Microfinance and Investment: A Comparison with Bank and Informal Lending

World Development Vol. 39, No. 6, pp. 882–897, 2011 Ó 2011 Elsevier Ltd. All rights reserved 0305-750X/$ - see front matter www.elsevier.com/locate/worlddev

doi:10.1016/j.worlddev.2011.03.002

Microfinance and Investment: A Comparison with Bank and Informal Lending LUCIA DALLA PELLEGRINA * University of Milan-Bicocca, Italy Paolo Baffi Center, Bocconi University, Italy Summary. — Comparing the impact of different types of credit on households’ investment in Bangladesh, we find that loans from microfinance institutions are likely to be channeled toward non-agricultural activities while both informal and bank lending are associated to a higher expenditure in agricultural inputs. Estimated effects are net of the differences in the amount borrowed, interest rates, and collateral. Results suggest that features which are specific to microfinance—such as tight repayment schedules and land-based eligibility rules—may reduce the suitability of this source of funds for the farming sector. Ó 2011 Elsevier Ltd. All rights reserved. Key words — microfinance, banks, informal lending, investment

1. INTRODUCTION

on a comprehensive definition of investment—expenditure in both working capital and fixed assets—and by including all available sources of credit. 2 The reason why we focus on investment is that, as opposite to consumption behavior, the former is more suited to provide insights on programs’ long-term effect on growth. For example, practitioners (Ba˚ge, 2004) stress that in order to achieve self-sustainability households in low income contexts should not myopically consume borrowed funds but rather invest them in productive activities. 3 Ahlin and Jiang (2008) also claim that the key to the success of MF long-term objectives rests in the fact of promoting the gradual accumulation of average returns in self-employment. These recommendations, though, might be ineffective when addressed to poorest people, since, having a higher time preference rate compared to richer individuals, they might be tempted to raise present consumption instead of acquiring production inputs (Lawrance, 1991). Hence—at least at first glance—it becomes important to verify to what extent borrowers from MF programs perform better than non-borrowers, and this is essentially what has been done by most of the empirical studies dealing with the impact of MF. However, this approach seems somehow restrictive since it classifies borrowers from other sources as non-borrowers. In fact, in the same vein as ignoring non-treated groups (see, e.g., Heckman, 1976), disregarding those who are treated with different types of credit may provide biased estimates when evaluating the impact of MF programs. Moreover, comparison with other types of lending is important to our purposes since it allows verifying whether there are features of MF

A large part of the literature on microcredit has been devoted to the analysis of its effectiveness in terms of poverty reduction. Several applied studies have investigated the impact of different programs operating on the basis of group lending on the behavior of households and firms, such as per capita consumption, labor supply, children school enrollment (Morduch, 1998; Pitt & Khandker, 1998), and business performance (Madajewicz, 2003a; McKernan, 2002), providing evidence of success. Standardized microfinance (hereafter MF) agreements, however, have been recently criticized since they are considered not properly suited to fulfill the needs of all sectors of the economy. Agriculture, in particular, seems to suffer from the absence of contractual flexibility (Christen & Pearce, 2005; Llanto, 2007; Meyer, 2002; Murray, 2001). In fact, it is a common practice for microfinance institutions (from now on MFIs) to lend to landless households and require reimbursement of the loan soon after it has been granted. In particular, tight repayment schedules may preclude borrowers from undertaking long-term investments, as is often the case in agriculture where the production cycle is longer than in several other activities. In addition, farmers may encounter difficulties to commit to regular installments due to the risk related to climate conditions (see, e.g., Caldwell, Reddy, & Caldwell, 1986). This might also bias MFIs in favor of the non-agricultural sector as consequence of reducing the risk of default more commonly associated with natural disasters. 1 Such a situation could be exacerbated by the eligibility rules for MF programs which rely on the lack of land ownership in order to identify poor borrowers. As opposite to other producers, in fact, farmers are more likely to have all their wealth concentrated in small land plots, and are thus less entitled to obtain loans. This paper examines the role of financing mechanisms on household investment decisions. In particular, we aim at verifying what kind of investment—in agricultural and nonagricultural activities—is promoted by different kinds of lending: micro-lending, informal lending, and bank lending. The article, in particular, adds to previous work by concentrating

