How health shocks and its relationship to repayments of loans from microfinance institutions can affect migrant households in Bangladesh

ABSTRACT

Increasing the feasibility of easy loan repayment is one of the objectives of providing microcredits for income generating activities requiring relatively small loans in the developing countries. However, evidence in the developing countries suggests that microcredits are often used by households for non-income generating expenditures needed to deal with shocks such as health shocks. Health shocks in particular have severe financial implications due to the absence of health insurance programs in the developing countries and also because they reduce income generating capabilities. The objective of this paper is to examine, using a survey data from Bangladesh, whether there is an association between the health shocks and the loan repayments performance of the borrowers of microfinance institutions. When issues involving mixed distribution of the outcome variable and endogeneity of health shocks are taken care of with IV Tobit and Two-part models, the results show that the health shocks lower both the probability and the amount of repayments. In addition, since a significant portion of the households in Bangladesh represents migrant households, this hints at important policy implications.

KEYWORDS:

• Health shock
• microcredit
• repayment
• Bangladesh

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Our analysis focuses on the relationship between health shocks and repayment performance for the microcredits. We could not compare it with repayment performance for other types of institutional credits because no information was available. Data used in this study were collected with the purpose of studying microcredits.

2. The estimation strategy to take care of this issue of endogeneity is explained in Section 4.

3. These microfinance members may have also been the members of other formal or informal lending institutions like Banks or Rotating Savings and Credit Association.

4. These households were excluded not because they did not respond to the repayment question, but because they were not participants of microcredit survey.

5. Questionnaire can be made available upon request.

6. The households spent roughly USD 144 (BDT 11,200) on average. The majority of the borrowers (98%) seeks care from a wide range of providers starting from traditional healer to tertiary level care at medical college hospitals. Details can be made available upon request.

7. Their businesses could be of something beyond businesses like opening or managing a grocery store and may often entail traveling to distant places.

8. For a robustness check, we have used another measurement where the loan repayment rate at the time of the survey, is the ratio of the amount of loan that is repaid (between the date when loan is received and the survey date) and the amount of the total loan that is scheduled to be repaid (during that period). The repayment plan for the microfinance in Bangladesh involves weekly payments. For example, for a loan of 10,000 BDT for a year, the borrower needs to pay roughly 192 (= 10,000/52) BDT every week starting from the week after the loan is taken.

9. This term is the inverse Mills ratio of the probability calculated in the first stage.

10. We have also found another variable called ‘Ownership of Health Insurance’ which could also be used as an indication of health consciousness. However, given that health insurance may be costly, it can affect the loan repayment performance. As a result, this variable cannot be an appropriate instrument. Actually, the data set has the information about ‘Salt Intake’ and the ‘Type of Water Supply (like from ponds or canal, deep tube-well and tube-well) to the Household’. Each one of them could be an appropriate instrument. However, we could not use them because there are quite a few missing observations.

11. Being ‘health-conscious’ helps in avoiding health shocks but it does not affect the willingness to pay the loan or the existing financial condition such as asset holdings. Also, using non-smoking tobacco is not known to affect earning capacity in the short run (the time period covered in this study).

12. Female borrowers mostly get involved in income-generating activities that are available near their house/locality. Thus, field level loan officers can easily monitor their activities.

13. When we have tried interaction between occupations and health shock, it does not have a significant effect.

14. Hall, Rudebusch, and Wilcox (1996) used Monte Carlo simulation to show that simply having an F statistic that is significant at the typical 5% or 10% level is not sufficient. Stock, Wright, and Yogo (2002) suggest that statistic should exceed 10 for inferences based on the 2SLS estimator to be reliable when there is one endogenous regressor. Our F statistics has exceeded that threshold.

15. Since the Wald test statistic is significant only at 10%, it leaves room for using the instrument. In other words, there is not sufficient information in the sample to reject the null hypothesis of no endogeneity at 1% and 5% level. Accepting the fact there could be some trace of endogeneity, our use of instrument is justified.

16. Weak Instrument test calculates Lagrange multiplier (LM) or minimum distance (MD) versions of weak-instrument-robust tests of the coefficient of the endogenous variable beta in an instrumental variables (IV) estimation. In an exactly identified model where the number of instruments equals the number of endogenous regressors, it reports the Anderson–Rubin (AR) test statistic. The structural parameter, beta, is the coefficient on the endogenous regressor; b0 is a hypothesized value for beta; the excluded instruments are Z; the assumption is that the instruments are exogenous i.e. E(Zu) = 0. In other words, a well-specified model is one in which H0: beta = b0 cannot be rejected for a narrow range of hypothesized values of b0 and the assumption of instrument exogeneity cannot be rejected for a wide range of hypothesized values of b0 (i.e. the exogeneity assumption is generally satisfied). Since we fail to reject the H0: b = 0 (where beta is the coefficient on the endogenous regressor; b is a hypothesized value for beta), our model is well specified.

17. In their study for Hyderabad, India, Banerjee and Duflo (2015) did not find any developmental impact of microfinance on health. Although our study deals with a different issue, the results show that loans from MFI are used for healthcare.

Additional information

Notes on contributors

Md. Shahadath Hossain

Md. Shahadath Hossain is a Graduate student at the Department of Economics, Binghamton University, USA. He has a Master of Arts (MA) in economics from Central Michigan University, USA. His research interest includes migration and capital formation, women empowerment and child health, and economic impact of health shock.

Malabika Sarker

Malabika Sarker is an Associate Dean & Director Center of Excellence for Science of Implementation and Scale up at the BRAC James P. Grant School of Public Health at BRAC University, Bangladesh. She is a physician with a Master’s in Public Health (MPH) from Harvard University and a PhD in Public Health from University of Heidelberg, Germany. She is also an an adjunct faculty at the Institute of Public Health, Heidelberg University. Her scholarly interests include impact evaluation of universal health coverage with minimal economic compromise and the role of information communications technology in improving health information systems.

Shaila Nazneen

Shaila Nazneen is a Research Fellow at BRAC James P Grant School of Public Health, BRAC University, Bangladesh. She is a dental surgeon with a Master’s in Public Health (MPH) from North South University, Bangladesh. She studies psychosocial well-being and quality of life, adolescent and child health and rights.

Bharati Basu

Bharati Basu is a professor in the economics department of Central Michigan University. She studies international trade and economic development.

Fatema Binte Rasul

Fatema Binte Rasul is a PhD candidate in Heidelberg University, Germany. Her doctoral thesis is about behavioral aspects and related out-of-pocket expenditure regarding chronic non-communicable diseases in rural Bangladesh. She is also a lecturer in BRAC JPG School of public health, BRAC University, Bangladesh. Her main interests are chronic non-communicable diseases, behavioral aspects of health and elderly health.

Hossain I. Adib

Hossain I. Adib is a practitioner with experience of working in several development organizations both at national and international levels. His domain of work covers community-centric approaches to program design and delivery in the fields of microfinance, healthcare financing, technological innovations, disaster risk reduction and humanitarian response both in rural and in urban settings. He is currently working at Practical Action Bangladesh Country Program as the head of program implementation.