Does capital and financing structure have any relevance to the performance of microfinance institutions?

Abstract

This paper aims to explore the impact of capital and financing structure on the performance of microfinance institutions (MFIs) from an agency theoretic standpoint. GMM and IV estimations with instruments have been performed using a panel dataset of 782 MFIs in 92 countries for the period 2000–2007. Results confirm the agency theoretic claim that an increase in leverage raises profit-efficiency in MFIs. The study also finds that cost efficiency deteriorates with decreasing leverage. Likewise, the negative significant impact of leverage on depth of outreach can also be explained. However, the study finds that capital structure does not have any noticeable impact on breadth of outreach and neither is it significantly related with women’s participation as loan clients.

Keywords:

• microfinance institutions
• outreach
• financial performance
• panel data estimation
• GMM estimation
• cross-country study

JEL Classifications:

• G21
• G32

Acknowledgements

The author would like to thank two anonymous referees for very helpful comments. Any remaining errors are, of course, the author’s.

Notes

1. Although MFIs are specialized financial institutions with social and financial missions, similar definitions of financing and capital structure should be used for them.

2. The institutionists stress the importance of profit-orientation, efficiency and self-sustainability in microfinance operations so that MFIs can generate enough revenue to meet their operating and financing costs. They argue that MFIs’ may attain these goals through massive scaling up that ultimately make them resilient to increased competition. In contrast, welfarists are less interested in MFIs’ efficiency and self-reliance. They emphasize self-employment and immediate improvement in the well-being of poor clients – especially women.

3. Modigliani and Miller (1958) provides numerous propositions to develop the theoretical foundations of the crucial link between capital structure and firm value, which is still believed to be relevant given that several strong assumptions (such as the existence of perfect capital market, homogeneous expectations, absence of taxes and zero transaction cost) are needed for it to hold. It held the position that capital structure is irrelevant and has basically no impact on firm performance. However, although a few also supported their views, this theory was later challenged by others as corporate governance theory served to understand the complex financing behaviour of corporate bodies (for detailed discussions see, for instance, Cassar and Holmes 2003; Kyereboah-Coleman 2007).

4. For brevity’s sake, basically, we skipped detailed discussions on relevant literature. Among others, for example, Berger and Bonaccorsi di Patti (2006), Kyereboah-Coleman (2007) and Cassar and Homes (2003) provide elaborate details on previous works.

5. The available MFI profiles and data are available in the public domain: www.mixmarket.org.

6. The Mix Market classifies MFIs in accordance with the level of information disclosure (within 1–5) provided. The higher the level provided the better the quality of data.

7. In our sample, several observations for ROE and Z-risk scores have been excluded because of their negative or zero values. MFIs with negative debt-equity ratios were also eliminated to avoid bad leverage. Some of the variables including DER, ROE, Z-risk and PAR30, for example, have been transformed through adding or multiplying a positive constant to every MFI’s observation so that the natural log is taken on a positive number. Natural logarithms have been taken so that normally distributed residuals are obtained and that the t-tests are valid.

8. The developing regions, according to the World Bank classification, are: East Asia and the Pacific (EAP), Eastern Europe and Central Asia (EECA), Latin America and the Caribbean (LAC), Middle East and North Africa (MENA), South Asia (SA) and Sub-Saharan Africa (SSA). Results of these time-constant dummies could not be reported since we have finally chosen the FE models.