: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

3

‘Tanzania at the Turn of the Century’ (CEM, 2001) examined the country’s economicperformance during 1960s – 1999 and concluded that the Government of Tanzania’s(GoT’s) intensification of macroeconomic policy reforms between 1995-1999,established a stable macroeconomic environment, especially low inflation, conducive forinvestment and growth. As GoT, also implemented structural reforms in parallel, manygross economic distortions that earlier constrained growth were eliminated effectively.The strengthening of private investment incentives was expected to boost privateinvestment and establish a strong manufacturing sector. Since the late 1990s, FDI flowsincreased and were related largely with mining and privatization. The share or growth ofprivate investment or FDI in manufacturing is not known. Overall, the share of privateinvestment (and indeed total investment in GDP) has plummeted from over 20 percent inthe mid-1990s to 12 percent in recent years. While it is true that manufacturing’sperformance improved with growth rising to 8 percent in 2002-03, it is more difficult todiscern if this accelerated performance rests on a sound foundation for growth anddefines a trend. It is unclear whether the sector’s performance is driven by steadyinvestment and employment creation, or whether much of the high growth is the outcomeof once-off efficiency gains associated with the reforms in the late 1990s. It is too early totell.

Recent reports (CEM 2001 and RPED 2004) compared Tanzania’s manufacturingperformance with its immediate neighbors, Kenya and Uganda, but this does not shedlight on country-specific factors. Several key issues remain unanswered:• Besides low inflation and macroeconomic stability, what are the determinants of

manufacturing growth in Tanzania? Recently the GoT implemented reforms in thefinancial and telecommunications sectors. How have these affected the manufacturingsector?

• What are the main obstacles to growth in manufacturing since 1999? What are themain policy levers that GoT can use to sustain growth in the sector? Can they beprioritized?

• If a small open economy like Tanzania is able to produce competitively, it should beable to grow rapidly by exploiting exports as a key source of growth. What is therelationship between growth and exports in the manufacturing sector?

• The PRSP noted that the benefits of macroeconomic reforms have not trickled downsufficiently to the poor. Expansion of the Tanzania’s manufacturing sector has thepotential to directly create jobs that fetch wages/incomes higher than those generatedin the rural or informal sectors. Is manufacturing a source of vibrant employmentgrowth in Tanzania?

Meaningful answers to most of these questions require a microeconomic framework thatdelves below the macro to assess the impact of reform on firms. This paper adds to theexisting knowledge-base on Tanzania by grappling with these questions using RPED firmsurveys for the 1999 - 2003 period. By using 1993 as a reference point, it assesses whereTanzania’s manufacturing sector stands more than 10 years after it initiated seriouseconomic reforms.

Page 4: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

4

The organization of this paper is as follows. Section 2 provides a brief description ofmanufacturing activity in Tanzania in the 1990s, but especially since 1999 when thebroad macro reforms were in place. Sections 3, 4 and 5 address the issues listed in theprevious section in the context of investment and growth, exports and employment in thesector. The starting point of each section is issues identified by the CEM of 2001. Thediscussion begins from the broad macro trends, especially since 1999. This is followed bymicro evidence. International comparisons are also provided. The performance of thesector is evaluated in terms of pertinent reforms, especially those since the late 1990s,and their impact. Section 6 concludes with policy implications and recommendations.

1. BACKGROUND

RECENT TRENDS AND DEVELOPMENTS IN THE MANUFACTURING SECTOR

At 8 percent, the share of manufacturing in Tanzania’s GDP is presently small. It has alsoshrunk in the past 10 years. In 1985, it was over 9 percent of GDP (Table 1, chart 1). Inthe past few decades, agriculture has remained the key driver of overall growth, andservices, especially construction, have also grown fast. Between 1964 and the late 1980s,when the government imposed a ceiling of 12 percent on the share of private investment,the latter was restricted to a narrow range of activities such as construction, transport,agricultural equipment and miscellaneous manufacturing.

Today, while the size of the manufacturing sector remains relatively small, its structure isfairly diversified (see Table ---). Production is concentrated in Dar es Salaam in threetypes of firms: agro-processing and food (beer, spirits and cigarettes), textiles and otherlight industry such as furniture, and heavy industry producing metals (aluminum, ironsheets), cement, paints and plastics. Industrial capacity utilization increased to 7 hours aday in 1998 from just 4 hours a day in 1995 (BOS). Recent growth has occurred mostlyin consumer goods such as food products and beer, edible oils and detergents that haveattracted new investment. Growth in heavy industry such as chemicals (fertilizers andpaints) and metals has declined.

What remains remarkably unchanged and a source of concern about the nature ofmanufacturing growth since 1999 is the disconnect between it and private investment.Several developments explain the persistence of this disconnect over the past threedecades. In the 1970s- late 1980s, it was explained by productivity losses anddisincentives for private investment associated with government’s socialist polices. Thespillover from these was blamed for the disconnect in the early 1990s. During the 1990s,its perpetuation was attributed to declining public investment ensuing with the reductionin fiscal deficits in the early to mid-1990s (CEM 2001). Public investment wasconsidered complementary to private investment. The reforms of the late 1990s wereexpected to promote a business friendly environment and re-establish the natural linkbetween investment and growth. This did not happen.

Page 5: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

5

Public investment increased from about 3 percent in 1999 to almost 8 percent of GDP in2003, but aggregate private investment did not pick up correspondingly. Instead, from ashare of 18 percent of GDP in 1992, it plummeted after 1996, settling at 11 percent ofGDP in 2000-03. In manufacturing, private investment declined even more steeply.From 35 percent of total private investment in 1992, it fell to about 25 percent in 2000-02(Table …).

During 1997- 2001, following the range of fiscal incentives for investors, mining was themain beneficiary of private investment growth of over 16 percent per annum (EIU, 2003).Recent reforms have attracted FDI flows in telecommunications as the industry wasprivatized, and into ports, airports, utilities and the financial sector (EIU 2004/05 page 26– 27).

These trends suggest that, at least in manufacturing, the recent acceleration in growth isrelated not with a corresponding rise in private investment, but is a combination of theimproved quality of private investment and increased utilization of existing capacity. It isconsistent with the channeling of private investment flows into mining and construction.At a micro level, an analysis of firm surveys for 2000-02 also show that the associationbetween private investment and growth in manufacturing remains elusive.

• If the data is to be believed, the sustainability of the recent growth withoutcorresponding investments in manufacturing is questionable, particularly in themedium to longer term. In the short term, reforms that further reduce regulatoryobstacles and corruption will continue to yield significant efficiency gains. Payoffsfrom the latter can probably support the prevailing growth rates in manufacturingover the next 2-4 years. However, without a significant and sustained increase inprivate investment, this trend is unlikely to be sustained for too long.

RECENT TRENDS IN MANUFACTURING EMPLOYMENT

In 2001, out of a labor force of approximately 18 million, 16 million were employed. Ofthese, Tanzania’s non-agricultural private sector across all sectors employed only 4.5percent. Manufacturing’s contribution to total employment was even smaller at 245,000or 1.5 percent of the employed. As a share of non-agricultural private employment(industry and services), manufacturing accounted for about a third. The pattern of labordemand reflected the technical skill-intensive nature of manufacturing. Of the totalemployed in all industries, manufacturing absorbed about 34 percent of all craft andrelated-skills workers, 32 percent of all machine and plant operators and 11 percent of allclerks. The manufacture of goods such as textiles and garments or furniture and non-metallic products does not occur in formal manufacturing alone. For example, only 19percent of the manufactures of wearing apparel, spinning, weaving and finishing, and 35percent of furniture making and the production of non-metallic products occurs in formalfirms that employ paid workers; the remainder is produced by the self-employed (ILFS2001).

Page 6: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

6

The CEM (2001) notes that in addition to higher growth, a wider sharing of the benefitsof growth was deemed equally vital for poverty reduction. This implied the creation ofgainful employment opportunities for the poor, given the enormous pressure created by ahigh rate of population growth (3 percent per annum) and a dependency ratio of 50percent .1 Manufacturing has the potential to contribute to better employment creation asit generates better paying jobs relative to agricultural earnings that typically fall belowthe poverty line. For example, in 2001, the median2 monthly manufacturing income wasTZ Sh. 40,000 compared to 8000 in agriculture, 30000 in mining, 36000 in constructionand 20000 in trade. The industrial average was TZ Sh. 30,000. Those who were self-employed in manufacturing earned TZ Sh. 21,429 per month. For policymakers interestedin sustained poverty reduction through job creation, promoting growth in manufacturinghas a clear attraction. Whether this has indeed transpired cannot be confirmed byaggregate data. Using firm survey analysis, we seek to answer this issue as well as whatconstrains employment creation in Tanzania’s manufacturing sector.

LONGER TERM VISION FOR GROWTH AND OBSTACLES SKILLS

In 2001, the longer term goal of a modern, export-led economy was based on severalfactors favorable for Tanzanian firms in manufacturing. Among these were easy access tothree sea ports that were operating with greater efficiency; expectations of improvedpower supply from increases in hydro and thermal sources; potential sources of mineralsand coal, sustained reduction in policy and administrative barriers, especially ininternational trade and a stable macroeconomic environment.

The imperatives for scaling up growth in the medium term rested on an increasedinvestment rate of 25 percent of GDP. This was to be achieved through better quality andlower-cost infrastructure services to be improved through private-public partnerships. Inaddition to

1) infrastructure, the other critical obstacles to investment were2) a weak institutional and legal framework regarding property rights, relatively

high business risks,3) low level of skills and educational attainment that hampered acquisition of

knowledge, use of superior technologies and FDI and hence, limited productivitygrowth,

4) the HIV/AIDS epidemic eroding the productive, prime-age cohort of 15 – 59 inthe labor force,

5) a weak and primitive system of financial intermediation, and6) a governance system that was not sufficiently transparent and accountable. (CEM

2001).

1 Overall, between 1990/91 – 2000/01, the labor force increased by 58 percent. The urban labor force grewby 14.3 percent compared to a decline in the rural labor force of 3 percent (ILFS).2 The comparable means were TZ Sh. 103,407 in manufacturing; 15234 in agriculture, 76277 in mining,49693 in construction, 31,301 in trade. The industrial mean was TZ Sh. 49,954. (ILFS 2001, Tables 9.4 and9.5)

Page 7: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

7

In 2004, with overall investment stagnant at 20 percent of GDP, GoT was engaged in arange of reforms in the area of tax and the financial sector, improvement of publicenterprise performance especially in energy, combating corruption, and strengthening thelegal and institutional environment for private sector growth (IMF, July 2004).

POST-1999 REFORMS RELEVANT FOR REDUCING THE OBSTACLES TO GROWTH INMANUFACTURING

Macroeconomic stability, exchange rate and trade reforms. The post-1995-96 reformshelped to establish macroeconomic stability, lower inflation to single digits (about 5percent in 2004), and led to a steadily depreciating and liberalized exchange rate fortrade-related transactions. The continued depreciation of the real exchange rateapproaching its equilibrium levels in 2003-04 provided favorable conditions formanufacturing export growth. Additional benefits were anticipated from further tradeliberalization through tariff cuts and the establishment of a customs union betweenTanzania, Kenya and Uganda (EAC). By 2009, the latter is likely to eliminate remaininginternal barriers to trade (IMF, 2004) and should boost manufactured exports further.

Structural reforms. Several structural reforms - civil service, privatization, tradeliberalization - reigned in fiscal discipline, reduced distortions, improved economicefficiency and speeded up manufacturing growth to 6 percent per annum during 1999-2003. Apparently, starting in 2005, the next critical area of reform is customsadministration which through reductions in the transactions costs of trading will bestowpotential productivity gains for the sector (IMF, 2004). While the scope for continuedefficiency gains is significant in the near future, without additional investment, growth inthe sector may not be sustainable. The challenge for policy makers is to explore if, inaddition to efficiency-improving reforms, they can do anything more to improveproductivity and attract investment in the sector.

Financial sector reform. Between 1967 – 1991, the state controlled the commercial banksin Tanzania. In recent years, private participation has steadily increased. In 2003, therewere 20 banks, including most international banks that operate in Africa. To reduce thecontinued hurdles to private credit for business investments, in 2001, GoT’s otherplanned reforms included: privatization of the Cooperative and Rural Development Bank,institution of commercial courts to legally enforce loan repayments and improve banklending to businesses. Although the effect of these reforms on investors in manufacturingis not known, they had several potentially beneficial effects for the sector. Overall, theratio of non-performing loans to capital fell from 22 percent in 2002 to 9 percent in 2003(IMF, 2004). The reduction in interest rate spreads from 18.4 percent in 1997 to 11.4percent in 2003 led to a doubling of bank credit to the private sector. The latter rose from3.5 percent of GDP in 1997 to 7.6 percent in 2003. In 2000, for many manufacturingfirms, at 14 percent the interest rate spread was a deterrent to investment (CEM 2001).

To promote economic development, GoT designed several non-bank financialinstitutions, merchant banks and the Tanzania Investment Bank designed. In 1998, theDar es Salaam stock exchange (DSE) was established to attract investors; since 2002,

Page 8: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

8

foreigners have also been permitted to invest in the DSE. At present, the listed companiesinclude Tanzanian Breweries, Tanzanian Oxygen, Tanzanian Tea Packers, TanzanianCigarette Company and Tanga Cement. Five stockbrokers have been licensed to operatebut the market remains constrained because of limited stocks and low liquidity.

The sale of GoT’s majority stake in the National Commercial Bank (NCB) is planned formid-2005. Combined with the recent establishment of a credit bureau to screenborrowers, modernization of the land registry to enable use of land as a collateral, andplans for restructuring of the state-owned People’s Bank of Zanzibar, Tanzania PostalBank and Tanzania Investment Bank, the efficiency of the financial sector is expected toincrease, improving GoT’s efforts to increase access to credit (IMF 2004).

Infrastructure - electricity. In 2001, Tanzania had a large and unexploited potential forpower generation, with a relatively well developed but costly hydroelectric system, byAfrican standards. The source of the cost inefficiencies were related with a weakdistribution system and low revenue collection. 85 percent of the hydroelectricity wassupplied by the National Grid; the remainder was thermal. The power company,TANESCO, was financial unviable, depending on GoT for large operating transfers. Withthe anticipated coming online of the Songo Songo gas reserves which would enable theswitch from diesel to gas based power, the costs of electricity were expected to decline.The coming on line of the Kihansi hydropower project was further expected to boostpower supply from 350 MW to 530 MW (CEM 2001). In 2004, GoT indicated thattransfers to TANESCO would decline in 2005/06 due to ongoing financial managementreforms, and cost reductions facilitated by the switch from diesel to gas (IMF 2004). Inthe medium term, GoT would continue to prepare the company for eventual sale.

Infrastructure reforms.o In 1998, GoT established a Road Fund; around 2001, it established an

independent road agency to manage and maintain the roads network (CEM2001).

o In 2000, the anticipated privatization of telecommunications, water transportand power were to improve supply and lower costs (CEM 2001).

In 2001, only about 70 percent of the installed water capacity was fully used. 68 percentof the urban areas were covered but 52 percent of the urban coverage was disrupted bytechnical and commercial losses. By 2002, water sector reforms were expected tosignificantly improve coverage and reliability. [Did these happen?]

Combating corruption. In 2005, GoT expects to legislate a new anti-corruption law (IMF2004).

Improving the business environment. The launch of the Investors’ Roundtable (IRT) hasenabled dialogue between GoT and the private sector, focusing attention on critical issuessuch as the revised Land Act and a new business licensing system (IMF 2004).