* I am grateful to two anonymous referees, Gani Aldashev, Jean-Marie Baland, Vittoria Cerasi, Lisa Crosato, Asif Dowla, Hadi Esfahani, Christopher Flinn, Eliana La Ferrara, Robert Lensink, Niels Hermes, Matteo Manera, Jean-Philippe Platteau, Hans Seibel, and to seminar participants at University of Namur (FUNDP), University of Milan-Bicocca, Bocconi University, University of Lecce, and 2007 Conference on Microfinance at the University of Groningen. I am particularly grateful to Shahidur Khandker for allowing me to use the World Bank dataset. The usual disclaimer applies. Final revision accepted: October 30, 2009. 882

MICROFINANCE AND INVESTMENT: A COMPARISON WITH BANK AND INFORMAL LENDING

contracts which are likely to penalize some categories of borrowers, such as farmers. If this is the case, we should observe that these categories more frequently apply to other credit providers having different characteristics. 4 We use data from a survey of the World Bank carried out during the years 1991 and 1992 in Bangladesh. The survey contains information about credit from MFIs and nongovernmental organizations providing group lending, loans obtained from landlords, input suppliers, shopkeepers, employers, relatives and friends (which we define as the informal lending channel), and banks. There are a number of reasons why it is still useful to investigate data related to a period when group lending was predominant in the context analyzed in the paper. First, although the Grameen Bank (GB) has significantly reduced the focus on the traditional five-person group in favor of the village organization, it still maintains the group structure (Barua & Dowla, 2006), so that borrowers’ incentives should not be significantly affected by these changes (Armendariz & Morduch, 2005, p. 101). Second, traditional group loans are even now the core of credit services provided by other institutions included in our dataset, such as the Bangladesh Rural Development Board. 5 Third, the group lending model as used by GB in the 1990s is still the dominant one in many developing countries, and especially in Africa (see, e.g., Basu, Blavy, & Yulek, 2004, on a variety of African counties, Brandsma & Chaouali, 1998, on North Africa and the Middle East; Hermes, Lensink, & Mehrteab, 2005, on Eritrea). The empirical analysis has been carried out through techniques that use instrumental variables to reduce the endogeneity bias generated by the correlation between the choice of the credit source and non-measurable characteristics affecting investment, such as for example, household members’ ability. Moreover, we concentrate on a core relationship involving investment, on the one hand, and the probability of borrowing from each source, on the other hand, while the impact of the amount borrowed, interest rate, and collateral is analyzed separately. This helps isolating the effect of other characteristics of credit contracts such as, among others, the repayment system. Results show that, conditional on measurable features of credit agreements, borrowers from MFIs are likely to invest more in non-agricultural activities while both informal and bank lending are associated to a higher investment in agricultural inputs. According to what has been discussed so far, this seems to provide evidence in favor of our hypothesis concerning the lower suitability of MF standardized lending programs for the agricultural sector. It is worth stressing, however, that there may be other factors driving our empirical findings. Joint or individual liability (Ghatak & Guinanne, 1999; Hermes & Lensink, 2007; Hermes et al. 2005; Madajewicz, 2003b; Paxton, Graham, & Thraen, 2000; Sharma and Zeller, 1997; Wydick, 1999), the quality of monitoring (Armendariz, 1999; Banerjee, Besley, & Guinnane, 1994; Madajewicz, 2003a; Stiglitz, 1990; Varian, 1990, among others) and the pattern of sanctions (Besley & Coate, 1995) may induce different attitudes toward investment depending on the contract chosen. Threat of future credit denial, which is typically used in MF, could also work in a similar way. Nevertheless, there seems to be weaker evidence, as well as a less intense debate, as to whether these elements are likely to unevenly affect different sectors of the economy. The rest of the paper is organized as follows. In Section 2 we illustrate the dataset. Section 3 concentrates on the estimation techniques and instruments adopted. In Section 4 we discuss the results. Section 5 concludes.

