Religiosity and threshold effect in social and financial performance of microfinance institutions

 Mohammad Ashraful Mobin

What is the Issue ?

• Effect of religion on Microfinance Institutions’ Performance.

• Testing the paradox of the microfinance hypothesis.

WHY It is an Issue ?

World Poverty Map

Prevailing Religion

WHY it is an Issue ?

• Several theories and studies argue that religious beliefs represent an important yet neglected factor in development.

• 16.1% of worldwide population lives with less than one US dollar a day and 57.6% with less than 2.5 USD.

• 44% of this extreme poor people is concentrated in the Muslim countries.

• About 65 million people around the world received small loans without collateral through microfinance.

WHY it is an Issue ? • 72% of the people in Muslim countries do not use formal financial services.

• 20%-40% people do not use conventional microfinance to avoid interest.

• Currently more than 300 Islamic Microfinance Institutions are working to alleviate the poverty.

• Little academic inquiry on the relationship between Islamic based institutions and development. Moreover, there is even less in the area of Islamic microfinance and development.

What are the theories supporting this issue ?

• Neo-classical

The Neo-classical growth model emphasizes the importance of savings in order for a country's economy to grow. According to Todaro et al. (2003), one of the main constraints for poor households in developing countries is the lack of access to financial services. ( Financial Inclusion )

• Welfarist theory

The Welfarist theory focuses on reducing poverty through credit, often provided in conjunction with services such as skills training and teaching of literacy and numeracy, health, nutrition and family planning.

What are the theories supporting this issue ?

• Micro-Finance Unity Theory

This theory advocates for joint liability in the repayment of microfinance loans. The argument fronted by this theory is that joint liability could improve repayment rates and the welfare of credit-constrained borrowers.

• Empowerment Theory of Micro-finance

The empowerment theory shows how microfinance loans can empower people to improve their standard of livings by enabling them to start businesses. The theory points out that women account for nearly74 percent of the 19.3 million of the world’s poorest people now being served by microfinance institutions.

 What is the meaning of Religion , Development and Microfinance ?

Religion and development:

• The notions of development and religion have so much in common

• Hindu idea emphasizes harmony with the living environment.

• Muslims believe that the ultimate aim of life is to return humanity to its creator in its original state of purity.

• Christians believe that personal transformation – inner change – is the key to the transformation of society.

What is the meaning of Religion , Development and Microfinance ?

Microfinance:

• To provide credit to working poor who otherwise would not have access to credit services.

• MFI created for poverty alleviation and empowerment of the poor.

• Relying on traditional skills, poor people use financial services to start or expand self-supporting businesses.

What are the variables ?

Dependent Variable • OSS ( source: microfinance consensus guidelines; CGAP, World Bank)

Operational self-sufficiency (oss) = Operating revenue /(Financial expense + Loan-loss provision expense + Operating expenses

Independent Variables

• Total Equity • Average loan Size • Female clients • Portfolio yield• Loan loss reserve Ratio• Cost Per Borrower  

Title 13

Why This Model ?

• Preferred Model by CGAP, World Bank

• Used by most of the microfinance rigorous study

• Research Agenda from Previous Study

Reference :

Mersland, Roy, Bert D’espallier, and Magne Supphellen. "The effects of religion on development efforts: Evidence from the microfinance industry and a research agenda." World Development 41 (2013): 145-156.

Title 14

Why System GMM ? Why not Difference GMM

One potential problem with the ‘differenced’ equation is that the lagged ‘levels’ of regressors may be weak instruments for the ‘differenced’ variables.

• The problems can be tackled by the ‘System GMM’ proposed by Blundell and Bond (1998). This estimator combines within a system the regression in first differences with the regression in levels.

• The ‘first differenced’ moment conditions in Standard GMM are augmented by ‘level’ moment conditions in System GMM for more efficiency in estimation(Blundell and Bond, 1998).

Title 15

What is the finding : Difference GMM

* p<0.05, ** p<0.01, *** p<0.001t statistics in parentheses Observations 49 34 49 34 (.) (.) (.) (.) Constant 0 0 0 0

(.) (.) (.) (.) Dummy Islamic 0 0 0 0

(-0.99) (-1.93) (-1.05) (-1.86) lequity -8.468 -25.83 -8.468 -25.83

(0.70) (1.10) (0.99) (1.99) lloans 6.082 11.69 6.082 11.69*

(-2.72) (-2.45) (-2.87) (-2.73) lloanloss -11.88** -12.64* -11.88** -12.64**

(0.01) (1.26) (0.02) (1.43) lloansize 0.140 21.50 0.140 21.50

(-0.37) (-0.30) (-1.29) (-1.59) Femaleclients -0.00000176 -0.00000153 -0.00000176 -0.00000153

(1.11) (1.55) (1.42) (2.33) ROE 16.82 36.48 16.82 36.48*

(2.66) (0.21) (2.51) (0.35) Portflio 0.406** 0.208 0.406* 0.208

(-2.69) (-1.81) (-2.55) (-2.84) CPB -299.3** -259.3 -299.3* -259.3**

(-1.92) (-3.41) L2.OSS -0.623 -0.623***

(0.76) (0.23) (0.89) (0.51) L.OSS 0.170 0.0595 0.170 0.0595 Model 1 Model 2 Model 3 Mo~4 diff22robust Difference GMM