Page 9: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

9

Human capital. Human capital that provides a near-guarantee for better payingemployment in a poor country like Tanzania, displayed disappointing trends. In additionto a large share of primary school dropouts3, in 2001, Tanzania had the lowest rates ofsecondary and higher education achievements in sub-Saharan Africa. In 2001, the grosssecondary school enrollment ratio was only 5 percent, far lower than Sub-SaharanAfrica’s 27 percent and East Asia’s 69 percent (update). In 1990, the share of itspopulation with at least some formal post-primary education was lower (3.2 percent) thanat independence in 1964 (5 percent). In spite of universal primary education, failure ofpost-primary education to keep pace with population growth, emigration of skilled labor,and the high cost of secondary and tertiary education relative to income levels inTanzania are some of the reasons for weak human capital accumulation. The HIV/AIDSepidemic is further eroding the productive and skilled segment of the labor force. In2001, the cost of treating one AIDS patient was equivalent to educating 9 primary schoolpupils.

As a large and fast growing manufacturing sector can be critical in achieving sustainableeconomic growth and in creating better paying jobs in Tanzania, using the micro-evidence, we need to resolve at least three issues for policy makers in sections 3 - 5:1) if private fixed investment has been declining, what explains manufacturing growth

rates of 8 percent per annum in 2002-2003? Is the recent spurt of growth reflectingthe efficiency gains associated with private investment climate reforms? How longwill this last?

2) Is the delayed response from private investors due to factors that policy makersshould have done/can do something about? This could imply accelerating outstandingreforms or provision of critical public goods such as physical infrastructure, laborskills that rapidly eroded during 1995-2000 as government implemented stabilizationreforms?

3) In light of a dampened private investment and employment response inmanufacturing, do we have evidence that makes a case for re-thinking the growthstrategy for manufacturing in Tanzania? Is there a compelling case for additionalactions that policy makers can take to reign in private investment and productivitygrowth?

2. TANZANIA’S MANUFCTURING SECTOR - A MICROECONOMIC PERSPECTIVE

STRUCTURE

The 2003 RPED firm survey for Tanzania covers the period 2000-2002 and comprises ofa sample of 276 manufacturing firms in 8 sectors – food and agro-industry, chemicals andpaints construction materials, metals, furniture and wood products, paper, printing andpublishing, plastics and textiles, garments and leather products. The sample can also bedisaggregated by firm size measured by the number of employees: micro (1- 5), small (6

3 In 1998, 70 percent of the primary school pupils failed to achieve a passing grade (CEM 2001).

Page 10: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

10

– 29), medium (30 – 99) and large (100 plus). The distribution of the firms across sectorsand sizes is shown below.

Page 11: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

11

Year Size 1 –Micro

Size 2 -Small

Size 3 -Medium

Size 4 -Large

Allfirms

1- 5 empl. 6 – 29 empl. 30 – 99empl.

100 plusempl.

Number of firms andshare in brackets

2002 16 (6.1%) 109 (41.3%) 72 (27.2%) 67 (25.3%) 276(100%)

Sales growth p.a.(%)

2001 -3.5 4.0 14.5 17.9 8.8

Sales growth p.a.(%)

2002 2.2 8.5 10.5 20.9 11.4

Sectoral share insample (%)

Sales growth in2001

Sales growth in2002

Agro industry 29.3 8 18Chemicals 9.8 14 13Construction 3.9 12 7Metals 10.5 0 21Furniture 23.5 6 3.5Paper, printing, publishing 9.1 10.0 10Plastics 2.5 12.0 16.6Textiles and garments 11.2 18.5 17.0All firms 100 (number =276) 8.8 11.4

Non-exporters Exporters Non-exporters ExportersSales growth Median median mean meanin 2001 8% 24% 66% 91%in 2002 11% 23% 41% 49%

71 percent of the firms are located either in the capital Dar es Salaam (39 percent) orother cities/towns with a population of over 1 million (33 percent). Most of the firms inthese larger urban centers are in heavy manufacturing such as plastics (71 percent of theindustry), construction materials (64 percent of the industry), and chemicals and paints,and metals (almost 50 percent of the industry). Towns with smaller populations of250,000 – 1 million house about 18 percent of all firms4 and even smaller towns withpopulations of less than 250,000 house the remaining 11 percent of all firms. More laborintensive-industries are located mostly in smaller towns with populations of 1 million orless. About 72 percent of food and agro-industry processing agricultural produce, 66percent of the furniture and wood products industry and 65 percent of the textiles andgarments industry is located outside the larger cities.

4 All numbers are derived from the sample (representative?).

Page 12: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

12

The average age (median) of a manufacturing firm is about 13 years. Only 7 out of asample of 276 firms are older than 50 years. The oldest firm, aged 82 is engaged inproducing food and agro-processing.

The firms survey is a good instrument for understanding private manufacturing sectordevelopment in Tanzania. Except for 2 percent that are publicly held and 1 percentowned by the GoT, the remainder are privately owned. Only 16 percent of all firms werepreviously state-owned; of these, 75 percent were privatized after 1996 when GoTaccelerated the privatization process. 40 percent of the privatizations occurred in 1997-98, 8 firms were privatized after 1999.

One fifth of Tanzania’s manufacturing firms have some degree of foreign ownership.Presently, the average share of foreign equity is 72 percent. Only about 5 percent of thefirms are 100 percent foreign-owned. The majority of the firms’ shareholders areTanzanian nationals (60 percent). The shares of non-African nationals is 14 percent,Kenyans, Ugandans and other Africans between 2 –3 percent each. The ethnicity of theprincipal owners is predominantly African (44%), followed by Asians (26 percent),Europeans (96 percent) and Lebanese (4 percent).

Start-up capital for about 72 percent of the firms came from the owner’s savings orinternal funds from some other source. Bank loans financed the start up of 15 percentand equity or sale of stock only 10 percent of the firms. Family and friends providedstartup capital for about 6 percent of the firms. Money lenders and informal sources areless important.

PERFORMANCE OF MANUFACTURING FIRMS

Aggregate growth in GDP per annum during 1999-03 was 5.4 percent, while that inmanufacturing was about 6 percent (Table 5).

Firm survey data for value added in manufacturing is unreliable; however, trends in salesgrowth during 2001-02 are consistent with aggregate growth trends.5 Firm sales grew at11 percent6 in 2002 and 9 percent in 2001 but the variance was large, ranging from (-) 26to (-) 98 percent for the bottom quartile to over 66 percent for 5 percent of the firms thatgrew the fastest. At least a quarter of the firms in both years grew at rates of –9 to – 99percent per annum.

Are large firms growing faster? Between 2000 – 2002, growth in sales increased with thesize of the firm7 measured by the number of employees, a trend that was also prevalent inTanzania in the early-mid 1990s. Large firms and exporters out-performed others withsales growth at about 19 percent in 2002. Sales in micro firms grew at only 2 percent. In 5 A reason why aggregate trends were probably slightly lower was that the aggregates also cover a largenumber of micro firms that were under-represented in the sample.6 Due to large variations, the measure of central tendency is typical the median in this paper, unlessotherwise stated.7 All results reported here are statistically significant as shown by linear and non-parametric tests ofstatistical significant (Pearson’s, Spearman’s, Kruskal Wallis etc.)

Page 13: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

13

general, between 2001 and 2002, except in medium firms, growth accelerated across allfirm sizes.

Are firms in some sectors growing faster than others? Growth across manufacturing sub-sectors was not uniform. The fastest growing sectors were metals (20%), agro-industry(18%), plastics (17%) and textiles/garments (16%). Furniture sales grew at only 3percent. To some extent, developments in other sectors of the economy reinforced thesetrends. Agriculture, which drives the largest sub-sector (agro-industry withinmanufacturing grew at about 5.5 % during 2001-03. Growth in mining and quarryingsectors that drive growth in metals, picked up in 2002-03.

DETERMINANTS OF GROWTH

For the period 2000 – 02, the key factors in Tanzania driving firm growth, defined asgrowth in sales, are age, growth in investment, growth in exports, and use of newertechnologies (Table ---). Firm size8, sector or firm location (large cities or smaller towns)in general do not affect growth9. Nor do other firm characteristics such as the nationalityor ethnicity of the firm owner seem to matter.

Older firms grew more slowly than younger ones. Firms that experienced higher exportgrowth also grew faster; for each percentage point increase in export growth, sales grewby 0.87 percent. This suggests that for globally competitive firms, the ability to export isa critical source of growth. It relaxes the demand constraint imposed by the size ofTanzania’s domestic market on firms who are unable to export. Firms with access tonewer technologies measured by the percentage of the workforce that used computers,grew faster too. For a one percent increase in the computer users in the workforce, a clearproxy for technical skills in the labor force, sales grew by 0.5 percent.

While aggregate trends for 1960 - 1999s show a disconnect between growth (proxied bysales here) and investment (CEM 2001), in contrast firm data suggests that during 2000-02, there was a positive, albeit weak, relationship between growth in sales andinvestment. An explanation for the appearance of this positive but weak link in the post-1999 reform period could be that the reforms are starting to appeal to investors in themanufacturing sector, even though slowly. While this analysis explains what drivesmanufacturing growth in Tanzania, it does not answer at least two related issues. One,why are the large variations in sales growth not associated with sector- or size-class-specific effects? And, two, can we explain whether the growth-investment relationshipapplies uniformly across all firms or it is peculiar to certain firm characteristics?

A quantile analysis of the determinants of growth in sales examined the variance ingrowth by focusing on the fastest (75th ) and slowest (25th) growing segments of the firmdistribution. While it did not answer the first question very well, i.e., whether certain

8 These findings contradict the traditional theory of firm growth (Evans (??), Nugent and Nabli (??)) whichargues that firm growth is a simple function of size and age.9 One explanation why firm size or sector do not seem to matter is that within each of these categories, thevariance in sales growth is large, and neutralizes the effect of a higher average growth rate.

Page 14: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

14

determinants of growth affect faster and slower growing firms differently, it confirmedthat export growth, technology and the firm’s age are robust determinants across all firms(Table --). For the fastest growing firms in the 75th quantile, location in larger cities andtowns, relative to firms located in towns with under 50,000 population negativelyaffected growth. Location did not seem to affect growth in any other quantile. Investmentgrowth did not explain the variance in growth in the tails of the distribution.

Table: Determinants of firm growth (sales) in Tanzania – 2000-02– a quantile analysis

The second issue related with the weak link between investment and growth is partiallyunraveled by firms’ age. An econometric investigation shows that firms aged over 25years had lower investment rates, lower rates of technological pick-up as given by the useof computers among the workforce and grew more slowly than their younger counterparts(Table ---). For example, sales in firms aged 25 or more grew at 8 percent or lesscompared to rates nearly twice these in younger firms. Firms aged 10 years or less grewat 19 percent. Growth of younger firms in the below 25 age bracket is strongly relatedwith investment growth rates; for a one percent increase in investment, sales grew byalmost 0.43 percent. In contrast, for firms over 35 or 45, investment growth was notassociated with firm growth. Interestingly, export growth does not seem to be sensitive

QuantilesAll firms 25% 50% 75%

Constant 14.7 ***(3.64)

1.33(5.22)

13.8***(3.2)

69.52**(20.07)

Age of firm -0.62***(0.65)

-0.53*(0.27)

-0.29**(0.13)

-0.55***(0.16)

% of computer users 0.46***(0.16)

0.29*(0.17)

0.25*(0.17)

0.33**(0.17)

Invest/asset02 0.19+(0.13)

0.07(0.14)

-0.09(0.06)

0.31(0.23)

Export growth ‘02 0.86***(0.14)

0.69***(0.16)

0.68***(0.08)

0.63**(0.30)

Location: default is towns with a population less than 50,000Loc : capital city 12.52

(28.37)-37.3**(20.18)

Loc : city pop.> 1 miln. 20.54(27.96)

-40.8**(18.4)

Loc: town pop. 250,000– 1 mln.

-4.58(31.96)

-45.04**(19.7)

Loc: town pop. 50,000– 250,000

-0.96(42.64)

-54.09***(20.9)

Notes: *** significant at the 1%** significant at 5%* significant at 15%+ significant at 16%Standard errors are shown in brackets, bootstrapping was applied.

Page 15: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

15

to firm age and was robust across age categories, but the use of computers is. For firmsunder 15 years, a one percent increase in computer use in the workforce increases growthby 0.64 percent. The contribution of computer usage halves with firm age but wasinsignificant for firms under 10 years.

Table: Link between firm age, investment and growthin Tanzania’s manufacturing sector – 2000-02

LessThan10yrs

LessThan15 yrs

LessThan20 yrs

LessThan25 yrs

LessThan35 yrs

LessThan45 yrs

Intercept 9.96** 6.8** 8.9*** 8.09*** 8.38*** 6.4**% of compusersin workforce

0.04 0.64*** 0.48** 0.38** 0.39** 0.31**

invnass02 0.57** 0.39** 0.38** 0.43** 0.13 0.18gexp02 0.69*** 0.64*** 0.66*** 0.68*** 0.71*** 0.73***

Number o ffirms tested

68 91 110 124 147 162

*** 1%** 2 – 10%

Issues for policy analysis: the determinants of growth in Tanzania’s manufacturingsector are firm age, growth in investment, exports and the level of technical skills in theworkforce. There is little that policy makers can do about firm age. Hence, the three keypolicy issues we now turn to are:

1) What are the determinants of investment in manufacturing and what can policymakers do about it?

2) What are the determinants of the ability to export and propel export growth inmanufacturing and what can policy makers do about it?

3) What can policy makers do to promote technical skills accumulation in the workforcein Tanzania’s manufacturing? Policies that support the adoption of newertechnologies such as those that enable the use of computers among the force promoteproductivity and growth.

3. INVESTMENT GROWTH IN TANZANIA’S MANUFACTURING SECTOR

1999 seems to mark a breakpoint in Tanzania’s recent investment trends, reversing therelationship between growth and investment. As noted earlier, growth in investment andGDP in manufacturing and the overall economy were unrelated during the 1960s – mid-late 1990s. From 30 percent in 1992-98, aggregate manufacturing investment measuredas a share of total private investment dropped to 23 percent in 1999 (Table 15, EconomicSurvey 2002). This was consistent with the drop in total private investment from 18

Page 16: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

16

percent of GDP in 1992 to 11 percent in 1999-03. In absolute 1992 prices, aggregateprivate investment in 1992-2003 was still below (TZ Sh. 240,679) its 1992 level (TZ Sh.243,527).10 In spite of the consistent drop in private investment, growth in manufacturingand overall GDP picked-up significantly.

In 2000, there was an upward level-shift in manufacturing investment as a share of thetotal private investment. The latter increased to 28 percent in 2000 and has since held atthat level. As this step increase in manufacturing investment in the last 3 – 4 years wascomplemented with higher growth in manufacturing, and indeed the overall economy, itprobably reflects an improved economic environment stemming from the reforms,especially in the late 1990s. In spite of the emergence of this positive albeit belated nexusbetween investment and growth, the level of private investment in Tanzania remains toolow to be a driving force of improved expansion in manufacturing in the future.

Micro-evidence from firm survey data validates that on average, investment growthstagnated during 2001- 02. When measured by the change in investment as a share offirm assets during 2000-02, investment grew by 0.2 percent per annum in 2001 and 2002.About half of all manufacturing firms invested in 2001-02 while the other half did not.These results also apply to other measures, such as, a change in investment as a share ofthe firms’ replacement or resale value of firm assets

In general, investment growth in younger firms aged 20 or less increased with firm age.Firms less than 10 years invested almost 10 times more than the average. Investmentgrowth in micro and small firms was nearly zero; in medium firms about 0.5 percent andin large firms it was 2.1 percent per annum. The range of investment growth for firms inthe top ten percentile in each size class was large, varying from 19 – 119 percent.Investors did not reveal a particular preference for any sector. Agro-industry, chemicals,construction, paper and plastics experienced investment growth of 1 – 3 percent perannum. Investment growth among exporters was 1.6 percent per annum in 2002.

