![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
Abstract
Microfinance Institutions (MFIs) are used to claim that their impact goes beyond money since rescuing from exclusion uncollateralized poor borrowers significantly affects their dignity, self-esteem, social recognition, future economic perspectives and, through it, life satisfaction. Our article aims to verify the validity of this claim by evaluating whether access to microfinance loans has significant direct impact on life satisfaction beyond its indirect impact via current income changes. Empirical findings on a sample of poor borrowers in the suburbs of Buenos Aires show that, after controlling for survivorship, selection and interview bias, microfinance membership has a significant and positive effect on life satisfaction.
Acknowledgements
We thank Beatriz Armendariz, Kaushik Basu, Andrew Clark, Margaret McConnel for their useful comments and suggestions. We are also gratefully indebted to Adriana Abardia for her precious assistance in collecting data for the research. The grant of the Euricse foundation is gratefully acknowledged.
Notes
1 Accessed 28 October 2009.
2 Within this debate a line of thought argues that happiness indicators represent a unique, subjective and ‘nonpaternalistic’ measure of wellbeing which cannot be suspected of imposition from external experts and should reflect the real desires (driven by individual tastes which may be self driven or affected by social norms) of those who are targets of a policy intervention (Sugden, Citation2007). The critique to this position is well expressed by the ‘happy slave paradox’: if individuals are so deprived of their rights, they may be in a condition of not even desire their emancipation and therefore remain satisfied of their condition of slavery (Sen, Citation2005). The ample empirical literature however shows that happy slave paradoxes are irrelevant when drawing inferences on large samples which always reveal a strong positive correlation between life satisfaction and capabilities (Frey and Stutzer, Citation2002b).
3 Consider however that part of this enhanced trustworthiness may also have sound economic effects since, in a framework of contract incompleteness, many aspects of business relationships may be modelled under the form of an investment game (Berg et al., Citation1995). More specifically, the relationships between business partners, between an entrepreneur and her partners generally assume a sequential structure. One of the two players takes the initiative first by sharing something (knowledge, physical or financial assets) and, after it; the counterpart may decide whether to do the same or to abuse of the trust of the first mover. As in the typical investment game the counterparts joint decision to share (the trustor) and not to abuse (the trustee) has super additive effects and a higher outcome than the two suboptimal equilibria in which the first player shares and is abused or the first player decides not to share because she is afraid of the risk of being abused. As a consequence, enhanced reputation has monetary but also nonmonetary effects.
4 In the relationship with Protagonizar borrowers can have access to a new loan once they have paid the previous one. For credit cycle we therefore mean the number of subsequent loans that the borrower had with Protagonizar whatever the time interval occurred between two consecutive loans.
5 To account for survivorship we add to the original sample made by these two groups a third group of Protagonizar drop-outs in a number proportional to the historical dropout rate of the MF organization.
6 Evidence supporting the paradox is also reported by Blanchflower and Oswald (Citation2004) for the UK, Frey and Stutzer (Citation2002b) on a large sample of countries using data from the World Database of Happiness and the US Bureau of Census and Veenhoven (Citation1993) for Japan over the period 1958–1987. In spite of it, the Easterlin paradox is not in itself a regularity always confirmed across countries and time. When Castriota (Citation2006) repeats the Easterlin exercise on Eurobarometer data for some European countries in the last decade he actually finds that the paradox applies to Germany but not to Italy where a quite strong positive relationship between the happiness and per capita income is found.
7 Even though the question whether life satisfaction responses measure flow or lifetime utility is still open, life satisfaction measures nonetheless represent the closest empirical proxies to the standard utility concept in economic theories.
8 See also Penn (Citation2009) on the importance of nonresponse in self-reported measures of satisfaction.
9 For its earliest formulations see Gossen (Citation1854), Jevons (Citation1886) and Menger (Citation1994).
10 Extracted from the ‘microentrepreneurs’ stories’ section of the Protagonizar handbook (Protagonizar, Citation2005).
11 Real interest rates are not particularly high if we consider unofficial inflation rates. Consider in fact that several authors judge Argentinean poverty lines grossly undervalued due to a downward bias in computing domestic inflation. One of the main independent research centers, Ecolatina, estimates that prices rose 65% from 1 December 2006 to 31 July 2009, compared with the 20% increase calculated by the statistical institute (to follow this debate see http://www.bloomberg.com/apps/news?pid=newsarchive&sid=aKQUiLozzZko and http://www.bloomberg.com/apps/news?pid=newsarchive&sid=a5joiySC_mXc).
12 The questionnaire is omitted for reasons of space but is available from the authors upon request.
13 We include a third village (Villa de Mayo) in which Protagonizar activity has just started and there are no treatment group observations (MFI borrowers). This is typically done in impact studies in order to reduce the noise generating potential spill-over effects from treatment to control group in the two other villages. The econometric results of the article are however robust in a check in which we exclude respondents of Villa de Mayo from the control sample. Results are omitted here for reasons of space and available upon request.
