Long-term Credit Contract

Title:

Long-term Contract for Ultra-poor Microcredit Borrowers (with Seiro Ito, Kazushi Takahashi and Takashi Kurosaki)

Published Paper:

How Does Contract Design Affect the Uptake of Microcredit among the Ultra-poor? Experimental Evidence from River Islands of Northern Bangladesh (with Seiro Ito, Kazushi Takahashi and Takashi Kurosaki). Journal of Development Studies. (Forthcoming, 2016). [pdf]

Synopsis:

Although successful in reaching out to millions of the under-banked in developing countries, microcredit is typically not offered to the poorest of the poor or the ultra-poor (Armendariz, B. and J. Morduch, The Economics of Microfinance, Second edition, Cambridge, Mass.: MIT Press, 2010). In an attempt to enhance the welfare of the ultra-poor through microcredit provision, therefore, we implemented a four-year research project in FY2012-15 (JSPS Grant-in-Aid for Scientific Research (B)-24402023) to understand the effectiveness of “packaged” microcredit on the ultra-poor as an alternative to standard microcredit. Specifically, we randomly provided different microcredit schemes that were designed to unbundle the underlying mechanisms such as the lack of entrepreneurship and a long gestation period of income-generating projects. The first-stage research outcome regarding the uptake decision of borrowers has been published in a reputed field journal (Takahashi, K., A. Shonchoy, S. Ito, and T. Kurosaki, “How Does Contract Design Affect the Uptake of Microcredit among the Ultra-poor? Experimental Evidence from the River Islands of Northern Bangladesh,” Journal of Development Studies, DOI:10.1080/00220388.2016.1156092. Published online: 29 April 2016). The next-stage research outcome is about the impact of those credit schemes on repayment, livelihood strategies, and household welfare.

Our research involves the first experiment that directly gives the loan to the ultra-poor from the initial stage. Inspired by “Challenging the Frontiers of Poverty Reduction-Targeting the Ultra Poor” (CFPR-TUP) by BRAC－world’s biggest NGO based in Bangladesh－, several existing programs targeting the ultra-poor are all an asset (usually livestock) transfer model rather than a loan. The basic assumption of these existing programs is that the ultra-poor are too poor to be creditworthy, and a gift of asset can help them “graduate” from extreme poverty and convert them into a client of microcredit in the future. Our study challenges these conventional views and examines whether the ultra-poor can manage loans well and be better-off if appropriate credit designs are introduced.

Thus far, we have conducted three rounds of household surveys in the study area (i.e., the baseline surveys and two follow-up surveys) under the JSPS funding. The research objective of this project is to evaluate the long-term impact of microcredi as well as loan repayment performance.

The research outcome is expected to contribute to the academic understanding of how microcredit works in reducing the extreme poverty and to the improvement in microcredit designs in development practice. Regarding the former, we can particularly deepen our understanding of the role of entrepreneurship and asymmetric information in credit markets.

Research Design:

The research method is based on the randomized controlled trial (RCT) conducted under JSPS-24402023. The RCT was conducted in the river island areas, known as Chars in Bengali, of northern Bangladesh. The area is characterized by poverty and vulnerability due to frequent river erosion and floods and poor infrastructure and public service access. The sampling of our RCT involved double-stratified two-stage clustered sampling; in the first stage, we selected Chars, and in the second stage, we selected moderately-poor and ultra-poor households. Once our baseline survey was completed, we implemented the randomized credit offer in two levels. First, we randomly allocated 80 groups of Char villages into one of the four treatment arms. Second, within each Char, the credit was given only to 10 randomly selected households out of the 20 members in each group in the initial phase. More specifically, we randomly chose 3 moderately poor (MP) and 7 ultra-poor (UP) households per cluster, as we wanted to capture the impacts on general poor population that the previous research have not been able to study, and we would like to know if the impacts are different between moderately poor and the ultra-poor.

The four treatment arms are: 1) a regular Grameen-Bank style micro-credit (RC) of three consecutive loans of 5,000 BDT with a duration of one year, 2) a large credit (LC) of 15,000BDT with a duration of three years, 3) same large credit as previous with an initial one-year grace period (LC + GP), and 4) same large credit with a one-year grace period disbursed in-kind instead of cash (IK + GP). In each group, half of the members were randomly selected to receive the offered credit or credit package now and half was kept as a control to receive the product later.

From the quantitative analysis, we expect to see the impact of the lack of entrepreneurship and a gestation period on borrowers’ welfare. More specifically:

1. To estimate the impacts of unobservable entrepreneurial ability, we will contrast the packaged loan arm with large loan arm. The difference in outcomes will give the impacts of difference in required level of entrepreneurial ability in two arms.

2. To estimate the impacts of waiting period, we contrast the small loan arm with the large loan arm. The difference between the two are loaned amounts, existence of grace period, and maturity, all of which are components in determining the waiting periods. So this comparison will give the impacts of longer waiting period or nonconvex technology on outcomes.

3. As a byproduct, measuring the profitability of moderately poor borrowers and ultra-poor borrowers will give an opportunity to run a verification test of popular perception that the latter is riskier than the former. This by itself is of a great interest to practitioners of microfinance institutions and policymakers in all over the world.

4. By contrasting moderately poor and ultra-poor in each comparison, we can observe the heterogeneous impacts by the poverty strata. This is another interesting contrast as it will show the differences between moderately poor and ultra-poor, and hence whether such grouping is justified on the policy effectiveness ground.