DETERMINANTS OF INVESTMENT GROWTH IN TANZANIA

As in any other country, investment decisions in Tanzania are affected by a variety ofintricately linked factors. There are over100 distinct factors that affect investor decisions.As most of these factors are also strongly correlated with each, they create empiricalcomplications. 11 For example, the firms that are large as also likely exporters, less

10 With the exception of 1994, private investment levels remained below the 1992 level (Table 7, EconomicSurvey).

11 As the number of observations is far fewer than the number of explanatory variables, there is a highdegree of multi-colinearity and few degrees of freedom. These econometric problems are relativelycommon in firm survey data relative to household or labor force surveys that have a larger number ofobservations. The fact that there are a limited number of observations and a large number of explanatoryvariables in the dataset makes it difficult to model a clean and comprehensive investment story in one step.There is a high level of intra-group and inter-group multicollinearity which diminishes the predictive powerof a single multivariate investment model. A variety of innovative econometric approaches are used tograpple with this problem and arrive at empirically rigorous results.

Page 17: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

17

financially constrained, and have or can afford better infrastructure, better technologyetc.. In the absence of knowing their relative importance, they can easily result in alaundry list of policy recommendations for increasing investment in Tanzania’smanufacturing sector. This approach is not very useful for policy purposes as it makes itdifficult to identify policy priorities required for the proper sequencing of reforms andinformed policy decisions.

A muti-pronged methodology was used to resolve this issue. As a first order of priority,the relative importance of various policy reforms is examined. To maintain analyticaltractability and empirical rigor, the factors affecting investment decisions are organizedinto 7 – 8 different groups of determinants which are related to a broad area of policyreform such as financial, infrastructure, technology, labor, demand, etc.. Within eachgroup, there are a large number of factors which differentially impact overall investment.For example, within the financial group, the policy implications of issues related to start-finance are quite different from those related to access to working or investment capitalwhich are quite distinct from how the level of the interest rate affects investors, and soon. Hence, the second order of priority is to zoom in on the most binding constraintswithin each policy reform. This helps to identify which areas of the financial sector, tocontinue the example, need more attention and can contribute more to investment growth.Last, in keeping with tradition, the full investment story is pulled together. Because of theeconometric complications, an effort is made to keep the model simple with at most oneor two key variables from each policy area.

IDENTIFYING POLICY PRIORITIES FOR HIGHER INVESTMENT AND GROWTH

Time and resource constraints make it difficult for policy makers to focus on andimplement a large number of policy reforms at the same time. In the first round, therelative importance of each determinant-group that influences investment decisions in themanufacturing sector was assessed econometrically (Table ---). Firm demographicsdefined by age, size, sector, location and nationality of the owners explained about 13percent of investment growth across all firms 2000-02, and 28 percent in younger firmsaged 20 or less.12 In general, across firms, investment growth declined with age andincreased with firm size. The location of firms in towns with a with a population of amillion or less had a negative impact on larger firms but did not seem to affect smallerfirms. Relative to textiles and garments, firms engaged in metals and metals producingsectors had lower investment growth. Interestingly, across all firms, owners’ with ethnicorigins from Lebanon contributed positively to investment growth. In smaller firms,while owners whose nationality was Ugandan made a positive contribution to investmentgrowth, those of Tanzanian, Kenyan or non-African nationalities had the opposite impact.

12 As the age of manufacturing firms significantly affected the nexus between investment and sales growthin recent years (Table ---), the determinants of investment were analyzed for the cluster of allmanufacturing firms as well as those aged 20 years or less.

Page 18: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

18

Table ---: Indicative reforms priorities for increasing investment in manufacturing

All firms

Firmsunder 20or less

yrs.

Rankfor allfirms

Rankfor

firms20 orless

R2 R2

1Demographics – age, size, sector, nationalityof owner, location etc. 0.13 0.27

2 Demand 0.059 0.150 4 33 Financial group 0.286 0.220 1 1

4Assets & technology,equipment, capital intensity 0.117 0.176 2 2

5 Infrastructure group 0.113 0.216 3 16 Investment climate group 0.046 0.147 5 47 Labor skills, productivity, HIV etc. 0.112 0.228 3 1

Source: Authors’ estimates based on econometric investigation

Table – indicates the relative importance of the 6 key policy-related determinants ofinvestment in manufacturing pertaining to a particular area of policy reform. Firmdemographics is an exception as policy makers can do little about most of thecharacteristics, at least not directly. Within each reform area, there are as many as 15 – 20different factors that explain investment growth. The broad determinant groupings are:

1) market demand measured by sales in the domestic market to different customers sucha state owned enterprises, large firms, MNCs, or in export markets which generallynot constrained by the size of the domestic market;

2) finance as given by access to credit for start-up, investment and working capital, andinterest rates etc.;

3) assets and technology reflecting the age of the equipment, access to superiortechnologies, extent of computerization, capital intensity etc..

4) infrastructure as noted by the paucity of power, water, transport, telecommunications,and the extent of substitution between private and public infrastructure and its costsetc..;

5) quality of the investment climate as measured by the level of corruption, bribery,confidence in the judicial system, etc..

6) human capital – skills and education of workers, managers, HIV-related costs and thelevel of productivity.

Special attention was paid to younger firms aged 20 or less as these displayedsignificantly higher investment and sales growth in recent years and were the maindrivers of growth in Tanzania’s manufacturing sector. The top 4 priority policy reforms

Page 19: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

19

which affect older and younger firms were related with the financial sector, technology,infrastructure and human capital. For younger firms, the next priority area was themarket demand constraint.

Access to finance for start-up, investment and working capital was severe across theboard. Foreign firms had a slight advantage either because of parent companies easilyfinanced investment or because they had easier access to bank loans. Most domesticfirms were severely constrained in terms of access; those that had access found theinterest rates too costly to finance investment.

Good infrastructure and good labor skills are equally critical in explaining investmentgrowth in manufacturing firms in Tanzania. The nature of the determinants and theirsignificance levels underscore the complementarity between human capital and physicalcapital. For example, while transport and utilities (such as electricity and water) fallunder infrastructure, physical capital also encompasses superior technologies ofproduction reflected in the type and age of the machinery, use of modern informationtechnologies, level of computerization etc.. But all these, in turn, are critically dependentupon human capital embodied in the ability of workers to operate such technologies andlearn how to use superior technologies of production.

While market demand is ranked 3 or 4, the data makes a rigorous case for incentives forexport promotion. There are inherent spillovers between export-ability and investment inthe quality of physical and human capital, including better technologies. Apart fromenlarging the size of the market, exporting imposes on firms global competitivenessstandards that encourage the choice of technology associated with higher laborproductivity, dynamic learning processes and constant technological upgrading tocompete in the global marketplace.

INFRASTRUCTURE13 - A CONSTRAINT TO INVESTMENT GROWTH

Within the infrastructure group, the four main constraints are related with electricity,water, road transport and others factors such as telecommunications, internetconnectivity, waste disposal facilities etc.. Overall, infrastructure-related factors explainabout 15 percent of investment growth; for younger firms aged 20 or less with the fastestinvestment growth, this statistic rises to 23 percent, mostly because of the exacerbatedimpact of water and transport-related constraints (Table __). In order of importance, thetop three problem areas are (1) irregular electric supply leading to investments in theprivate provision of power for those who can afford it. (2) production losses related topoor transportation, frequently resulting in private provision of transportation. (3) Internetconnectivity for all firms boosts investment growth. In younger firms, investment inwater infrastructure to compensate for the irregular supply of public water has aparticularly negative effect on investment growth (Table ---).

13 See RPED Investment Climate Assessment, 2004 for details on the firm surveys.

Page 20: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

20

Table: Relationship between investment growth and infrastructure – identifying the bindingconstraints.

All firms Aged 20 or less

Coeff.Sig.

Level Coeff.Sig.

LevelIntercept 41.7 *** 25.60 **Demographic group - Age -0.43 *** - Small 3.40 * 6.62 *** - Medium 4.30 ** 10.51 *** - Large 8.4 ***Metals and related products -3.80 *Infrastructure group Electricity Own generator dummy -30.72 *** -18.15 ** Generator age* -0.76 *** -0.41 *Water Share of water provided privately -0.04 * -0.06 ** Share of water from public resources Own well dummy Own water infrastructure -4.95 ***Other Infrastructure Transportation Own road dummy 5.71 ** 8.53 *** Own waste disposal dummy 2.79 * 3.72 ** Own freight transport dummy -3.45 * -5.49 ***

Production losses due to poor public transport 0.23 *** 0.12 * Percentage of average cargo lost -0.21 * -0.11Telecommunications costs 0.000007 * 8.6E-06 ** Internet access dummy 4.04 ** Located in an industrial estate -4.85 *** -2.66 *Number of observations 231 171R- squared 0.15 0.23 Source: Authors estimates *** = sig, level 1 - 5% ** = sig. Level 6 - 15% * - sig. Level 16 - 30 %

Poor public infrastructure has driven many firms to invest in private infrastructure inTanzania. Apart from efficiency considerations related with the loss of scale economieswhen each individual small or large firm installs a private generator, for the privateinvestor the cost of installing private infrastructure is an additional cost of doing businessin Tanzania. Every Shilling invested in a private generator or water source or road is aShilling of forgone investment in expanding production and contributing to growth andemployment.

Page 21: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

21

About 62 percent of all firms are located in special industrial estates,14 presumablybecause of some public policy that either allocates industrial sites within these designatedareas to firms or attracts them there for some other reason. In any case, firms locatedwithin these estates do not appear to benefit from better infrastructure. Power outages arefrequent and impose high financial costs on firms. Measured in terms of a single spelllasting 8 hours (the equivalent of one working shift), power outages occur on average atleast 18 working days per year. Some firms experience disruptions for as many as 56working days per year. As many as 54 percent of the firms have installed private powergenerators at costs ranging from Tz.Sh. 172,700 – 812,500 per employee.15 Mostgenerators are about 7 – 8 years old. Generators lose efficiency with age and have to bereplaced. In sum, the recurrent costs of installing and maintaining private generators inthe face of power outages inflict high costs on investors, constraining investment growth,as evident from the large and significant coefficient on the electricity determinants intable --. Power outages and fluctuations also damage production equipment.

Industrial water is available to manufacturing firms from a variety of sources includingpublic. Due to irregular supply, about 40 percent of the industrial water is obtained fromprivate sources. While this affects investment growth adversely, as Table – shows, wateris relatively less constrictive for investors than power and transport. Its negative impacton investment growth is directly related to two factors: (i) the larger the share of privatelyprovided water, the more negative the effect on investment; and (ii) private investment inwater infrastructure has a particularly negative impact on investment growth in youngerfirms (Table --). Tanzanian firms obtain 58 percent of their industrial water from thelocal municipality, 28 percent from a private well, 10 percent from private vendors andthe remainder from communal sources. Private provision of water entails additional costsof construction to channel water from its source to factories. About 34 percent of thefirms own a well, 31 have invested in their own water infrastructure and about 12 percentshare a water source with the local community. The data does not permit comparisons ofthe costs of water from private and public sources, but as Table – shows, privateinvestment in water infrastructure substitutes for direct investment that could lead to firmgrowth. On average, firms paid water bills ranging from Tz. Sh 71000- 32700 peremployee in 2002 for the use of industrial water from all sources. The fact that a largeproportion of firms have not invested as heavily in private industrial water as powershould not be mistaken with water not being a constraint but rather its being a seriousdeterrent to investment growth in water-intensive sectors like textiles and garments whichare labor intensive or agro-industry and food which are potentially high-growth activitiesin Tanzania.

For a country with a rich rural hinterland that produces a variety of agricultural produceas input for agro-industry, the shortage of reliable freight transport by road, rail and air isa direct constraint to growth. About 41 percent of the firms lose anywhere between 1 – 50percent of their cargo in transit due to spoilage, breakage etc., adversely affectinginvestment growth. About 15 – 20 percent of all firms lose as much as 2- 3 percent ofproduction and sales respectively due to poor transport facilities. In extreme cases, as

14 Presumably, the industrial estates provide better or less costly infrastructure and possibly, tax breaks.15 The higher range applies to firms in the upper 10 percent of the distribution.

Page 22: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

22

much as 80 percent of total sales are lost. Empirically, transport-related lost productionor sales, and the private provision of roads seem to affect investment growth positively.One explanation for this could be that existing investors have factored in transport-relatedlosses in their investment decisions. After incurring the fixed cost of investment in aprivate road, the latter is an asset which has a positive effect on future investment growth.The same does not apply to infrastructure that requires large but recurrent investments.Investment in generators, water infrastructure, and freight transport that need regularupgrading or replacement negatively affect investment growth in the firm. At least 10percent of the firms invest in private roads; 18 percent have invested in freight transportto cart goods back and forth. About 26 percent of firms have invested in transportationfor workers to compensate for lack of good public transport to and from the workplace.

Another area with problematic infrastructure is telecommunications. Only 49 percent ofthe firms have internet connectivity, but power outage-related spillovers diminish theability of these firms from accessing information when they want. Firms pay about Tz.Sh. 11428 per employee per year in telecommunications bills. While this may be seem ameager amount, the opportunity cost of regular and easy access to informationtechnology is yet another factor that constrains Tanzanian firms from linking better withglobal markets in pursuit of faster growth.

Key policy issues: power, transport., especially freight, roads, telecommunications(internet); for younger firms, also water.

CAPITAL/BANK CREDIT – A CONSTRAINT TO INVESTMENT GROWTH

Constraints related with the financing of investment in Tanzania are numerous. Thedegree to which investment financing is a constraint is largely determined by thenationality, ethnicity of the investors and firm owners, the share of foreign equity in thefirm, or the firm’s ability to export. Most of these are demographic characteristics thatpolicy makers cannot do much about. Such factors allow investors to transcendconstraints like access to bank finance and its price, the interest rate otherwise imposedby the financial sector on domestically owned firms. As the financial sector hasundergone several important reforms in the recent past, in this section, we focus onaccess to bank loans as a source of startup, investment and working capital. The role ofinterest rates is discussed descriptively as the data is bad.

In 2001-02, the banking system played a relatively negligible role in meeting thefinancial needs of the manufacturing sector. Our analysis shows that (a) only about 20percent of the domestically owned firms, mainly non-exporters, use bank loans forinvestment and operational purposes. Exporters or foreign owned firms relied more onprivate foreign finance, especially for investment purpose. (b) Exporters rely more onbank loans as a source of working capital. (c) Availability of bank finance had a positiveand significant effect on investment growth (table___). Two key policy issues arise. First,increasing the level of accessibility to bank finance for domestic firms by widening anddeepening of the banking system is necessary to facilitate investment growth in themanufacturing sector. Greater coverage and new products for different needs are needed.

Page 23: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

23

Second, a decline in interest rates would facilitate investment growth by reducing the costof working capital for exporters and the overall cost of financing and debt for all firms.

