Written Toshio Kondo and based on a survey study carried out by the Asian Development Bank (ADB), released in September 2007 as an ADB Operational Report, 24 pages, available on November 2, 2008 at: http://www.adb.org/Documents/IES/PHI/IES-PHI-Impact-of-Microfinance.pdf.
The paper uses an economic approach to evaluate the impacts of the Rural Microenterprise Finance Project (RMFP) in the Philippines. RMFP is a government-sponsored program that aims to support rural microfinance institutions (MFIs) and other lenders. In the long-term, the government hoped the program would reduce poverty, create employment opportunities and increase the incomes of the poorest 30 percent of the rural population. Rural banks, cooperative rural banks, cooperatives, thrift banks and non-governmental organizations (NGOs) participated in the program, which ended in December of 2002.
When RMFP was completed in December of 2002, the program had reached 618,906 clients, of whom 97 percent were women (p. 7). As of June, 2006, government records showed that the program had reached some 1.6 million borrowers who on average were in their 75th month of borrowing and seventh loan cycle. On average, borrowers had taken out about PHP 70,000 (USD 1,538) in loans. About 9 percent of exiting clients were identified as “problem clients” while approximately two percent were identified as successful “graduates” (p. 7).
The research paper also reports the demographic characteristics of the respondent households (p. 8). The respondents were on average 44 years of age. About fifteen percent of household heads were female, and about 92 percent of respondents were female. Less than one percent had acquired no education; 31 percent had acquired some elementary education; 46 percent had acquired some secondary education and 23 percent had acquired tertiary education. Respondents had lived in the barangay for about 19 years, and the respondents had homes of 63 square meters on average.
Framework for Analysis
Kondo uses a one-time survey and a quasi-experimental pipeline design used by Coleman (1999) in his study of microfinance in Thailand (p. 1). Each “treatment” (i.e. RMFP-involved) barangay, or village, is matched to a different “control” barangay according to procedures outlined by Coleman. The treatment barangays had been involved with lending for some time, while the control barangays were similar on many angles but had not yet received any program loans. Unlike Coleman, Kondo includes in his group of client households several clients consisting of graduates and “problem” households. For a complete explanation of the economic framework and analytical models, please see pages 2 to 7. The rest of this paper wrap-up focuses instead on the significant quantitative and qualitative findings identified by Kondo:
Problems with Client Identification Procedures
Interestingly, the paper found that only ten percent of respondents were poor as defined by 2006 official poverty thresholds for the country (p. 8). Kondo provides a histogram showing the distribution of the difference between the respondents’ per capita income and the official poverty threshold (p. 9). While a large proportion of the respondents’ incomes are around the threshold, a much larger proportion are significantly above it. The histogram looks similar for both old and new clients, and Kondo argues that this disproves the notion that the program itself may be responsible for the participants’ rising incomes over time. Another surprising finding is that the non-participating households, the households that the community considers qualified for the program, exhibit a similar income distribution (p. 10).
Kondo comes to a number of significant conclusions based on the above findings. First, the procedures used to screen existing and new program clients, assuming they have been applied correctly, did not correctly identify poor clients as defined by official poverty guidelines. Second, the fact that non-participating households’ incomes present a distribution similar to that of participating households implies that the community incorrectly identified some potential clients, namely those above the poverty threshold, says Kondo. A third, logically-deducible conclusion is that the stakeholders to the program—field personnel, center leaders and barangay leaders—conveyed the message that the official poor may not be the targets of government microfinance programs.
Impact on Per Capita Income, Expenditure, Savings, and Expenditure on Food
Kondo uses a linear fixed-effects model to study the primary measures of household welfare—the per capita measurements of income, expenditure, food, total and savings (p. 10). The four treatment variables used in the study were: (1) a binary variable for whether someone received a program loan; (2) the number of months the program was available to the barangay (based on the first loan released to the barangay); (3) the value of loans (cumulative total amount of loans) released; and (4) the number of loan cycles. Using regression analysis, Kondo discovers that only the first of these four, the loan availability/access treatment variable, has a measurable effect on primary measures of household welfare (p. 12).
On page 12, Kondo provides a table summarizing the impact of availing of program loans on participants’ per capita income, per capita expenditures, on two definitions of per capita savings and per capita food expenditures. Kondo finds a low but significant probability (a significance level of 10 percent) that availing of program loans positively impacts per capita household income (p. 11). Taking out a loan on average provides an additional PHP 5,222 (USD 115) relative to those who have not taken out a program loan. Considering that on average households have cumulatively utilized PHP 70,000 (USD 1,538) in loans over six years (about PHP 11,000 or USD 241 per year), every PHP 100 or USD 2 in loans should increase income by PHP 47 or USD 1, says Kondo (p. 11). In the same table on page 12, Kondo shows that per capita expenditures are also positively affected by a family’s access to program loans. Following the same line of reasoning presented above, this translates into a PHP 38 or USD .8-increase in per capita consumption for every PHP 100 (USD 2) loan availed (p. 11). Kondo places his estimates on the impact of program loans on income and consumption in context with those of other well-known studies (p. 11).