883

2. DATA Data were collected in a survey carried out on 1,798 households in rural Bangladeshi villages by the Bangladesh Institute of Development Studies at the World Bank in 1991–92. The survey was conducted in three rounds, approximately corresponding to the harvesting of the rice crop. The first round (November 1991–February 1992) corresponds to the Aman season, the second (March–June 1992) to the Boro, and the third (July–October 1992) to the Aus. The original sample consists of three randomly selected villages from each of the 29 districts (thanas) surveyed. In 24 of these districts, a microcredit program had been in operation for at least three years. A total of 20 households in each village were surveyed. We mainly concentrate on the first round since a great deal of information went missing during the remaining two. In particular, we work under a cross-sectional setup, 6 using the other rounds to gather information on investment in fixed assets (i.e., to compute the difference in the stock), since data do not provide a direct measure of this variable. 7 Households engaged in self employed activities number 1,276. Almost all of them (1,192) are farmers or fishermen, 797 are non-farmers, and of these 713 are engaged in both farming and non-farming activities. Investment in working capital corresponds to total operating costs in the year preceding the survey. 8 In the case of farming, these include expenditure on seeds, fertilizers, pesticides, water, tillage, rented labor, and veterinary costs. Non-farmers’ operating costs 9 are constituted by raw materials, rented labor, fuel for transport, and other expenses for equipment maintenance. Investment in fixed assets is computed as the incremental value of physical capital between the first and the third rounds of the survey. Capital consists of bullocks, cows, sheep, poultry, and agricultural equipment if the household engages in farming activities. As for non-farmers it is mainly constituted by buildings, machinery, rickshaws, sewing machines, and other durables. We exclude inherited assets since they are not bought by the household, neither through credit, nor by self-financing. Land is also omitted—it is instead used as a control variable for reasons that will be discussed in the following sections—while rent paid for land is accounted for as part of operating expenditure. The importance of distinguishing between fixed assets and working capital is a pure statistical issue. It lies in the fact that the distributions of these two variables differ considerably (see the following section). More precisely, fixed assets may be dismissed—hence a negative investment is possible—and also show a considerably higher variance as compared to working capital. Jointly considering the two variables would not account for the heterogeneity of the latter, since all the effects would be driven by the relationship between credit and the more highly-volatile component of investment. 10 Table 1 reports summary statistics on investment. Working capital expenditure is on average 1,297 taka for farmers compared to 1,191 for non-farmers. Also investment in fixed assets is higher for farmers (766 taka, against 69 for non-farmers if negative values are excluded; 350 taka, against 3,958 for non-farmers if negative values are included). 11 T-tests of mean comparison reported in Table 1 suggest that there is no significant discrepancy between the operating costs of farmers and non-farmers. The difference in means is instead weakly significant for investment in fixed assets. We consider all loans granted to household members in the year preceding the survey. The sources of microcredit are the GB, the Bangladesh Rural Advancement Committee (BRAC), and the Bangladesh Rural Development Board (BRDB). All

884

WORLD DEVELOPMENT Table 1. Investment in working capital and fixed assets (in Taka): farming and non farming activitiesa Obs.

Mean

Std. Dev.

Min

Max

Farmers Investment in working capitalb Investment in fixed assets after 12 monthsd Investment in fixed assets after 12 months correctede

1,192 1,192 1,007

1,297 350 766

2,263.745 2,589.825 2,325.911

0 20,000 0

31,420 30,000 30,000

Non-farmers Investment in working capitalc Investment in fixed assets after 12 monthsd Investment in fixed assets after 12 months correctede

797 797 224

1,191 3,958 69

6,472.574 13,228.07 638.5657

0 225,000 0

108,461.5 7,950 7,950

T test of farmers and non-farmers mean comparison:

Diff. in means

Std. Dev.

T statistic

Degr. of freedom

Investment in working capital Investment in fixed assets after 12 monthsd Investment in fixed assets after 12 months correctede

105 4,307 697

203.8573 393.9166 481.6863

0.5151 10.9338f 4.4484f

1,987 1,987 1,229

First round of interviews: 1991/11–1992/2. Second round of interviews: 1992/3–1992/6. Third round of interviews: 1992/7 to 1992/10. a Based on respondents’ statements about the main activity carried out by the household. b Corresponds to the average expenditure of one production cycle in the year preceding the (first round of the) survey. c Corresponds to the expenditure one year preceding the (first round of the) survey. d Corresponds to the difference between fixed assets in the third round of the survey and fixed assets in the first round of the survey, including negative values. e Corresponds to the difference between fixed assets in the third round of the survey and fixed assets in the first round of the survey, excluding negative values. f Denotes significant difference in means.