Start-up finance. Firm ownership drives with access to sources of investment andworking capital. For examples, firms that have access to bank loans for start-up also haveaccess to bank loans for investment and working capital (Annex 8). Among non-exporters, about 77 percent of the equity was held by the domestic private sector; about16 percent by foreign owners and less than 5 percent by government. For these largelydomestically-owned firms, 7 percent of the startup capital came from informal sourcesand 10 percent from the banking system. In contrast, exporters obtained nearly twice asmuch equity (31%) from foreign firms who also facilitated access to capital forinvestment and working finance. Most exporters relied less on debt capital - only 5percent of their start-up capital was loan-financed relative to 10 percent for non-exporters. Table __ shows that there was a positive and significant relationship betweenbank loans and investment growth in 2001/02, especially among younger firms aged 20or less. However, for start-up, relative to bank loans, the importance of equity or privatecapital is almost five times more important. And the latter is limited by personalconstraints. This makes clear that focusing on increasing firms access to bank loans as asource of start-up can go a long way towards increasing investment growth in Tanzania’smanufacturing sector. Annex 8 clearly demonstrates that this applies across all types offirms.16

Bank loan financed investment and working capital, and interest rates. Althoughrelatively small, the importance of bank loan financed firm capital is critical formanufacturing firm investment and growth. Although it does not emerge in themultivariate analysis presented in Table __ below due to severe multicollinearityproblems, there is a strong, significant and positive relationship between investmentgrowth and access to working and investment capital from banks (see Annex 8), and theirprice or interest rate.

As the two types of capital are essential for firm operations and growth, both areconsidered here. Firms who have access to bank loans for investment capital also haveaccess for working capital and vice versa, and are sensitive to interest rates. Only 20percent of the domestic firms have access to bank loans for investment and workingcapital. For this small proportion, bank loans financed about 14 percent of totalinvestment and 10 percent of all working capital. The balance of investment capital islimited to retained earnings. The remaining domestic or non-exporting firms whorepresent over 80 percent of all manufacturing firms (Note, exporters comprise less than20 percent of all firms) do not have access to bank capital. Loan applications carry a hightransactions cost and the negatively relationship between applications and investmentgrowth indicates that most of those who applied, were rejected and could not invest. Thesmall proportion of forms who do have access to bank loans are sensitive to the interestrate for investment purposes but indifferent with respect to working capital. The policyimplication of this is that 16 Due to severe multicollinearity, this does not show up in the multivariate in Table __- but is clear fromAnnex 8 that examines this relationship rigorously at the bivariate level.

Page 24: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

24

1) a reduction in interest rates can boost investment growth for firms who haveaccess.

2) Additionally, extending access to 80 percent of the domestic firms who arelargely non-exporters will also help to increase investment in manufacturing.

Exporters use bank loans to finance working capital more than investment capital (forwhich they rely more on their foreign partners/parents or retained earnings), but are alsosensitive to the level of the interest rate (Annex 8). Over 25 percent of their workingcapital relative to only 10 percent for non-exporters is financed by bank loans. The cruxis that even if investment capital is relatively cheaper from their foreign parents/partners,expensive working capital can deter investment growth. Exporters use fewer bank loansto reduce the overall cost of debt-financed investment in spite of considerably lowerinterest rates. In 2001-02, exporters paid interest rates of about 10 percent compared over16 percent for firms that were more dependent on the banking system. Nevertheless, incomparison to all time low world interest rates, 10 percent was relatively high anddeterred bank financed investment for expansion. The story of exporters reinforces thecase for a lowering of interest rates to facilitate cheaper working capital for exporterswho are the key drivers of investment and growth in Tanzania’s manufacturing sector.

Exporters or firms with links to foreign firms also enjoy several exclusive benefits thatfacilitate working capital, especially at a short notice, through overdraft privileges withdomestic banks. Nearly 49 percent of the exporters had overdraft privileges with thebanking system in 2002 compared to only 28 percent for non-exporters. Table __indicates that for younger firms, such a privilege is particularly conducive for investmentgrowth.

Other non-bank factors that affect investment growth are:1) firm profits - the larger the share reinvested, the higher growth in investment (table

__).2) Being a ‘priorsoe’ applies to a firm that was formerly state-owned, and is measured

by the number of years since its privatization. So, the longer a firm has beenprivatized, the more closely it would mimic a private firm that was never an SOE.The negative sign makes sense – firms that were only recently privatized experiencedlarge inflows of investment capital but the older ones are similar to the majority ofprivate firms whose investment levels were flat.

3) There is a significant and positive relationship between investment growth and theshare of foreign ownership in the firm (Table --- shows a). A higher share of foreignownership facilitates private foreign financing besides fetching several otherfinancial benefits associated with foreign ownership.

Page 25: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

25

Table: Relationship between investment growth and finance for investment –identifying the binding constraints

HUMAN CAPITAL AND INVESTMENT GROWTH

Analysis of the contribution of human capital to growth, especially in countries with alargely unskilled, and less educated labor force presents the problem of a sample bias –by definition, all existing firms have self-selected to work with the available workforce

All firms Aged 20 or less

Coeff.Sig.

Level Coeff.Sig.

LevelIntercept 4.83 **Demographic group - Age 0.16578 * -0.455 *** - Small 5.433 *** 2.8577 - Medium 5.586 ** - Large 3.7 * *** Metals and related products -5.67 ** Owner's ethnicity - Lebanese 12.33 *** 13.519 *** Owner's nationality Ugandana Sources of start-up finance Owner's saving /internal fund Equity, sale of stock 0.347 *** 0.165 *** Bank loan 0.06 * 0.099 *** Family or friends Share of private domestic ownership Share of private foreign ownership 0.040 * Prior SOE -0.302 *** Overdraft facility dummy 3.325 *Applied a bank loan dummy -4.23 ** -4.385 **Share of working capital financed: Private resouces Bank loans Public resouces/funds Share of investment capital financed: Private resouces Bank loans Public resouces/funds Share of profits reinvested 0.050 *** 0.06 ***Interest rate on bank loans Number of observation 151 116R-squared 0.439 0.327Source: Authors estimates *** = sig, level 1 - 5% ** = sig. Level 6 - 15% * - sig. Level 16 - 30 %

Page 26: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

26

while potential employers who could create better paying jobs by employing skilledworkers have chosen to invest elsewhere. It is therefore not surprising that thecontribution of human capital to investment growth in Tanzania, relative to that of otherdeterminants, is understated in the analysis that was conducted (Table ---).

TABLE: RELATIONSHIP BETWEEN INVESTMENT GROWTH AND HUMAN CAPITAL –IDENTIFYING THE BINDING CONSTRAINTS.

All firms Aged 20 or less

Coeff.Sig.

Level Coeff.Sig.

LevelIntercept -4.69 6.27 *Age of firm -0.033 -0.542 ***Males less than 30 0.00127 -0.0586 *Females 30 - 45 years 0.270 ** 0.211 **Effect of only education of employees Primary education - complete 0.118 *** 0.112 ***Secondary and vocational education Tertiary education or a diploma Graduate and post-graduate education Worker training 3.03 * 2.228 Effect of HIV in skilled workers -0.017 -0.00096 Effect of HIV in unskilled workers 0.008 -0.00622 Productivity (sales revenue) per non-mgworker -4.760-E8 * -4.2E-08 **Manager's education 1.20 * 0.02539 Manager's prev. export exp -3.25 -0.277 Manager's work experience Number of observations 152 103 R-squared 0.093 0.177 Source: Authors estimates *** = sig, level 1 - 5% ** = sig. Level 6 - 15% * - sig. Level 16 - 30 %

The key human capital determinants of investment in Tanzania from the self-selectedsample of firms are age and education levels of workers, worker training in addition toeducation, productivity of the non-managerial workforce, education and experience ofmanagers, and the adverse impact of HIV on firm profits. Severe multicollinearityproblems complicate the modeling of these highly correlated factors in the multivariatemodel in Table…. In spite of this problem some factors such as workers age, education,and skills training emerge as being significant and positive but others like productivityhave the wrong sign and yet others are masked by econometric problems and areinsignificant. In any case, Table __ shows that among variables with policy relevance,primary education, worker training, and managers education affect investment growthpositively. Worker productivity measured as sales per non-managerial worker shouldhave a positive sign but appears with a negative sign and a coefficient that is almost zero.Closer examination through bivariate analysis shows that worker productivity is

Page 27: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

27

positively associated with investment growth, as would be expected. Before dismissingthe remainder of the human capital determinants as being unimportant for investmentgrowth because they fail to appear in Table ___, further econometric scrutiny wasconducted and is reported below.

About 44 percent of the manufacturing workforce had completed primary education,another 25 percent had secondary education, 12 percent had some vocational educationand 7 percent had tertiary or a diploma. In firms with higher investment growth such asexporters, the proportion of graduates or post-graduates was twice as high at 10 percentcompared to 5 percent in firms that sell domestically. Worker training at the firm levelwas also more prevalent in exporting firms (71 percent of the workforce had receivedsome type of formal training compared to only 47 percent in non-exporters). Thesedescriptive statistics suggest a positive relationship between investment growth andhigher education, especially graduate and post-graduate and formal training in themanufacturing sector.

Rigorous bivariate investigation shows that there is a strong, positive and significantrelationship between investment growth and primary, secondary, vocational, diploma(technical) and graduate education in Tanzania’s manufacturing sector. The same alsoapplies to worker training (Annex 9). Primary education is linearly related withinvestment while other types of education and training have a non-linear relationship.Moreover, worker training at the level of the firm is strongly associated with the share ofthe workforce that has secondary, graduate, diploma or other types of technical or tertiaryeducation – a fact that may explain why firms in Tanzania or other sub-Saharan Africancountries invest less in training compared to firms elsewhere (East Asia for example).The presence of a large workforce with mostly primary education is hardly the suitableinput for investment in firm level training which is costly. This finding has serious policyimplications:

1) If emphasis on education is not extended beyond primary and secondary levels,firms in manufacturing will not invest in training and investment growth willcontinue to remain low.

2) Low skills or no skills are also constrains labor productivity which requiresworkers with superior skills to operate new technologies or learn better processes.

3) Tanzania’s manufacturing sector will either not attract or only attract investorsshopping for low-skilled workers. The type of low-wage employment that thepaucity of skills creates is inconsistent with the country’s vision of movingworkers out of poverty. To move out of poverty through manufacturing requiresmoving into higher value added production but the latter needs more skilledworkers.

4) More importantly, Tanzania’s low labor cost advantage is not permanent. Withincreasing integration into global markets, its ‘low skills--low cost’ workers willhave to compete with ‘high skills –low cost’ workers from other competitors,especially those in Asia. Early signs of this are clear from the stagnant investmentrates in manufacturing, a further testimony to the sector’s unsustainable growth,the recent 8 percent notwithstanding.

Page 28: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

28

Unfortunately, data problems do not permit a meaningful analysis of the impact of HIVon firms investment rates. But from other sources of information, it is apparent that thepandemic is inflicting severe costs on firms through decreased labor productivity.

TECHNOLOGY, QUALITY OF FIRM EQUIPMENT AND INVESTMENT GROWTH

An earlier section of this paper noted the strong relationship between overall growth inmanufacturing and the use of better technologies of production in relation to other firmsin the sector. A good example is the proportion of computer-users in the workforce whichindicatives the level of worker skills and the sophistication of the machinery andequipment in use (i.e., automated processes with computerized controls as opposed tomanually driven machines and implements).

Table: Relationship between investment growth and technology andcapital stock – identifying the binding constraints

All firms Aged 20 or less

Coeff.Sig.

Level Coeff.Sig.

LevelInterceptDemographics 1.9 - Age -0.0319 -0.527 *** - Small 5.70 *** 7.75 ** - Medium 6.734 *** 9.44 ** - Large 4.031 * 4.83 Metals and related products -4.451 ** -6.54 **Technology parameters:Share of machinery less that 10 yearsold 0.055 *** 0.029 *Assets employment ratio (K/l) -4.66E-8 *** -3.9E-08 ***Capacity utilization -0.038 * -0.004 Investment in technology 5.41 *** 4.24 **Share of computer users 0.04 0.0012 Investment in R&D -0.157 0.82 Number of observations 241 161R-squared 0.265 0.17Source: Authors estimates *** = sig, level 1 - 5% ** = sig. Level 6 - 15% * - sig. Level 16 - 30 %

Page 29: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

29

For manufacturing firms, technology and the quality of assets are the second mostimportant determinants of investment growth (Table --- above). This makes sense as thequality of machinery or equipment and the technology in use jointly affect the level ofproductivity and profits. The three key policy related determinants of investment inTanzania’s manufacturing sector are (i) the proportion of a firm’s machinery that is 10 orless years old – the larger the share, the better the performance, and the greater the impacton investment growth; (ii) the capital assets to employment ratios reflecting the capitalintensity of production – the higher this ratio, the more investors need to invest in capitalto obtain the same unit of output. Not surprisingly, this has a negative impact oninvestment growth; and (iii) the proportion of firms that invest in the use of newtechnology. This factor contributes the most to investment growth and is strongly relatedwith the proportion of computer users in the firm, as well as the level of investment inR&D to develop or acquire new technologies (table --- above). The last two variables arenot significant due to severe multicollinearity in the model.

Another factor such as the level of capacity utilization (that averages about 60 percentacross the sector) should also have a positive correlation with investment but appearswith a negative sign and is insignificant. This is probably due to multicollinearity in themodel (Table__). At the bivariate level, there is a robust relationship between the two.Further analysis reveals that capacity utilization and the age of the firm’s machinery thatis less than 10 years old are strongly correlated (the source of the multicollinearity in themodel!). This makes sense.

o About 48 percent of all firms including 65 percent of all exporters had invested innew technology during 1998 – 2000. Micro firms made hardly any investmentwhile between 61 – 77 percent of the medium and large firms invested in newtechnology. Agro industry, chemicals and plastics were the leading sectors in

Page 30: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

30

which over 60 percent of the firms invested in new technologies. The key sourcewas technology licensed from a foreign owned company – on average about 37percent of the large firms used this source while the numbers got smaller forsmaller firms.

o Across all firms, only 5 percent of the workforce could use computers but thisnumber applied mainly to medium and large firms for whom the average was 8and 15 percent respectively. Interestingly, relative to exporters, more workers innon-exporting firms use computers. Among sectors, textiles and garments andwood and furniture has the least proportion of computer-users.

o About 22 percent of all firms invested in R&D to develop new technologies. Thenumbers rose with firm size. Micro firms did not invest at all, but as many as 43percent of the large firms invested in R&D to develop new technologies. About30 percent of exporters compared to 19 percent of the non-exporters invested,indicating that export-orientation introduces incentives for firms to maintain theirglobally competitive edge. The average spending on R&D varied from Tz. Sh. 0.3mln in small firms to 5.0 mln in large firms. Relative to an average of Tz. Sh. 3.9mln for all sectors, firms in agro industry and metals invested an average of Tz.Sh. 10.0 mln each, while the outlier was the furniture and wood sector in whichfirms invested as much as Tz.Sh. 300.0 mln..

o On average, 12 percent of the firms were ISO certified with the highest proportionamong large firms (37 percent) and exporters (24 percent). Firms manufacturingplastics, textiles, construction materials and agro products had higher levels ofISO certification than others.

These descriptive statistics suggest that nearly half of Tanzania’s firms recognize thebenefits of new technology and have invested in it to some extent, acquiring it fromvarious sources. This is a positive trend. However, if the outcomes are measured by theproportion of computer users in the workforce and the level of ISO certification, the lattersignaling global competitiveness, these firms have a long way to go. This point isunderscored by the level of ISO certification among exporters (43 percent). Conspicuousis the fact that the technology wave has nearly bypassed smaller firms who neitherinvested in, or use modern technology in production. Given the critical importance ofthis source for overall firm growth, as well as investment growth, it seems worthwhile forpolicy makers to take on the challenge of figuring out what ,if anything, can be done bygovernment to promote the adaptation and adoption of superior technologies at the firmslevel. In this context, the experience of other developing countries, outside the Africabox may be instructive (see new study on the How to of Technological Change for FasterGrowth, forthcoming.). Policy makers need to explicitly focus on the following issues toboost investment growth in non-traditional ways through the use of better technologies:

Page 31: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

31

1) given that capacity utilization is constrained by the age of the machinery and thatyounger machines also had a significant and positive impact on investment growth,are there any tax incentives that policy makers can offer firms to facilitate thereplacement of outdated technologies to improve productivity and growth? Ifimplemented efficiently, such a policy should more than pay for itself throughincreased sales and tax revenues.