Kondo’s study finds no significant effect of program loans on savings, as calculated using two different methods. The first method subtracts expenditures from income, while the second adds back in expenditures on education, health and durable furniture because these are not expected to be consumed in one period. However, Kondo does find a significant and positive effect (an extra PHP 1,333 or USD 29) on per capita expenditure on food.
Impact on the Number of Enterprises and Employment
The survey also asked respondents about the enterprises and employment in these enterprises of program clients and other household members. Table 12 on page 14 shows that about 93 percent of existing client respondents were engaged in household enterprises. This is higher than the percentages for new clients (87 percent) and non-participating households (78 percent). Among homes with household enterprises, the average number of enterprises was 1.8, and the total number of employed people was 2.4.
The paper focuses only the effects of program loans on total enterprises and total employment in order to avoid analytical complications arising from substitutions between program clients’ enterprises and those of other household members. Using a non-linear Poisson model, Kondo finds the impact of the program on both the number of enterprises and the number of employed persons in these enterprises to be very significant and positive. Table 13 on page 15 shows that households with program clients have 20 percent more enterprises and 17 percent more employed persons than non-program households.
Impact on Household Assets
This section attempts to answer how microfinance affected the households’ acquisition of assets such as land, equipment, livestock and poultry and household amenities. Table 14 on page 16 shows that about 20 percent of survey respondents owned land assets with an average current value of PHP 557,000 (USD 12,242). Fifteen percent of respondents owned farm equipment with an average current value of PHP 55,000 (USD 1,209). About 53 percent of respondents had livestock and poultry assets with an average current value of PHP 46,000 (USD 1,011). Finally, about 97 percent of respondents owned some household appliances, whose value on average was PHP 73,000 (USD 1,604). Using a non-linear Tobit model, Kondo finds that the program had little significant impact on total household assets.
Impact on Human Capital Investments (Education and Health)
To assess the program’s impact on educational attainment, the paper examines several variables corresponding to school attendance proportions for school-age children (6 to 12 years, 13 to 16 years and 17 to 24 years) in addition to educational expenditures per school-attending child. Kondo also tracks several health variables: the proportion of household members who are ill or injured, the proportion of those ill or injured who sought medical treatment, the proportion of children (0 to 5 years) who have received full immunizations and per capita health expenditures.
Table 15 on page 17 shows that about 95 percent of children 6 to 12 years of age, 87 percent of children 13 to 16 years of age, and 31 percent of children 17 to 24 years of age were attending school at the time of the survey. On average, households spent PHP 7,239 (USD 159) per school-attending child. Table 16 on page 17 shows that the proportion of either sick or injured in the six months preceding the survey was about nine percent, while the proportion of households with at least one ill or injured member was about 23 percent at the time of the survey. Sixty-nine percent of these people sought treatment. Annualized per capita health expenditures were about PHP 740 (USD 16).
Kondo used a fractional logit model to estimate the proportion of school-age children attending school and a linear fixed-effects model to estimate the expenditure per school-attending child. His estimation results showed that availing of program loans did not strongly affect either school attendance (among all age groups) or expenditures per school-attending child (p. 17). Likewise, the estimates for health indicate that availing of program loans did not significantly affect the proportion of ill or injured family members, the proportion of those seeking treatment, the proportion of fully immunized children (0 to 5 years of age) or the household’s per capital medical expenditures.
Miscellaneous Findings
In his paper, Kondo also finds that availing of program loans did not strongly affect the incidence of hunger among program families (p. 18). He also presents an interesting technical analysis of the program’s effect by income quartiles of respondent households (p. 19). Overall, the program had a regressive impact on family incomes. While Kondo finds a significant positive impact for households in the top quartile, there was a significant negative impact on all poorer households. He offers a number of possible explanations for this finding: (1) the problem clients may have been concentrated among poorer households; (2) the average loan size may have been smaller for poorer households; (3) there may have been a preponderance of diversion of loan proceeds from production to consumption among poorer households; and (4) if there was no diversion, the projects of poorer households may have been less productive. Kondo finds evidence only for explanations 1 and 2 (p. 19).
Similar Posts:
- MICROFINANCE PAPER WRAP-UP: “Green Energy Finance and Gender Disparity: The Case of Mountain Areas in Bangladesh,” by Sakib Bin Amin et al, Published by Copenhagen Business School
- MICROCAPITAL BRIEF: 60 Decibels Microfinance Index Data from 32 Countries Indicates Higher Confidence Among Group Borrowers, Greatest Debt Burden in Cambodia
- MICROFINANCE PAPER WRAP-UP: “Child Labor, Rainfall Shocks and Financial Inclusion: Evidence from Rural Households,” by Carolina Bernal and Razvan Vlaicu, Published by IDB
- MICROFINANCE PAPER WRAP-UP: “Impact of Microcredit on Household Consumption and Assets in Nepal;” by Shalik Ram Pokhrel
- MICROFINANCE PAPER WRAP-UP: “Mind the Gap in Financial Inclusion! Microcredit Institutions Fieldwork in Peru;” by Pilar Lopez-Sancheza, Elena Urquia-Grande