14 Borrowers’ seniority is evaluated according to their credit-cycle. Since borrowers must first reimburse the previous loan in order to ask for a new one, a higher credit cycle is a proxy of a better borrower's repayment record. Given a median credit-cycle of 17, borrowers with a credit-cycle higher than (or equal to) 17 are categorized as ‘veteran’, while borrowers with a credit-cycle below the median as ‘new’.
15 We selected a number of dropouts from each area which is proportional to historical exit rates of borrowers from the organization.
16 During the survey period (July 2009–September 2009), the average malnutrition and poverty thresholds are set by the INDEC (National Statistical Agency of Argentina) at 4.88 and 11.04 pesos/day respectively, which are in turn equivalent to 3.84 and 8.70 PPP-US$ according the PPP country's factor evaluated by the World Bank in 2005. When considering the country's implied PPP factor in 2009 (US$ 2.033, Source: IMF), both the malnutrition and poverty lines fall to 2.40 and 5.43 PPP-US$ per day, respectively. Consider however that several authors consider Argentinean poverty lines grossly undervalued do a downward bias in computing domestic inflation (see footnote 8).
17 The variables considered in the principal component analysis are ownership of a clock, radio, CD, Fridge, TV set, DVD, VCR, bicycle, motorbike, sewing machine, tool kit, car, truck, personal computer, telephone line, cell phone, internet connection. The first component explains around 16% of the variability. It is positively correlated with all the variables mentioned above (the highest correlations are with personal computer, 36%, and internet connection, 34%).
18 We ask in the questionnaire about the presence of loans from the following additional sources different from Protagonizar: Non-Government Organizations (NGOs), trade credit, Rotating Savings and Credit Associations (ROSCAs), private banks, trade credit, family or acquaintances. Note that only 5% of individuals (10% among nonmembers and 2% among members and drop-outs) have access to these additional financing sources (mainly moneylenders for nonmembers and NGOs or trade credit for dropouts).
19 The question is standard in survey run in both high and low income countries and the same as the one posed in one of the main databases used for life satisfaction studies such as the German Socioeconomic Panel.
20 Van Praag (2007, p. 18) simply argues that ‘All these specifications amount to different specifications of the labeling system of the underlying indifference curves, but the indifference curves themselves are unchanged and are these indifference curves which are estimated, either by Ordered Probit, Logit or what else’.
21 Methodological problems such as scale heterogeneity in interpersonal comparisons using self-reported measures at the individual level (Harsanyi, Citation1955) are widely debated in the life satisfaction literature. There is however widespread consensus that these potential problems do not make empirical evidence on life satisfaction meaningless so that the related empirical literature has evolved and conquered space in top economic journals. More specifically on the issues at stake Di Tella and MacCulloch (2006) consider that, even though accepting the existence of heterogeneity in individual scales, we do not need to believe a priori that such heterogeneity is systematically affected by drivers of life satisfaction. More recently, Beegle et al. (Citation2009) with their vignette approach, document that heterogeneity cannot be ruled out but: (i) it is uncorrelated with happiness regressors; (ii) vignette rankings are not correlated with the residual of the standard life satisfaction regression; (iii) life satisfaction regression results do not change when self declared life satisfaction is rescaled with vignette results.
22 Among the first microfinance papers dealing with these issues see Hulme and Mosley (Citation1996), Pitt and Khandker (Citation1998) and Coleman (Citation1999).
23 Note that if we regress life satisfaction for clients and drop-outs only on the income variable we get a positive income coefficient with a t-stat. of 1.68 (2.16 if we do the same for the overall sample) which drops to 1.31 if we just introduce a second regressors such as the number of credit cycles (correlation between the two variables is positive and equal to 0.12). Hence credit cycles may proxy (together with nonmonetary factors) future expected revenues for members arising from the financed projects and therefore proxy permanent income better than current income.
24 The point is well resumed by Frey and Stutzer (Citation2002a, p. 59) claiming that ‘the level of education, as such, bears little relationship to happiness. Education is highly correlated with income [. .]. Education may indirectly contribute to happiness by allowing a better adaptation to changing environments. But it also tends to raise aspiration levels. Further, it has been found that the highly educated are more distressed than the less educated when they are hit by unemployment (Clark and Oswald, Citation1994)’.
25 Individuals who declared the highest level of satisfaction are about 28% of the sample. The propensity score is estimated using the following probit specification:
26 See also Blatmann and Annan (2010), Rosenbaum and Rubin (Citation1983) and Imbens (Citation2003) for a similar approach.
27 To have a bias of 60%, U should increase the relative probability of having Y = 1 by a factor greater than 10 and the relative probability of having T = 1 by a factor greater than 17. Similarly, to have a bias of 120%, U should increase the relative probability of having T=1 by a factor greater than 24. The presence among unobservable factors of a confounder with similar characteristics can be considered implausible in the present setting (where the set of matching variables W is quite rich).