2) As the capital intensity of production negatively impacts investment growth, it maybe worthwhile for policymakers to think of policies that can improve laborproductivity to compensate for the substitution of capital for labor. This requires amulti-pronged approach across the domain of human capital improving policies thatcan accelerate the production of technical skills, and the level of computer literacy.The larger public goods’ aspect of these goods requires coordinated public action toincrease investment growth.

3) Given the role of new technologies and the relationship between the latter and largerfirms and exporters on the one hand, and the higher rates of investment growth inthese groups on the other, policy makers may want to consider what types ofinterventions other developing countries have used to promote better technologies toin turn, promote export growth and overall manufacturing growth. As it stands, theseefforts are far too small and ad hoc. It is not clear whether the investment in R&D forbetter technology is productive and whether research at the individual firm level isthe most efficient. Most successful countries have exploited the late comeradvantage to rapidly accelerate use of new technologies with efficient support frompublic policy. There are numerous unexplored options available for Tanzanian policymakers to pursue this hitherto neglected source of investment and overall growth.

MARKET DEMAND AND INVESTMENT GROWTH

Lack of adequate demand for goods manufactured in Tanzania is the fourth mostimportant constraint to investment growth for all firms. For younger firms, this constraintis ranked number three. In light of the relative inward-orientation of manufacturing firms,this finding is neither surprising, nor at odds with the state of manufacturing firms inmost of Sub-Saharan Africa. Only about 27 percent of the firms are able to export someamount of their output and no more than 18 percent at least 10 percent of it. Hence,Tanzania’s manufacturing firms are constrained by the size of the domestic market thatis further limited by low purchasing power, a manifestation of the country’s low incomestatus. Even exporting firms are heavily dependent on the domestic market where theysell an average of 80 percent of their output.

TABLE: RELATIONSHIP BETWEEN INVESTMENT GROWTH AND DEMAND – IDENTIFYING THE BINDING CONSTRAINTS.

All firms Aged 20 or less

Page 32: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

32

Coeff.Sig.

Level Coeff. Sig. LevelDemographics - Age 0.009 -0.37 *** - Small -5.398 * -7.76 ** - Medium 3.415 * 2.83 * - Large 2.783 2.54 Metals and related products -5.088 ** -5.31 *Export growth 0.088 *** 0.10 ***Proportion of output solddomestically 0.05 *** 0.10 ***Sales to MNCs 0.021 -0.01 Sales to SOEs -0.17 * -0.14 Sales to large domestic firms 0.0847 ** -0.05 Number of observations 182 119R-squared 0.19 0.317Source: Authors estimates *** = sig, level 1 - 5% ** = sig. Level 6 - 15% * - sig. Level 16 - 30 %

Other than continued macroeconomic stability, especially a favorable exchange rate,there are few direct policy levers available to raise aggregate demand in the domesticmarket. However, policies that can accelerate export growth are more likely to affectinvestment rates. As in other small-open economies, investment growth in Tanzania’smanufacturing sector is nearly twice as responsive to higher export growth (as opposed tothe share of output exported) than the level of domestic demand (measured by the shareof output sold domestically, Table ---). Sales to sate-owned enterprises hamperinvestment growth. In contrast, sales to large domestic firms increase investment byalmost as much as export growth. This is hardly surprising given that larger firms andexporters share many common characteristics. Among younger firms, investment growthis equally and strongly responsive to export growth and sales in the domestic market.

The policy implications for raising aggregate demand to increase investment growth areclear: given that domestic demand will continue to remain limited by households’ level ofincome/purchasing power, policies that can accelerate manufacturing export growth arethe strongest policy levers available to decision makers. In addition to a competitiveexchange rate, such a strategy requires a focused approach to improving the global, asopposed to the regional competitiveness of Tanzania’s manufacturing firms. Amongvarious options available are ones that can raise productivity through the use of superiortechnologies, aggressive marketing reforms to help firms penetrate new markets and astrengthening of policies that can spur nascent exports. The fact that the majority of firmsselling in the domestic market perceived competition from cheaper imports (not lack ofcustomers) as a threat to growth underscores this point.

Page 33: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

33

INVESTMENT CLIMATE AND INVESTMENT GROWTH

Adequate data to assess the overall quality of Tanzania’s investment climate given by thetime taken to obtain various clearances to operate in the country, and the official and un-official costs (bribes) of doing business is not available from the survey. In most cases,the number of firms that responded to this section is too few. As the investment climate isan important factor in investor decisions, two approaches are used to work with thedeficient data. First, to give policymakers an idea of the systemic inefficienciesassociated with the time costs (delays) and financial costs associated with the clearancesneeded to manufacture in Tanzania, the information provided by firms, albeit sparse, ispresented descriptively. This provide direct pointers for further streamlining efficiency-improving procedures to reduce the general costs of doing business in the country.Second, to econometrically estimate the nexus between investment growth and thequality of the investment climate, we give government departments responsible forproviding clearances the maximum benefit of the doubt by treating firms that did notprovide answers as having no problems, i.e. they do not have to wait for too long, the feeassociated with each procedure is not too high, they do not pay any bribes etc., i.e. allmissing values are treated as being equal to zero. It may be noted that as this approachgrossly understates the problem: for a balanced view, any judgments regarding thebusiness climate should jointly consider the descriptive and econometric evidence.

1. The data on delays in customs clearance is good. 81 percent of all exportersresponded. For customs clearances for exports and imports, the normal waitingtime is 20 days (average); the longest time it took during the previous year was,on average, 36 days. For a few firms, the delays were as long as 97-150 days.

2. The delays in obtaining inspection certificates seem reasonable for non-exportersbut not exporters who have to wait more than twice as long. About 40 percent ofthe latter reported that it took them 12 days (maximum of 72) to obtain clearancesfrom 6 different inspectorates (tax, labor and social security, fire safety,sanitation, policy, environment etc). 71% of the exporters who responded said ittook government an average of 27days to give them clearances (one firm reporteda maximum of 410 days). Except for micro, all other firms pay bribes to obtaininspection clearances.

3. On average, it takes about 7 days for a firm to obtain each of the followingconnections: water (33% of firms), import licenses (33% of firms) and operatinglicenses (77% of firms). While the 7 day period seems fairly reasonable, there aretwo issues that suggest that improvements could reduce these time costs: one,unless, government departments coordinate the service delivery connections, ittake a firm as long as 5 weeks to get all connections and become operational; andtwo, averages mask the enormous delays incurred by some firms. The maximumtime taken for any of these connections varied between 180 – 360 days, sufficientto drive the potential firm out of the market! The bribes paid to obtain theseconnections on average ranged from Tz. Sh. 35,000 for a water connection to Tz.Sh. 150,000 for an import license.

Page 34: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

34

4. Obtaining a telephone connection takes as long as 14 days and an electricconnection takes about 30 days (33% reported). While the responses were lowerfor bribes paid, the numbers range from Tz. Sh. 50,000 for a telephone line to Tz.Sh. 60,000 for an electricity connection.

5. Registering or re-registering a firm, obtaining a basic permit, construction permit,or any other permit on average takes between 7 – 10 days each (about 30 – 77percent of firms answered each question). The problem is with firms, typicallymedium and large ones, who often have to wait as much as 180 days.Construction permits take the longest. For one firm, the wait was as long as 998days. The bribes paid for such permits are correspondingly high ranging from Tz.Sh. 50,000 – 180,000 with the outliers in the range of Tz. Sh. 0.5 – 2.4 million.84% of all firms hire a company/agent to assist with registration.

After adjusting the data for missing values, the time costs associated with delays inobtaining clearances and permits seem to affect investment growth more negatively thanthe financial costs of either bribes or fees17 (Table ---). The number of days taken toregister or re-register a firm seem to affect investment growth in all firms more than thedelays associated with inspections or obtaining a service connections, although all areequally significant. For younger firms, none of the indicators of the cost of doingbusiness in Table one seem to affect investment growth adversely. In fact, therelationships are often positive though insignificant. Perhaps, one reason for these oddrelationships is that the costs of doing business have been internalized by incumbentfirms who treat them as fixed costs; hence, the rate of investment in manufacturing is notresponsive to them.18 This hypothesis is also complicated by the fact that the rate ofinvestment growth was almost stagnant during 2000-2002.

Whether evaluated descriptively or technically, the policy implications of the costs ofdoing business in Tanzania are significant and serious. These costs not only constraininvestment growth in incumbent firms but likely deter potential firms from entering thecountry’s manufacturing sector. This may be one explanation why investors have notrushed to invest and expand Tanzania’s manufacturing sector in several decades, asevident from its meager share of 8 percent of GDP in the past 3 –4 decades.

17 For all the reasons listed earlier, these results must be evaluated in conjunction with the descriptiveevidence in the preceding paragraph.18 This argument is also validated at the bivariate level when most of these relationships are positive orinsignificant with respect to investment growth. These results apply with or without adjustments formissing values.

Page 35: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

35

TABLE: RELATIONSHIP BETWEEN INVESTMENT GROWTH AND THE BUSINESS ENVIRONMENT– IDENTIFYING THE BINDING CONSTRAINTS.

All firms Aged 20 or less

Coeff. Sig. LevelCoeff.Sig.

LevelIntercept 1.78Age -0.039 -0.5 ***Small firms 6.45 *** 10.72 ***Medium firms 8.58 *** 13.22 ***Large firms 7.25 *** 10.22 ***Metals and related products -5.46 ** -4.76 *Dummy for number of bribespaid 2.02 *Average number of days toregister/re-register firm -0.10 *** 0.076 ***Number of days to clear customs(X&M) -0.014 Number of days required forinspection -0.06 *** -0.003

Total value of bribes paid to getservice connections Tz. Sh. Number of days required to getservice connections -0.059 *** -0.008 Total official cost to register/re-register Tz. Sh. Total value of bribes paid toregister firm Tz. Sh. -0.00001

Total cost of inspections Tz. Sh. Total days of delay for allclearances to do business incountry 0.056 *** 0.086 ***Total cost of obtaining allclearances to operate in countryTz.Sh. 4E-06

Confident that judicial systemwill protect property rights Member of business association-dummy

Number of observations 256 173R-squared 0.23 0.313Source: Authors estimates *** = sig, level 1 - 5% ** = sig. Level 6 - 15% * - sig. Level 16 - 30 %

Page 36: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

36

OVERALL DETERMINANTS OF INVESTMENT GROWTH

As evident from the preceding discussion, since each policy area covers many policyissues, the policy-specific analyses helped to identify and rank policy levers that have thegreatest impact on investment growth. For example,

In the context of infrastructure, the priorities are: electricity, roads, freighttransport and access to the internet- these affect investment growthdisproportionately more than other services. For younger firms, access toindustrial water is an additional problem. While the establishment of industrialestates has attracted many manufacturing firms, they have failed to deliver thebenefits they were supposed to. (Help: what was the purpose of the industrialestates? )

In the context of the financial sector, availability of bank finance for start-upcapital, particularly for younger firms, is critical. Recent financial sector reformsand lower interest rates notwithstanding, the insignificance of bank loans forinvestment and working capital is evidence of the negligible role played by banksin Tanzania’s manufacturing sector.

In the context of human capital, the critical policy priorities for raising investmentare: higher rates of completion in primary education, formal training for theworkforce, manager’s education, and policies that can raise worker productivity.With respect to primary education, this is a biased result, as the absence of acritical mass of more educated workers has led incumbent investors to self-selectinvestment in low value-added activities that generate low skills-low wage jobs.For higher worker productivity, growth and better quality jobs, policy makersneed to consider longer term investments in technical education.

In the context of the use of superior technologies that can raise workerproductivity, public policies that can assist firms to acquire/replace newmachinery/equipment, invest in new technologies are key priorities.

In the context of reducing the costs of doing business, policies that can reduce thedelays caused in registering/re-registering firms, customs clearances, inspectionsand service connections are key.

In the context of market demand, policies that can promote more and faster exportgrowth are key.

Table – presents the combined effect of the determinants of investment growth inTanzania during 2000-02. Most of the results are consistent with the policy-specificinvestigation of the earlier sections.19 Investment grew least in micro firms and thefastest in medium sized firms. No particular sector mattered but investment growth wassensitive to location. Firms that were formerly state-owned (SOEs) dampened investmentgrowth (coeff –0.31). The more recent the privatization leading to a large boost of newinvestment, the smaller the new investment. The lumpy nature of investment validatesthis result, e.g. firms that invested in 2000 did not invest again in 2001.

19 In several cases, the signs are the opposite of those in the policy specific analysis (formal worker trainingfor example) or the coefficient is not significant (electricity generator dummy or age of the generator).These are attributable to the technical problems (severe multicollinearity discussed earlier).

Page 37: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

37

TABLE: THE DETERMINANCTS OF INVESTMENT GROWTH IN TANZANIA

- identifying the binding constraints

All firms Aged 20 or less

Coeff. Sig. Level Coeff. Sig. LevelIntercept Age 0.32 *** -0.29 *Location in industrial estate -2.80 -0.667

Location in small town (pop. < 50,000) -8.12 * -5.07 *Micro firm -7.88 * -9.397 **Medium firm 4.10 *Export growth rate 0.038 0.11 ***

Share of output sold domestically 0.025 0.119 ***Sale to large domestic firms 0.116 ** Share of foreign firm equity 0.016 Investment capital - bank -0.06 * -0.037 Working capital - private 0.06 ** -0.019 Prior SOE -0.31 ** Age of equipment < 10 yrs 0.09 *** 0.044 **Capital-labor ratio -1.02E-11 *** -3.4E-12 *

Workers education - primary complete 0.08 *** 0.052 **Formal worker training provided -4.72 ** -0.118Manager's education -1.439 **

Investment in new technology dummy 8.95 *** 5.95 ***Age of the electricity generator -0.14 Own generator -4.21 1.98Cost of telecommunications 0 0.00002 *Invested in private road 3.41 8.725 ***In transit cargo losses -0.48 ** Provide own freight transport -4.15 * -3.12

Transport-related production losses 0.35 ***

Dummy for number of times bribes are paid -4.08 **

Total bribes paid to do business in country 0.00009 *

No. of days to clear customs (X& M) 0.147 ** Averagenumber of days to register/re-registerfirm 0.248 ***Confident that judicial system will protectproperty rights -0.972 -0.736

Number of observations 142 107R-squared 0.47 0.52

*** = sig, level 1 - 5% ** = sig. Level 6 - 15% * - sig. Level 16 - 30 %

Table – highlights the importance of several policies levers, including those listed earlierin this section. Firm location has a strong impact on investment growth. Firms in larger

Page 38: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

38

cities/towns invested more than those in smaller towns with populations less than 50,000(coefficient of –8.12), suggesting that there is a case for urban agglomeration-relatedeconomies (in the form of better infrastructure, labor force, markets etc.). Firm locationin industrial estates had a strong but negative effect (coefficient - -2.80) on investmentgrowth. It appears, that the latter were created for some purpose but have ceased to serveit anymore (tax breaks? Better infrastructure?).

Policy issue : Infrastructure is among the most important determinants ofinvestment growth in manufacturing but it appears difficult for government toimprove the quality and quantity of infrastructure economy-wide, at least in themedium term. Is there a case for developing a few urban agglomerations (wouldthe industrial estates perform this function??) where government can target largerquantity (as in more electric and water connections, roads etc.), and better qualityindustrial infrastructure to attract investors? As most firms already finance mostservices to some extent, some cost recovery should be possible. Lessons fromexport processing zones could be applied more widely in this case.