three institutions operated through group lending at the time of the survey. The class of informal lending, instead, is represented by suppliers and merchants, landlords, relatives, and neighbors. On the one hand, loans from suppliers and merchants are almost exclusively short-term based and collateral-free, and are recovered through the purchase of the output at a price agreed in advance. On the other hand, landlords are typically wealthy persons who set very high interest rates and often require physical collateral, as opposite to relatives and neighbors who lend at lower interest rates and rarely require collateral. In general, the main characteristic of borrowing from the informal channel consists in the advantage of obtaining immediate approval and flexible amounts of money (Timberg & Aiyar, 1984). Finally, the banking sector consists of commercial banks, specialized banks, and cooperative banks. Statistics on credit are reported in Table 2. The number of households borrowing from MFIs under group lending during the year previous to the survey is 297, while those borrowing from the informal sector and banks number 111 and 40, respectively. Average micro-loans are 6,386 taka. A slightly lower principal is accorded by informal lenders (5,445 taka), while banks provide substantially higher amounts (11,795 taka). Interest rate is 16% on both MF and bank contracts, 12 compared to a mean of 52% on informal ones. In particular, informal rates differ considerably across the sample with a standard deviation of 62%. 13 Informal moneylenders and banks require collateral on 10% and 45% of loans respectively, whereas MFIs never require any guarantee. We build a dummy variable taking the value of 1 in case collateral is required by the lender, although there are no observations concerning its value. Loan duration at the time the loan is granted 14 is almost the same for MF and informal credit (392 and 384 days, respectively), while the duration of bank contracts is considerably higher. At the time they were surveyed, the percentage of borrowers having repaid the loan ranges between 12% and 15% regardless the type of contract, and should presumably correspond to loans obtained at the beginning of 1991. As we consider loans that have been obtained earlier in time, that is 1.5

and 2 years previously, the repayment rate increases. As one can see, despite having the same duration of informal lending, MF loans seem to perform better in this regard. Banks seem to perform the worst, but since bank contracts are of longer duration they cannot be compared with other sources. As for the timing of loans and investment, we assume that working capital expenditure takes place within 12 months from the loan, while disbursement for fixed assets is assumed to occur up to twenty-four months from the loan. The choice of time lags relies on the fact that buying working capital does not require a lengthy and detailed evaluation process, and normally occurs at the time—or right after—the loan is obtained. Instead, longer periods may elapse from the time the loan is granted to its actual disbursement for purchasing physical capital. We explain this with the fact that fixed assets normally represent important expenditures, 15 and, therefore, buyers and sellers are likely to engage in lengthier negotiations— which often involve cash advances—before the transfer of ownership occurs. Sometimes agreements may even fail to complete, so that entrepreneurs must further delay purchases since they need to find other potential sellers. From a breakdown of the main loan characteristics in terms of borrowers’ activity it emerges that typically MFIs and informal lenders grant similar sums to households engaged in farming and non-farming activities, while banks provide substantially higher amounts to farmers. This may be due to the fact that farmers own land, which, according to the statements of respondents, is almost the only asset accepted as collateral. Therefore, farmers seem to have a better access to bank lending but also appear to be precluded from accessing the MF sector, perhaps because of land eligibility rules (see below). For the purposes of this paper, it may be useful to combine information on credit and investment discussed so far. A rough measure of the ability of different credit sources to raise investment, in fact, is the ratio of expenditure to the amount borrowed for each type of loan. This is a proxy of the share of invested funds, that is, not used for consumption purposes. 16 Table 3 reports computed values of this measure, separating working capital from fixed assets. At first glance, what seems

MICROFINANCE AND INVESTMENT: A COMPARISON WITH BANK AND INFORMAL LENDING

885

Table 2. Credit: amount borrowed, interest rates, and collateral Number of households Group lending (MFIs) Amounta (taka) Interest rate (%) Collateral (% loans requiring-) Duration (days)

Mean

Std. Dev.

Min

Max

6,386 16 0 393

6,262 0.63 0 192

1,000 16 0 0

54,014 20 0 2,338

0.12 0.45 0.81

0.33 0.49 0.38

0 0 0

1 1 1

5,445 52 0.09 385

5,924 62.10 0.30 509

1,000 0 0 12

40,000 240 1 3,710

0.13 0.17 0.44

0.34 0.38 0.51

0 0 0

1 1 1

11,795 16 0.45 511

27,707 0 0.503 407

1,000 16 0 0

175,000 16 1 2,190

0.15 0.22 0.34

0.36 0.42 0.47

0 0 0

1 1 1

Diff. in means

Std. Dev.

t statistic

Degr. of freedom

940.865 5,409.36 6,350.23

686.6809 1,875.675 2,777.574

1.3702 2.8840 2.2863

406 335 149

297

Repayment rate Date loan 1991–2011 to 1991–2012 Date loan 1990–2007 to 1991–96b date loan 1989/1–1989/12c Informal lending Amounta (taka) Interest rate (%) Collateral (% loans requiring-) Duration (days)