What is the finding : Difference GMM

* p<0.05, ** p<0.01, *** p<0.001t statistics in parentheses Observations 82 86 82 86 (0.96) (-2.62) (0.94) (-1.03) Constant 114.3 -583.2** 114.3 -583.2

(0.92) (2.66) (0.95) (1.08) Dummy Islamic 25.52 138.3** 25.52 138.3

(-2.00) (-2.65) lequity -16.33* -16.33**

(1.21) (-1.52) (2.33) (-0.63) lloans 9.058 -14.75 9.058* -14.75

(-3.24) (-2.05) (-3.40) (-2.00) lloanloss -12.90** -9.996* -12.90*** -9.996*

(-0.44) (5.36) (-0.54) (0.95) lloansize -4.778 86.91*** -4.778 86.91

(-0.62) (0.42) (-1.45) (0.49) Femaleclients -0.00000275 0.00000278 -0.00000275 0.00000278

(0.83) (0.12) (1.01) (0.05) ROE 11.66 1.960 11.66 1.960

(2.90) (3.34) (2.55) (1.10) Portflio 0.414** 3.018*** 0.414* 3.018

(-2.99) (-2.46) (-2.62) (-2.07) CPB -309.1** -287.8* -309.1** -287.8*

(1.48) (0.96) L2.OSS 0.133 0.133

(-0.44) (-1.06) (-0.27) (-0.60) L.OSS -0.0621 -0.132 -0.0621 -0.132 Model 1 Model 2 Model 3 Mo~4 system22ro~t System GMM

Title 17

Robustness Check

• Sargan Test

Prob > chi2 = 0.2162 chi2(17) = 21.22822

H0: overidentifying restrictions are validSargan test of overidentifying restrictions

Title 18

Implications

1. Islamic rules and principles should have significant on Microfinance performance.

2. Donors can consider Islamic Microfinance Institutions too as it may provide a different results.

3. It is recommend for managers and donors to ensure that Islamic inspiration translates into benefits for clients.

Title 19

Can MFIs maintain both Social and Financial performance ?

• Financial Inclusion Vs Financial Performance of MFIs

• Poverty eradication Vs Strengthening Microfinance Institutions

• Indicators : OSS Vs Average Loan Size

• Reference : Li, Yan. "Threshold Effects of Poverty Targeting on Financial Sustainability and Social Outreach of Microfinance Institutions."

Title 20

What is the Paradox of Microfinance ?

Higher loan size , Small no. of borrowers , better performance Smaller loan size , large no. of borrowers, worse performance.

Microfinance would enable a group of people with widely different characteristics and / or unbanked out of the stranglehold of moneylenders, therefore including them financially and socially.

The underlying assumption considers that the seniority in access to microfinance services can improve living conditions and reduce poverty among beneficiaries.

What is the Paradox of Microfinance ?

It could lead beneficiaries into situations of debt, and therefore aggravate their already difficult fate.

Given this duality of approaches, we assume that there may be a "threshold" at which if nothing is done, the positive effects of microcredit become exhausted and result in the deterioration of welfare recipient

Title 22

How to test ?

• Threshold regression models or sample splitting models have wide application in economics and applied econometric practice (Hansen 1999; Hansen 2000).

• It allows to endogenously determine the threshold level(s) at which the sample is split. It treats the threshold value(s) as unknown instead of arbitrarily deciding the splitting point.

• Hansen (1996) and Caner and Hansen (2004) suggest a bootstrapping procedure to obtain the asymptotic P-value.

Title 23

Why Bangladesh ?

Known as Birth Place of Microfinance.

• Bangladesh is well-known for its "minimalist" Grameen program focusing mainly on livelihood enterprises.

• MFIs in Bangladesh serve a total number of over 20 million members, with Grameen Bank leading the flock with 7.2 million, ASA with 6.4 million and BRAC with 5.3 million members.

• The average size of the loan stands at BDT8,000, making the Bangladesh microfinance segment a vibrant US$2.3 billion industry.

• All the MFIs claim to have reported excellent repayment rates at 98 percent and above.

• Data is available

Title 24

What is the Finding ?

Title 25

Why this result?

• By applying the sample splitting technique of Caner and Hansen (2004), we get a threshold value of 1 at which the full sample is not divided into difference regime.

• when the percentage of the poorest is relatively low and MFIs are actually serving a large number of marginally poor or non-poor clients.

• It does not support paradox hypothesis 

Title 26

Implications

• Good News for Policy makers , MFIs and micro- entrepreneurs.

• MFIs ( Specifically in Bangladesh) should not be more concerned about financial performance rather than social performance.

• If financial Inclusion is achieved, financial sustainability will be achieved too.

.

Thank You!

For any future queries please email me

1400028@student.inceif.org

AlHuda CIBE FZ LLE - U.A.EP: + 971 56 9286664, + 971 55 938 99 00

AlHuda Center of Islamic Banking & Economics - PakistanPh: (92-42) 35913096 - 98, Fax: (92-42) 35913056

Email: info@alhudacibe.com Website: www.alhudacibe.com