The strongest contribution (coefficient of 8.95) to investment growth is associated withinvestments in new technologies that help firms to remain globally competitive and raiseproductivity. As was noted earlier, use of superior technologies contains several aspectsof public goods if it is to be promoted to develop the manufacturing sector as opposed toa firm, typically foreign-owned in Tanzania, who can afford it. Use of newer andyounger machinery, typically less than 10 years contributed positively to investmentgrowth (coefficient of 0.09), raised capacity utilization, and productivity.

1. Policy issue : Higher skills among managers and workers (technical andcomputers for automated processes), timely and current industry information,means of accessing such information in an affordable manner (i.e. access to anduse of computers), and the adaptation of known international technologies to localconditions are all aspects of technological capability building that have used theenabling hand of public policy in most countries to propel technological change inthe private sector. The same factors also raise worker productivity that attractsinvestment to Tanzania. Some technological upgrading can occur through theuse of newer, imported technologies in the form of machines and equipment.While this is a clear private good, given that the bulk of the firms are indigenous,policies that can assist firms to upgrade old/outdated technologies

Closely related to location was the negative contribution of industrial infrastructure formanufacturing firms. Consider the combined effect of electricity-related problems, goodslost in transit due to poor transport, especially freight, and the costs firms have to incurwhen they invest in roads is aggregated in table ---.

2. Policy implication: The aggregated effect of infrastructure deficiencies oninvestment growth is probably the single most important negative factor drivingthe low investment growth rates in the manufacturing sector. The policy prioritiesare clear: electricity, roads, freight transport and access to the internet.

The positive contribution of a workforce that has completed primary education(coefficient of 0.08) is more than offset by the negative effects of firm expenditures in the

Page 39: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

39

formal training of workers (coefficient of –4.72). One way to rationalize this outcome inthe context of Tanzania’s labor market characterized by a paucity of skills other thanprimary is that although the availability of trained workers enhances investment, for firmstraining imposes additional costs that are significant and high. At a broader level, it alsodeters investment in high value production due to the public goods aspect of training inan economy where skills have a high premium and lead to negative externalities.

Policy issue : In addition to primary education, public policy must focus ontechnical skills development and training to attract investment in high-valueproduction.

Bank financed investment capital (coefficient of –0.06) reduced investment growth ,probably because high interest rates made investment costly. In comparison, privatelyfinancing had an equally significant and positive effect (coefficient of 0.06) oninvestment growth. Besides addressing the direct financial constraint for investment orworking capital, there are several positive externalities associated with a financial sectorthat can reach out to more and smaller domestic firms at affordable interest rates. Forexample, access to superior technologies, private costs of infrastructure financing,investment in new machinery all require easy and timely access to affordable capital.

Policy issue : Access to commercial funds for investment and working capital,and lower interest rates are necessary to facilitate investment growth inmanufacturing – apart from direct investment in expansion of facilities, thisshould also enable firms to upgrade technology, purchase newer machinery andequipment, invest in private infrastructure and critical skills. If not addressed,investment growth will remain dependent on foreign firms who have access toparent company finance, or the availability of retained earnings.

For reasons discussed in the preceding section, the investment climate indicators are notsignificant in the overall investment model in table --- . Nevertheless, closer inspection inthe preceding section suggests that delays in obtaining licenses and permits are moreimportant than the financial costs.

1. Policy implication: Policies that reduce the overall time costs of obtaininglicenses and permits, preferably by training government officials in operating aone stop shop for all investors, will go a long way in reducing the costs forincumbent firms and likely attract new investors who presently opt for otherlocations with lower time costs. Further, polices that can instill more confidencein the judicial system to protect private property rights are needed. Longer termleases (100 years?) to attract fixed investment in land and buildings is one way togo. Commercial courts that can speedily resolve disputes is another.

Technical problems similar to those in the case of doing business resulted in masking thestrong and positive effect of export growth on investment growth. Of course, as only 18percent of the firms are exporters, being able to sell more domestically is critical tospurring investment but there are no policy levers for this.

2. Policy implication: To lift the demand constraint on investment growth inTanzania’s manufacturing sector, there is no substitute to public policies that canencourage firms to export more and faster. This issue is examined next.

Page 40: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

40

PERCEPTIONS OF INVESTORS

Table: Perceptions of firms - Percentage of Enterprises rating problems as major or verysevere constraints on enterprise operations and growth in Tanzania in 2003

Results from the 2003 RPED Survey Report list the rankings that firms assigned to themain obstacles to doing business in 2003 (Table ---). As in previous years, taxes and theiradministration were in the top five constraints.20 While the corporate taxes in Tanzaniaare comparable to Kenya or Uganda, the payroll taxes in Tanzania are nearly twice ashigh, suggesting that the tax burden may indeed be a problem for manufacturing firms(Table ---). Compared to the econometric analysis in the earlier section, the top fiveRPED rankings consistent with our results apply to (i) electricity: our analysis shows thatin addition to electricity, transport, especially freight, and roads also affect investmentadversely. (ii) interest rates and access to bank finance: for those who have access to bankfinance, high interest are indeed a problem, but for the large majority, access to bankcapital is a serious constraint to expansion. It is difficult to validate the high rankassigned to corruption by firms as most did not respond to the level of bribes they pay.

20 There is no data in the survey to confirm this but a comparison with neighboring countries indicates thatat 30% or higher, taxes are among the top five constraints in Uganda and Kenya.

73

59 58 56 51 4843

31 31 27 25 25 25 24 23 2012 12

0

20

40

60

80

Tax ra

tes

Electric

ity

Cost o

f Fina

ncing

Tax ad

ministr

ation

Corrup

tion

Acces

s to f

inanc

ing

Macroe

comon

ic ins

tability

Custom

s reg

ulatio

ns

Regula

tory u

ncert

ainty

Busine

ss lic

ensin

g

Crime,

theft a

nd di

sorde

r

Skills

of work

ers

Acces

s to L

and

Anti-co

mpetiti

ve pr

actic

es

Transp

ortati

on

Lega

l sys

tem

Labo

r regu

lation

s

Teleco

mmunica

tions

Source: Tanzania Investment Climate Assessment, 2003

Page 41: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

41

Table: Tax rates applicable to manufacturing firms in Tanzania, Ghana, Kenya and Uganda.

4. EXPORTS AND GROWTH IN MANUFACTURING

Manufacturing’s contribution to Tanzania’s exports has improved since 1999. Its share innon-gold exports (that were about 52 percent of total exports in 2003) increased to 17percent in 2003 (US$ 100 million) from 6 percent in 1999, just exceeding the averageshare of 16 percent during 1991-98 (Table --). Overall, Tanzania’s non-gold21 exportperformance was dominated by agricultural exports, especially non-traditional goods likecut flowers and fish products whose share rose from 19 to 46 percent during 1999 and2003.

21 Gold and other mineral exports accounted for 48 percent of total merchandise exports in 2003. In 1991-98, their share was 8 percent of total merchandise exports.

Government Tanzania Ghana Kenya UgandaCorporate tax on profits (%) 30.0 35.0 32.5 30.0Standard applied value added tax rate (%) 10.0 10.0 16.0 17.0Payroll tax contribution by employee (%) 10.0 5.0 5.0 5.0Payroll tax contribution by employer (%) 10.0 12.5 5.0 5.0Time spent with officialsnegotiating/obtaining licenses, permitsetc. (%) 14.4 12.55 10.47 14.77Source: Africa Competitiveness Report 2000/2001, World Economic Forum

Page 42: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

42

While manufactured export growth could not have occurred without a favorable trend inthe real exchange rate in the 1990s (Figures --& --) , an equally if not more importantfactor seems to have been wide ranging reforms that generated large efficiency gains andled to accelerated production in manufacturing and boosted the competitiveness of thecountry’s manufacturing sector.

Share of total exports - gold and non-gold (%) 1985-90 1991-98 1999 2000 2001 2002 2003 1999-03 TOTAL 100 100 100 100 100 100 100 100of which gold 0 8 13 27 39 43 48 37of which non-gold exports 100 92 87 73 61 57 52 63 Share of non-gold exports 100 100 100 100 100 100 100Traditional exports Coffee 21 16 17 12 7 8 12 Cotton 19 6 8 7 6 8 7 Sisal 1 2 1 1 1 1 1 Tea 6 5 7 6 6 4 6 Tobacco 6 9 8 8 11 7 8 Cashewnuts 11 21 17 12 9 7 13 Cloves 0 4 2 3 1 2 2SUBTOTAL 64 64 60 49 40 37 49Non-Traditional exports Petroleum products 2 0 0 0 0 0 0 Manufactured goods 16 6 9 12 13 17 12 Others 19 29 31 39 48 46 39SUBTOTAL 36 36 40 51 60 63 51 Source: Economic survey, Table 18 and 19

Page 43: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

43

Tanzania: Exports and the Effective Exchange rate, 1991=2003

0.00

200.00

400.00

600.00

800.00

1000.00

1200.00

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Source : Economic Survey Tab 19 and The IMF (Eff. Exch. Rate, annual average 1990 =100)

US

$ m

illio

n

0.0

20.0

40.0

60.0

80.0

100.0

120.0

140.0

160.0

Total Exports

Effective Exchange rate

Manufactured Exports

Traditional and Non-Traditional Exports (US $ Million)

0100200300400500600700800900

1000

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

Source : Tab 19 Economic Survey

US

$ m

illio

n

Petroleum Products Minerals Manufactured Goods Other ExportsTraditional Exports

Page 44: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

44

SIZE OF THE MARKET, EXPORT ABILITY AND OPPORTUNITY FOR GROWTH

Tanzania’s low per capita income levels constrain the size of the internal market for itsmanufacturing firms. For firms who are competitive, exporting relaxes the local demandconstraint, unleashing the potential for growth. Micro evidence validates this hypothesis(Table ). Only about 18 percent of the manufacturing firms export some proportion oftheir production. Defining exporters as those who export at least 10% of their production,unravels a large part of the growth story. Median22 growth in firms that are exporters wasabout 23-24 percent in 2001-02 compared to 8 – 11 percent per annum for non-exporters.

Table : Relationship between export growth and overall growth in manufacturingNon-exporters Exporters Non-exporters Exporters

Sales growth Median Median Mean Meanin 2001 8% 24% 11% 114%in 2002 11% 19% 61% 25%Investment rate2002

1.63% 6.9% 0.15% 6.36%

Source: Firm survey data

The nexus between the manufactured supply response and manufactured exports is strongin at least one way: as a small open economy, ceteris paribus (especially the exchangerate), in the short term Tanzanian exporters can sell all that they can produce. Thisimplies that if the domestic constraints to production can be relaxed, exporting canproduce some rapid short term gains, as evident from the double digit growth experiencedin recent years (Table ---). In the medium to longer term, the ability to export will bedetermined, as in any other country, by the global competitiveness of manufacturingfirms.

Table: Performance of the manufacturing sector 1980 - 2003 1980 1985 1990 1995 1999 2000 2001 2002 2003Exports of manufactures (Cur. US$mln.) 90.60 32.80 72.60 109.20 21.50 34.20 56.20 65.90 73.00Ratio of Exports to Production (%) 11.47 9.10 20.09 31.30 3.68 5.48 8.72 9.98 10.58Growth of manu exports % -34.29 -0.91 3.13 41.82 -27.85 59.07 64.33 17.26 10.77Growth of manuf VA % 2.08 5.28 12.93 1.81 7.00 3.20 2.44 4.56 Source: Regional database(SIMA)

At least 50 percent of the exporters sold 71 percent of their output to global marketseither directly or indirectly. Of the other half, 25 percent of the firms exported 20 percentof output while the remainder exported nearly 100. Clearly, reinforcing the ability toproduce and export is key to growth in Tanzania’s manufacturing firms. What are thedeterminants of exporting and what can policy makers do about them? We turn to theseissues next.

22 Given the high degree of variability, medians are more appropriate measure than means.

Page 45: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

45

DETERMINANTS OF EXPORTING

Is exporting a function of size?While exporting is more pervasive among larger firms, forthose smaller firms who are competitive, exporting accounts for a large share of theirbusiness. In 2001-02, about 11% of small firms, 33 percent of medium sized firms and55% of large firms were exporters. The amount of total production exported was notrelated to firm size, suggesting that if a firm is globally competitive, it can export most ofwhat it produces. Smaller firms exported more than medium sized firms. The medianproportion exported was about 56% in small firms and 82-85 percent in large firms.Among medium sized firms, the proportion exported tumbled from 48% in 2001 to 30 %in 2001.

Table : Share of firms that are exporters, by sector

Does location or sector matter? In general, firm growth is not affected by whether thefirm is located in Dar el Salaam or some other part of the country. However, forexporters, locations outside Dar el Salaam seem more conducive to higher growth. 73percent of all exporters are located outside Dar. Perhaps locations outside Dar play afacilitating role in providing easy access to local inputs such as agricultural produce forTanzania’s fastest growing export industry, i.e. agro-industry with the largest proportionof exporters (Table ...).

Does access to finance matter? In most exporting firms, about 61% of the equity isprivate and domestically owned; about 31% is foreign owned, and 3 % is governmentowned compared to non-exporters among whom the share of domestic private equity is75% and foreign equity is 16%. Among exporters, own resources and equity capital werethe main sources of start-up compared to non-exporters who had twice the access to bankcapital for start-up (11%). However, exporters enjoy access to commercial loans through

% of firmsthatexported inpast 5years

Agro-processing 46.2Chemicals and Paints 10.3Construction materials 1.3Metals 7.7Furniture/Wood 12.8Paper, Printing, Publishing 3.8Plastics 3.8Textile/Garment and Leather prod. 14.1Total 100.0

Page 46: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

46

overdraft privileges (49% compared to only 28% of the non-exporters); lower interestrates, and greater access to bank financed working capital (25% compared to only 10% ofthe non-exporters).

Demand for skilled labor matters. In contrast to non-exporters, exporting firms demandhigher education levels (graduate, technical and vocational) and higher skills. Forexample the skilled to unskilled worker ratio in exporting firms is 3.8 compared to 2.3 innon-exporters. Exporters also pay a premium for higher skills: for example, the averagepremium for managers is about 20%, for professionals about 37%, for technically skilledworkers about 19%. Unskilled workers in exporting firms, on average, earn 6% less whilenon-production workers earn the same. More exporting firms (71%) invest in formaltraining than non-exporters (47%). A larger proportion of managers in firms that exportare foreign, with more experience, especially in the export business. The proportion ofworkers with computer skills is on average, five times higher than in non-exporters (only2% of the workforce can use computers relative to 10% in exporters).

Infrastructure matters.

Table: Status of infrastructure in manufacturing firms

% of

exporters% of non-exporters

Location in industrial estate 66.67 60.56Firm has its own generator 82.69 47.22Firm has its own water well 54.90 30.00Firm has installed its own water infrastructure 49.02 27.75Firm shares communal water source 35.42 6.83Firm has invested in own road 28.85 6.02Firm provided worker transport 38.46 22.22Firm has invested in own freight transport 34.62 14.35Firm has invested in own waste disposal facility 55.77 36.11firm has invested in other infrastructure 5.77 3.7Firm has internet connectivity 80.77 42.13

Page 47: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

47

Table --- provides a clear indication of why the production response of exporters,measured by either sales or investment growth, is many-fold that of non-exporters. Inspite of overly long waiting times for clearances and registration permits, and theassociated official and unofficial costs, exporters have a clear comparative advantage – inthe absence of publicly provided infrastructure and basic services, they have moreresources to finance their own infrastructure - they simply have more and betterinfrastructure to work with. For most of the key types of infrastructure, there are at leasttwice as many if not more exporting firms than there are non-exporters.