111

Repayment rate Date loan 1991/1–1991/12 Date loan 1990/7–1991/6b Date loan 1989/1–1989/12c Bank lending Amounta (taka) Interest rate (%) Collateral (% loans requiring-) Duration (days)

40

Repayment rate Date loan 1991/1–1991/12 Date loan 1990/7–1991/6b Date loan 1989/1–1989/12c T-test of mean comparison on the amount borrowed: Group lending versus informal lending Group lending versus bank lending Informal lending versus bank lending

All data refer to credit obtained in the period 1991/1–1991/12 (one year preceding the survey). In order to compute the lending interval the central time of the first round of the survey (1991/12/31) has been considered as a point in time to move one year backward. a Cumulative amount borrowed by the household. b Credit obtained 1.5 years preceding the survey. c Credit obtained 2 years preceding the survey.

Table 3. Ratio of investment to the amount borrowed Farmers

Nonfarmers

t test of mean comparisona

Group lending Working capital Fixed assets after 12 months

0.143 0.024

0.153 0.142

0.2049 (592) 2.6885b (443)

Informal lending Working capital Fixed assets after 12 months

0.523 0.178

0.091 0.000

2.8943b (220) 2.1193b (178)

0.276 0.098

0.081 0.000

1.9487b (78) 1.6148 (58)

Bank lending Working capital Fixed assets after 12 months a b

t statistics are reported. Degrees of freedom in parentheses. Denotes significant difference in means.

interesting (see figures in italics) is that farmers who are financed by informal lenders tend to show the highest expendi-

ture ratio in both working capital and fixed assets, while the same occurs to non-farmers when borrowing from MFIs. However, univariate statistics may not account for important components such as the fact that borrowers have several different characteristics like, for example, wealth. More relevantly they do not account for causality. Among the other variables that may affect investment decisions, own wealth is particularly relevant since it may allow self-financing. It has been recognized that in less development contexts like the one we are analyzing, land seems the most reliable proxy of wealth. In particular, we consider cultivable land following the de facto approach described in Pitt (1999), although a measure of cultivated land is also accounted for. For most of the households the latter is different from land owned since many of them are sharecroppers or rent land for farming. In addition, it is also important to account for land tenure since the latter may affect incentives to increase investment (a typical example is the traditional sharecropping inefficiency described in Eswaran & Kotwal, 1985). To this purpose we build two dummies each one taking the value of 1 when the household head is either a sharecropper or rents land.

886

WORLD DEVELOPMENT Table 4. Descriptive statistics: dependent variables

Variable HH head: age (years) HH head: education (years) HH spouse: age (years) HH spouse: education (years) HH head: father lives in HHa HH spouse: father lives in HHa HH spouse: mother lives in the HHa HH spouse: mother lives in the HHa HH head is malea Religiona (1 = Islam) Feminine ratio No. of persons in HH No. of parents alive No. of siblings alive No. of other relatives alive Father: same activitya Total area cultivated (acres) Tenure: fixed renta Tenure: sharecroppera Land (acres) House value (taka) Transport value (taka) Injury last yeara Medical expenditures last year (taka) No. of days not working last year Transfers last year: cash value (taka) Transfers last year: other value (taka) Transfers last year: food value (taka)

Mean

Std. Dev.

Min

Max

40.24027 2.377642 29.00111 1.161846 0.023916 0.001669 0.154060 0.008899 0.943270 0.883760 0.493146 5.226363 1.838710 8.120133 6.739711 0.756396 0.608766 0.130701 0.263626 0.694768 1033.122 29.42158 0.334263 169.4914 5.903226 168.2864 16.19244 10.49194

12.42312 3.404029 14.46732 2.317328 0.152828 0.040825 0.361107 0.093939 0.231390 0.320602 0.172192 2.293305 1.288886 3.839145 5.771005 0.429375 1.138134 0.337167 0.440722 2.296784 6192.850 459.3392 0.471862 677.1430 10.11567 1485.269 148.2298 105.6856

16 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

85 16 70 14 1 1 1 1 1 1 1 19 4 25 36 1 15 1 1 52.5 160,000 15,000 1 15,600 80 33,000 5,000 2,000

Observations: 1798. a Dummy variables.