Page 48: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

48

The amount exported to each global destination matters. Two main regional marketsaccount for about 70% of Tanzanian exports. Western Europe and the regional Africanmarket, dominated by Kenya and Uganda absorb equal 35% (mean) of the total exportvalue (table ---). A fairly large number of exporters are involved in the process - about65 - 67% of the firms sell nearly 35% of the exports to these two main destinations,suggesting that at least of Tanzanian exporters are globally as competitive in WesternEurope as they are in the African market where competition from low income Asianexporters such as China is rising. Considering that there were only about 52 firms in thesample that exported more than 10% of their output, it is encouraging that exports todeveloped country markets are not dominated by one or two large firms, as is often thecase in many low income countries. Clearly, if the critical inputs can be put in place,there is scope for more entrants in the export of manufactures.

Table: Main destinations and share of output exported to each

% Valueexported toeach region

Mean Median

No. of firmsthat export to

thisdestination

% of firmsthat export to

thisdestination

West Europe 34.77 10 30 66.7East Europe & Central Asia 3.88 0 7 15.6North America (USA & Canada) 5.90 0 8 17.8North Africa/ Middle East 7.50 0 12 26.7Other Asia 9.10 0 13 28.9Kenya 17.90 0 29 64.4Uganda 3.08 0 12 26.7Other Africa 15.44 0 25 55.6Others 2.33 0 6 13.3

Source: Firm Survey, 2003

Page 49: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

49

Which government programs promote export growth? Table --- shows that ofapproximately 75 firms that export a positive amount of their output, less than 25% eachuse the 13 or so use the export incentives/programs. By far the most popular programsare retention of export proceeds in a foreign country (typically the export destination) andthe bonded warehouse or similar scheme. In order of importance, in the second rank arethe customs duty drawback, profit tax exemption and duty certificates used by 14 – 16percent of the exporters. Of the remaining, the export development fund (EDF), exportprocessing zone (EPZ), export adjustment fund scheme and foreign inputs facility areclear cases for rationalization or closure in their present form, given their low utilizationrates.

Table: Usefulness of existing export promotion programs

No. of firms thatresponded

No. offirms thatuse theincentive

% fo firms thatuse the incentive

Manufacture-in-Bond Scheme 73 5 6.8Customs Duty Drawback 73 12 16.4Duty suspension on imported inputs 75 4 5.3Bonded warehouse or similar scheme 75 16 21.3Profit tax exemption 73 10 13.7Export Credit Guarantee e.g. Nexim 74 4 5.4Export Development Fund (EDF) 74 2 2.7Retention of export proceeds in foreign country 73 18 24.7Export Processing Zone 74 2 2.7Export Adjustment Fund Scheme 73 0 0.0Foreign Inputs Facility 75 2 2.7Duty Certificates 74 11 15.0Other 57 3 5.3Source: Firm Survey 2003

Table ---- presents the key determinants of manufactured exports defined by theproportion of output exported by firms. Relative to garments and textiles, firms in agroprocessing contributed positively to exports but those in chemicals, furniture and wood,and paper and printing each reduced the proportion exported by almost equal amounts.Large firms, especially those owned by individuals whose nationality was Kenyan werethe main drivers of export growth, while firms with Ugandan owners performed poorly,reducing exports. In terms of ethnicity, firms with owners of African origin contributednegatively to growth. The relationship between export growth and firms with Asianowners was not significant. These results reflect the entrepreneurial skills and exportexperience of firm owners and make sense – unlike most sub-Saharan African countries,Kenya has the longest tradition of private enterprise; unfortunately, in the past decades,an adverse domestic climate has encouraged most of its entrepreneurs to cross the border

Page 50: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

50

and operate in neighboring countries such as Tanzania where the investment climate ispresently more conducive to exporting. In addition to the owners experience, firms asorganizations that have benefited from prior learning to export to global markets in thepast 5 years are able to export more. Similarly, given the complexity of exporting toincreasingly challenging global markets, new entrants into the Tanzanian market (but notnecessarily new to the global market) also seem to export more. They probably comeequipped with the knowledge that exporters need, and do not need much lead time tolearn the skills.

Page 51: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

51

TABLE: DETERMINANTS OF EXPORT GROWTH – IDENTIFYING THE BINDING CONSTRAINTS.

Coeff. Sig. LevelIntercept 7.83 **Age -0.19 ***Sector - agro proc 5.16 **Sector - chemicals -6.13 *Sector - furniture/wood -4.73 *Sector - paper/printing -7.35 **Size - large 10.63 ***Owner - Kenyan 26.8 ***Owner - Ugandan 12.98 *Ethnic origin - Asian -3.32 *Ethnic origin - African -5.16 **Machinery's age < 5 yrs. 0.03 % of workers that can use computers 0.22 ***Share of skilled workers in workforce -0.05 *Share of non-prod. workers in workforce 0.07 *share of workers with grad.&post-grad.edu. 0.22 ***Years when firm started exporting 0.76 ***Exporting experience in past 5 years? 7.15 **Bank loan financed start-up costs -0.09 **Access to bank financed working capital 0.06 *Access to bank financed investment capital -0.05 *Time taken to clear customs at border -0.26 ***Share of exports to specific destination Share of output exported to Western Europe 0.36 ***Share of output exported to Eastern Europe& Central Asia 1.03 ***Share of output exported to the U.S. 0.48 ***Share of output exported to other Asian countries 0.41 ***Share of output exported to Africa -0.11 **Trade promotion programs Manufacture-in-Bond Scheme -38.06 ***Duty suspension on imported inputs -32.66 ***Bonded warehouse or similar scheme 12.75 **Export Credit Guarantee e.g. Nexim -3.37Export Development Fund (EDF) -51.3 ***Retention of export proceeds in foreign country 14.77 ***Export Processing Zone 33.31 ***Duty Certificates 9.38 **

Number of observations 240R-squared 0.71

*** = sig, level 1 - 5%; ** = sig. Level 6 - 15%; * = sig. Level 16 - 30 %;

Page 52: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

52

The availability of workers with graduate and post graduate education and computerliteracy enhances exports. This is natural as more manufacturing processes, especially inheavy industry such as chemicals, paper, metals, machinery etc are becomingincreasingly computerized. However, the premium on skilled wages given the paucity ofskills in Tanzania imposes additional costs on exporting firms relative to non-exporters,and reduces export growth.

While Tanzania’s exporters have greater and easier access to the banking system, withthe exception of working capital that promotes exports, bank-financed investment andstart-up capital are costly and dampen export growth.

For technical reasons, infrastructure related variables were excluded from the exportgrowth model but Table .... provides a clear guide to how critical this factor is inreleasing the constraints to export growth in Tanzania.

Among the bureaucratic hurdles, the time taken to clear customs is an unambiguousdeterrent to export growth. There seems to be discrimination against exporters who haveto, on average, wait as long as 20 days compared to only 7 days for non-exporters toobtain customs clearance for exports and imports.

Exports destined to SADC or the local regional markets in Kenya and Uganda do notgrow as fast as those outside of Africa, especially to Western Europe, Eastern Europe, theUS, and other Asian countries where exporters can sell larger amounts without runninginto the market size constraints that characterize low income African markets, and wherecompetition from other regional African exporters selling similar products is high.

Among the export promotion programs that help to raise exports are the bondedwarehouse scheme, policies that permit the retention of foreign exchange earnings inanother country, EPZs and duty certificates. While the others make sense, given theirusage among exporters, the significance of EPZ presently used by only about 3% of firmsis curious. Further investigation into what is special about the EPZs in Tanzania is neededto inform this issue.

EXPORT GROWTH AND COMPETITIVENESS – WHAT DOES IT MEAN AND WHAT WILL IT TAKETO ACCELERATE EXPORT GROWTH?

To expand exports, Tanzania’s incumbent exporters in manufacturing need to exportlarger amounts, and more manufacturing firms need to start exporting. Compared to itsneighbors in Sub-Saharan Africa and other global competitors who are low wage – highskills producers competing with Tanzanian firms in most markets, there is sufficientroom for expansion, especially as the size and purchasing power in the domestic marketremain limited. Table C1 and figure F1 provide international comparisons to support thiscase. Not only do its Asian competitors such as India, Bangladesh and China export asignificantly larger proportion of the manufactured output in spite of much larger

Page 53: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

53

domestic markets with higher purchasing power, the share of manufacturing in theireconomies is also significantly larger and has grown over time.

Figure F1: Percentage of firms exporting at least 20 percent of production

0

10

20

30

40

50

60

Nigeria

Ethiop

ia

Bolivia

Moz

ambiq

ue

Eritrea

Ugand

a

Tanza

niaIn

dia

Nicarag

ua

Kenya

Zambia

Seneg

al

China

Bangla

desh

Mor

occo%

fir

ms

expo

rtin

g >

20%

of

prod

uctio

n

Source: Figure 2 in Eifert, Gelb and Ramachandran, 2005.

Table C1: Selected Economic Indicators, 2000–2002

Country

GNI per

capita, $Trade%GDP

Ag

%GDP

Investment (FDI),

%GDPManufacturing %GDP

(growth)

Manufacturing,% merchandise

exports

Eritrea 160 111 21 39 (5.3) 8 (5.4) -

Ethiopia 100 49 52 18 (1.2) 7 (5.0) 9.8

Nigeria 290 81 35 20 (2.4) 4 (3.7) 0.2

Kenya 360 57 19 14 (0.4) 13 (1.0) 22

Mozambique 210 79 23 40 (8.6) 13 (9.2) 7.5

Senegal 480 38 18 18 (1.3) 13 (7.3) 37

Tanzania 280 71 45 17 (3.7) 8 (5.9) 18

Uganda 250 40 31 20 (2.6) 10 (2.9) 6.5

Zambia 330 75 22 18 (2.9) 11 (4.5) 17

China 940 52 15 37 (3.7) 38 (8.7) 88

Bangladesh 380 33 23 23 (0.3) 16 (5.6) 92

India 480 31 23 22 (0.6) 15 (5.6) 77

Algeria 1,720 61 10 23 (1.6) 8 (-1.0) 2.3

Morocco 1,190 66 16 25 (4.2) 17 (4.0) 64

Bolivia 900 49 15 16 (9.3) 15 (1.9) 17*

Nicaragua 720 73 18 29 (5.0) 14 (1.2) 13

*was 40% in 1999, but oil & gas production has reduced it rapidly since

Page 54: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

54

Source: Table 4 of Eifert, Gelb and Ramachandran, 2005.

Page 55: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

55

In addition to the determinants of exports as shown in table ---, higher export growth inany country today is critically dependent upon the global competitiveness of the exportsector. In the medium to longer term, the ability to simply produce manufactured outputthat is not absorbed by the domestic demand does not necessarily imply that it can be soldin the global market place. This is particularly true of countries like Tanzania who arerelatively small players in an increasingly complex global marketplace with no obviousor established edge in any manufacturing activity.

Several factors define global competitiveness today.

1. To accelerate exports of manufactures, Tanzanian firms need to remaincompetitive with not just their regional counterparts (Kenya, Uganda or SADC)but especially the low income exporters in these regional markets. This issue issensitive to the lower purchasing power of the Sub-Saharan African market that ismore amenable to high skill-low wage producers like China who benefit fromhigher investment inflows attracted by higher labor productivity. The issueapplies equally to global markets in Europe and Asia which are destinations forTanzanian exports.

Table C2: Net value-added per worker and capital per worker, median by firm size

VA / L, $ K / L, $ country Micro small medium large very large Micro small medium large very largeEritrea 2,900 2,450 5,450 2,000 1,600 14,750 17,700 52,050 52,650 14,500Ethiopia 600 550 750 1,050 650 950 2,450 3,750 4,600 4,400Nigeria .. 1,400 1,500 2,850 3,100 .. 17,200 12,850 24,900 19,150Kenya .. 1,750 3,100 6,300 2,300 12,600 6,800 11,700 10,000 6,800Mozambique 350 1,250 2,800 2,200 .. 2,700 6,200 5,600 12,250 ..Senegal 6,150 7,500 17,100 15,600 14,500 9,450 6,900 11,300 11,950 1,000Tanzania 1,350 1,850 4,200 3,400 6,800 1,050 5,900 4,750 13,250 13,150Uganda 925 1,000 1,600 4,800 950 800 1,550 4,700 8,850 1,050Zambia .. 800 950 1,250 2,500 .. 9,650 14,000 6,700 13,750Bangladesh 950 1,300 1,650 1,200 1,150 400 1,450 1,650 800 1,150India 1,850 3,500 3,200 3,800 5,750 1,500 1,500 1,850 2,850 6,000Bolivia 1,300 1,700 3,050 2,450 7,350 1,200 2,150 4,700 5,350 11,050Nicaragua 2,100 2,450 3,650 10,200 3,700 950 1,300 1,250 5,700 1,350China 1,850 2,350 4,150 3,850 4,250 1,400 850 1,250 1,400 4,200

Morocco 4,100 4,850 3,600 5,250 5,750 4,350 2,150 2,550Source: Table 7 of Eifert, Gelb and Ramachandran, 2005.

It would be myopic to compete with neighboring Kenya or Uganda when thecompetitor may be China in the home and neighboring markets. In a nutshell, toattract more investment for higher export growth, Tanzanian firms need to raisethe productivity of the labor force and the returns to capital (profitability) toglobal levels in the exporting sector/industry. Table C2 shows that while thevalue added per worker in dollar terms in large and very large firms in Tanzania is

Page 56: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

56

competitive with its Asian counterparts, the capital intensity of production issignificantly higher in Tanzania. For each unit of labor, investors in Tanzaniamust invest more than 6 times as much as they invest in India or China and nearly12 times what they would invest in Bangladesh. The labor intensity of productionin Asia is clearly many times higher due to higher skills and increased production.Besides higher employment creation, this also raises the returns to capital (seeFigure F2. For global competitiveness, firms need both higher labor value addedand higher returns to capital. The latter is zero in Tanzania relative to returns of 5– 9 percent in Kenya and Uganda, and 50 percent or more in Asia which isattracting investors and exporting significantly more manufactures.

Figure F2: Median profit margin & median return on capital

-0.20

-0.10

0.00

0.10

0.20

0.30

0.40

0.50

0.60

Eritrea

Ethiop

ia

Nigeria

Zambia

Tanza

nia

Moz

ambiq

ue

Ugand

a

Kenya

Seneg

al

Mor

occo

Bolivia

Nicarag

ua

Bangla

desh

China

India

profit margin return on capital

*profit margin = (profits / sales – 1); return on capital = (profits / capital – 1)SOURCE: FIGURE 3 IN EIFERT, GELB AND RAMACHANDRAN, 2005

2. Tanzania’s manufactured exports fall into three categories: heavy industry(chemical, paints, metal products, paper, plastics etc.) based on its rich mineralsbase; agro-processing based on its rich natural resource base, and lightmanufactures (textiles and garments, furniture and wood products) that seem tobe declining because of competition from other lower wage producers, especiallyafter the expiration of the MFA. Firms producing these three categories ofmanufactures need different strategies to remain globally competitive and expandexports. As heavy industry relies on locally available natural resources, publicpolicies to support its expansion and raise productivity should be of high priority.Agro-processing is also natural resource based but of a different type. The

Page 57: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

57

processing of local non-traditional produce such as fish, fruits and vegetables oreven cashew is relatively new to Tanzanian firms and must comply with complexphyto-sanitary standards to gain entry in developed country markets. Whiledifficult, these exports also provide new opportunities for Tanzanian firms as theyare unlikely to run into market size constraints in developed countries. The scopefor growth is also enhanced by prospects of moving up the value chain. Forexample, with better technologies and higher technological capability, fishprocessing firms could graduate from simple fillets to higher value semi-preparedfish foods for the EU, presently their single largest destination for fish products.and quality and quantity of infrastructure as almost all other intermediate inputscan be imported. Expansion of light manufactures is more difficult as it entailscompetition with large MNC-managed global supply chains with a high level oftechnological sophistication and keen cost competition from low-wage – hghskills producers. Without entry into a global segments of these chains, there islittle scope to expand exports for small producers like Tanzania. All threestrategies require technological upgrading to raise productivity, attract newinvestment, reduce costs and expand exports.

The bottom line on competitiveness implies improving labor productivity throughtechnical skills and technological upgrading to achieve and continually maintain globalstandards. Two types of strategies are needed to achieve this.

1. First, public policies that (1) facilitate cost reductions by minimizing wastage –examples are the provision of more and higher quality infrastructure (reducingwaste due to poor freight transport; improve capacity utilization through regularpower supply, more roads etc.) so firms can re-allocate private resources towardsinvestment in firm expansion instead of the private provision of public goods. (2)public policies that enable the private sector to upgrade equipment, acquire bettertechnologies, improve firm level learning and human capital accumulation, andreduce the entry costs of exporting.

Figure F3 displays the business environment related losses in Tanzania asmeasured by power outages, crime, shipment losses and delivery delays are about4% of sales relative to less than 2 percent in Uganda and even Ethiopia and under2% in Tanzania’s Asian counterparts, India, china and Bangladesh. Clearly, thereis plenty of room for public policies that can contain wastage and reduce costs, asevident from the empirical analysis and these international comparisons. Anotherexample of this is evident from Figure F4 that illustrates that the indirect costs ofproduction, associated with private provision of infrastructure or poorly providedpublic infrastructure are almost 25% in Tanzania relative to 7 –10 % in Asia, and18% in Uganda. Table C3 underscores the same point using the example ofprivately financed generators that crowd out private investment in firm expansionwhen firms need to substitute for public goods such as power, a critical input ofproduction. China emerges as a unique example of a developing country thatuntil two decades ago was low income but today fares remarkably in powerprovision compared to even India and other sub-Saharan countries. This is part of

Page 58: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

58

the explanation for China’s rapid growth miracle that today challenges every lowincome exporter.

Page 59: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

59

Figure F3: Business Environment -related losses, % of sales, average of 5th to 95th percentiles ofsample

0

1

2

3

4

5

6

China

India

- Hyd

eraba

d

Bangla

desh

Nicarag

ua

Seneg

al

Ethiop

ia

Moz

ambiq

ue

Eritrea

Ugand

a

India

- Calc

uta

Nigeria

Tanza

nia

Zambia

Kenya

loss

es, %

of

sale

s

power outages crime, shipment losses, delivery delays

Source: Figure 4 in Eifert, Gelb and Ramachandran, 2005

Figure F4: Cost structures: % of total costs, average

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

china

nicaragua

bangladesh

senegal

india

morocco

bolivia

uganda

nigeria

ethiopia

kenya

tanzania

eritrea

zambia

mozambique

indirect labor capital inputs

Source: Figure 5 in Eifert, Gelb and Ramachandran, 2005

Page 60: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

60

Table C3: Share of Firms Owning Generator by Size

Country Micro Small / Medium Large / Very LargeBangladesh 0.28 0.53 0.88

Bolivia 0.13 0.11 0.31China 0.0 0.14 0.38Eritrea 0.38 0.43 0.63

Ethiopia 0.03 0.23 0.43India 0.23 0.76 0.91

Kenya 0.46 0.67 0.89Morocco 0.14 0.15 0.28

Mozambique 0.20 0.23 0.63Nicaragua 0.06 0.29 0.81

Nigeria 0.83 0.96 0.99Senegal 0.23 0.19 0.16Tanzania 0.18 0.60 0.89Uganda 0.04 0.44 0.87Zambia 0.30 0.28 0.61

Source: Table 12 of Eifert, Gelb and Ramachandran, 2005.

Table C4: Share of Firms with Access to Loans/Overdrafts by Size

Micro Small / Medium Large / Very LargeBangladesh 0.36 0.74 0.81Bolivia 0.75 0.76 0.84China 0.19 0.70 0.84Eritrea 0.25 0.62 0.75Ethiopia 0.29 0.66 0.88India 0.42 0.72 0.88Kenya 0.17 0.51 0.25Mozambique 0.32 0.36 0.46Nicaragua 0.53 0.67 0.72Nigeria 0.33 0.64 0.89Senegal 0.1 0.42 0.67Tanzania 0.08 0.37 0.66Uganda 0.07 0.35 0.72Zambia 0.1 0.54 0.80

Source: Table 13 of Eifert, Gelb and Ramachandran, 2005.

Page 61: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

61

Table C4 indicates an example of the facilitative role of the state in Tanzaniarelative to its global competitors in the arena of the financial sector. Compared toTanzania, firms across almost all size classes - micro, SMME and large firms - inAsia and most other sub-Saharan African countries had greater access to bankfinance, and over draft facilities for working and investment capital purposes.

2. The second type of strategy entails keeping up with external and globaldevelopments that define the size of the global market for a country’s exports.This requires a dynamic approach and partnerships with global firms throughincentives. Maintaining the competitive edge in even the packaging of cutflowers, vanilla or fish processing today is an externally driven challenge that ismost efficiently achieved with external players as is happening in the cut flowerindustry where European firms supply the latest technologies to flower growers inTanzania and Kenya. For example, moving up the agro processing value chain inthese industries cannot be achieved by existing domestic export promotionprograms or policies to attract FDI alone. More creative and comprehensiveapproaches are warranted.

5. EMPLOYMENT

As Table – indicates, in 2001, manufacturing share of total employment in Tanzania wasa meager 1.5 percent (245,000 jobs).

Table : Distribution of the Currently Employed Labor Forceby Industry (Standard Definition) 2000/01

Main Industry Number

Sex ManufacturingTotal for allindustries

Male 161,699 8,351,291Female 83,750 8,563,513Total 245,449 16,914,804

PercentageMale 1.9 100.0Female 1.0 100.0Total 1.5 100.0Source: Intergrated Labour Force Survey, 2000/01 - Table 3.4

Page 62: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

62

Manufacturing’s contribution to production-related jobs, i.e., craft and related workersand plant and machine operators in the country was significant – it accounted for a thirdof total jobs in each category. It also employed 10 percent of Tanzania clerical workersand 10 percent of the professional staff (mangers, professionals, technical staff ). Incontrast to agriculture, manufacturing created only 2 percent of the country’s elementaryjobs, clearly establishing that policies that are conducive to growth in manufacturing alsofavor the creation of better paying jobs (Table --). The policy challenge is of course morecomplex: in addition to investment increasing policies, policy makers must also addressthe creation of skills that can be employed by firms.

Table: Distribution of the Currently Employed population by Main Occupation (Standard Definition) 2000/01

Main Industry Percentage

Occupation ManufacturingTotal for allindustries

Legislators, Administrators &Managers 2.9 100.0Professionals 4.0 100.0Technicians & AssociateProfessionals 2.7 100.0Clerks 10.8 100.0Service & Shop Workers 1.2 100.0Agriculture & Fisheries Workers 0.0 100.0Craft & Related Workers 33.8 100.0Plant & Machine Operators &Assemblers 32.1 100.0Elementary Occupations 2.1 100.0Total 1.5 100.0Source: Intergrated Labour Force Survey, 2000/01 - Table 3.5

Survey data shows that during 2000-01 and 2001-02, median employment growth waszero; the mean was about 7 percent. Firms in the bottom 25% experienced significantlynegative growth rates, while the opposite was true for firms in the top 25 percent of thedistribution. On average, these offsetting effects yielded zero growth. Towns withpopulations of less than 50,000 experienced negative rates of about 2 percent. Amongsectors, plastics was the only one with positive average rates of 5 percent.

Employment growth among exporters was no different from that among non-exporters.But exporters employed workers with higher education, especially those with diplomas,graduate, post graduate and technical degrees and paid premia for higher skills ranging

Page 63: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

63

from 19% for skilled employees to 37 percent for professionals. They also employedmore female workers (12 percent of firms) compared to non-exporters (5%), and investedmore in formal worker training (70%) compared to only 47% among non-exporters.

DETERMINANTS OF EMPLOYMENT GROWTH

Table --- illustrates the determinants of employment growth in the manufacturing sector.Relative to textiles and garments, firms in metals and plastics create more jobs comparedto agro processing or chemicals. Relative to smaller towns, firms located in cities with apopulation of over 1 million help to raise employment.

Table : Determinants of employment growth in manufacturing firms

Coeff.Sig.

LevelIntercept 27.6 *Age 0.21 **Sector - agro proc -7.9 **Sector - chemicals -9.5 **Sector - Metals 17.4 ***Sector - plastics 25.7 ***Location in large city(pop. Over 1 ml.) 5.14 *Age of male workers < 30 yrs. 0.2 ***Age of female workers < 30 yrs. -0.08 **Worker with grad. & post-grad educ 0.2 *Managers education 2.39 ***Log of skilled wages -3.96 ***% of exports in total output -0.12 **Overall sales growth 0.15 ***

Number of observations 158R-squared 0.24

*** = sig, level 1 - 5%; ** = sig. Level 6 - 15%; * = sig. Level16 - 30 %;

Workers characteristics or the supply side of the labor market also affects job creation.The availability of younger males aged 30 or less contributes to job creation but the samecharacteristics in females has a negative effect. Higher education, especially at thegraduate or post graduate level, i.e. technical level, as well as higher education amongmanagers both affect employment positively. But the wage levels of skilled Tanzanianworkers are a deterrent to employment growth. Labor demand is significantly andpositively driven by overall sales growth but is negatively affected by the proportion ofoutput that is exported. A possible reason for this odd result is that exports levels andemployment growth were negatively related in 2000 – 02.

Page 64: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

64

Policy issue: Export growth has a positive and significant effect on overall firmgrowth and investment, but a negative, albeit weak, effect on employment growthin Tanzania’s manufacturing sector. This suggests that exporters in Tanzania areinvesting in less labor-intensive activities. The policy implications are clear:labor productivity at prevailing wage levels is too low to attract investors in labor-intensive export sectors. What can policy makers do to raise labor productivity?Improve the supply and quality of technical skills, adopt policies that enable firmsto acquire better technologies that raise labor productivity and promote firmlearning.

6. REFORM PRIORTIES FOR ACCELERATING GROWTH

In the first section, we established from existing official data that while salesmanufacturing sector has grown in the impressive range of 6 – 8 percent per annum since1999, this trend has not been matched by corresponding growth in investment inmanufacturing firms, either from macro or firm level data. In fact, investment growth wasnearly nil during 2000-02. In later sections, we found that employment growth inmanufacturing was also nil for the same period. Similarly, manufacturing exports grewat double digit rates, but unlike non-exporters, exporting firms also experienced higherinvestment growth. However, while manufactured exports contributed to overalleconomic growth, they did not contribute to employment growth. Three hypotheses seemare validated by our analysis:

1. recent growth in Tanzania’s manufacturing sector seems more of an outcome ofefficiency improving reforms in the late 1990s than an outcome of sustainedinvestment in manufacturing in the recent past. However, in contrast to the pastthree decades since the 1960s, there was a link, albeit weak, between growth inmanufacturing and investment in the 2000-02 period. This suggests that unlessinvestment is somehow boosted, these impressive growth rates will not besustained in the medium term. The link suggests that growth in investment, ceterisparibus, will support overall growth.

2. recent growth in manufacturing is strongly related to the export growth ofmanufactures – exporters experienced significantly higher growth rates than non-exporters. Policies that promote exports are good for growth.

3. whatever is good for growth is not good for employment in 80 percent of themanufacturing firms selling mainly in the domestic market. They did not add toemployment growth. Among exporters, investment growth is higher, but is notbeing channeled into labor-intensive production. This implies that capital asembodied in machinery and equipment is preferable to labor skills. The bottomline: machines raise profits more than workers. Labor productivity needs to beraised to establish or re-establish (?) the link between growth and employmentacross exporters and non-exporters.

Page 65: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

65

Our analysis in sections 3 – 5 shed light on the determinants of growth: investment andgrowth. In-depth analysis of each of these key determinants established the top policypriorities that can propel investment and export growth. International comparisons werealso provided to reinforce the hypotheses. We use some innovative techniques toprioritize the main areas for reform, given that policymakers are already overburdenedwith a laundry list of things to do. In section 6, we examined the determinants ofemployment to understand what drives it and explore why it is disconnected from growthand investment. Each section flagged key policy issues. In every section, two commonthemes emerged. There was no running away from the same recurring reality:

• for faster growth, more Tanzanian manufacturing firms will have to startexporting outside Africa where demand is high. For this, policymakers need to

1. facilitate greater investment, and2. use some innovative interventions that can raise productivity, improve the

global competitiveness of all firms, and hook them to global supply chains.

Is it critical to appreciate that while removing the obstacles to investment is necessary,this in itself does not guarantee that Tanzanian firms, especially indigenously owned,will become globally competitive and start exporting significantly more. Catching-upwith global markets is beset with complex public goods issues that cannot be achieved bythe private sector alone – the enabling, development-hand of the public sector is neededto accomplish public-private partnerships to achieve this. A discrete and new strategy isneeded to address this in addition to (i). This approach needs to be carefully orchestratedto include all firms, not just foreign owned ones, as it is believed that potential investorsare deterred by occasionally voiced views that companies should remain Tanzania-owned. International experience shows that technological catch-up necessary for rapidexport growth in manufacturing cannot ride on the shoulders of foreign capital alone.

Policy Recommendations

1. Policies that can facilitate investment growth are well known to Tanzanian policymakers who have been pursuing them for a while now. Our analysis helps to identify thebinding constraints in infrastructure, financial sector, human capital, businessenvironment etc. to enable even more focused interventions given the paucity of time,financial resources and capacity. These are listed at the end of each discussion in sections3 – 5. They are reviewed briefly below in the context of ongoing reforms to the extent weare aware of them. ……..to be repeated here.

Page 66: TANZANIA: GROWTH , EXPORTS AND …documents.worldbank.org/curated/en/852161468760788271/...TANZANIA: GROWTH , EXPORTS AND EMPLOYMENT IN THE MANUFACTURING SECTOR Vandana Chandra, Pooja

66

2. The use of innovative interventions to facilitate an increase in labor productivity,improve global competitiveness and faster export growth of manufactures is anapproach that is being recommended as a discrete new strategy. This could beadopted using clever policies and programs to put the private sector in the driver’s seat byenabling firms to participate in productivity-enhancing activities such as rapid technicaltraining of the workforce as opposed to waiting for the schooling system to churn out ameager number of skills, adapt and adopt newer technologies that can improve valueadded, scale up production (for example, from one shift a day to three or two) and mostof all, enable firms to catch-up to global standards, especially those required to connect toglobal supply chains in light and heavy manufactures, and comply with internationalstandards in non-traditional sectors such a agro-processing in which Tanzania has a clearcomparative advantage. The challenge is one of learning how to scale up nascent ornewly emerging sectors for export markets. International experience shows that this wasbest done through constructive public-private partnerships using the measuring rod ofexport growth to pre-empt the hazards of rent-seeking or putting the public sector backinto what private firms are doing; production. For more on this strategy, policy makersmay wish to refer to some recent research emerging on “The How to of technologicalchange for faster growth,” (Draft, World Bank, 2005).