The survey also provides information on household members, such as the age and education of the household head and spouse, the gender of the household head, his/her religion, his/her father’s activity, the number of family members, femininity ratio, and several other factors which are used as controls. Finally, since networking can also affect both investment choices and the form of financing (Wydick, Karp, & Hilliker, 2007), a set of controls aimed at capturing the relationship network of the household is included. These concern the number of various types of relatives of the household head and spouse who are alive, and in particular those living in the household. All variables are summarized in Table 4, while a more extensive definition is provided in Table 12 in the Appendix. 3. ESTIMATION One crucial step toward the identification of the impact of different credit channels on investment is to address the problem of the endogenous nature of the former, which would cause estimates to be inconsistent otherwise. Therefore, we first investigate the mechanism underlying the selection process in each type of credit channel in order to find suitable instruments. Then we test for the effect of the predicted value of credit on investment. We estimate both equations of investment in working capital (4) and fixed assets (5), conditional on borrowing from each type of lender (Eqns. (1)–(3)) and on a set of control variables, representing credit features, household preferences, and technology.

The complete set of reduced form equations is the following: CM ij C Iij C Bij

¼ a0M þ X ij aM þ Z Cij bM ljM þ eijM

ð1Þ

a0I þ X ij aI þ Z Cij bI ljI þ eijI a0B þ X ij aB þ Z Cij bB ljB þ eijB

ð2Þ

¼ ¼

Eij ¼ a0E þ X ij aE þ Z ij bE þ

^ M cE C ij

þ

ð3Þ ^ I dE C ij

^ B kE þ ljE þ eijE þC ij ^ M cA þ C ^ I dA Aij ¼ a0A þ X ij aA þ Z ij bA þ C ij ij B ^ þ C kA þ ljA þ eijA ij

ð4Þ ð5Þ

where i identifies the household, and j refers to the village. Eij and Aij are investment in working capital and in fixed as^I , C ^ B are dummy variables taking the ^M, C sets, respectively; C ij ij ij value of 1 if the household borrows from a MFI, an informal lender or a bank, respectively. Credit is treated as binary in this context. A separate variable reporting the amount obtained from whatever type of lender is exploited to isolate the money-value contribution of loans to investment. Two other measurable components of credit, such as the interest ^I , C ^ B are the fitted ^M, C rate and collateral, are controlled for. C ij ij ij M I B values of C ij , C ij , C ij , respectively, representing the probability of being financed from each type of lender. The associated parameters, which are our main concern, can be interpreted as the additional investment induced by borrowing from each credit source compared to non-borrowing. X ij is a vector of general characteristics of the household common to all equations, such as religion, age, gender and education of the household head, the feminine ratio, 17

MICROFINANCE AND INVESTMENT: A COMPARISON WITH BANK AND INFORMAL LENDING

variables seizing on the relationship network of the household, as well as technological features (total area cultivated and land tenure) and proxies of idiosyncratic shocks, such as illness and medical expenditures. Measures of wealth (land and other assets), transfers, and elements capturing the possibility that relatives interfere with investment decisions are also included in the vector X ij for reasons that will be explained further on. Z Cij are characteristics of the household that affect credit transactions but not other household estimated behaviors (instruments), while Z ij includes measurable characteristics of credit contracts which may influence the propensity toward investment, such as the amount borrowed, interest rates, and collateral requirements. 18 Note that these variables are not to be classified as instruments since we observe values that are different from zero only when loans actually take place. If we were to do so, we would observe, for example, that the correlation between interest rates and the likelihood of obtaining credit is positive, which is counter-intuitive. The same problem occurs when using whatever information available only for households participating in credit transactions. a0M , a0I , a0B , a0E , and a0A are constant terms; ljM , ljI , ljB , ljE , and ljA are village specific-effects, while eijM , eijI , eijB , eijE , and eijA are idiosyncratic errors, such as Eðeij jX ij ; Z Cij ; lj Þ ¼ 0 in Eqns. (1)–(3), and Eðeij jX ij ; C ij ; Z ij ; Z ij ; lj Þ ¼ 0 in Eqns. (4) and (5). The following is a covariance matrix of error terms of the reduced form of the model: 1 0 2 rM rMI rMB rME rMA C B rIB rIE rIA C B rIM r2I C B B rBM rBI r2 rBE rBA C B C B C B @ rEM rEI rEB r2E rEA A rAM rAI rAB rAE r2A We assume that rME , rMA , rIE , rIA , rBE , and rBA are zero. By doing this we allow for both unmeasurable household features that simultaneously affect all credit equations, and unmeasurable household features that simultaneously affect expenditure. There are several reasons for these assumptions. Regarding credit equations, one is the possibility that there are individual characteristics of the households which induce some of them to borrow from multiple sources. Jain and Mansuri (2003), for example, explain this practice with the fact that some of them get refinanced by informal lenders in order to meet a more rigid structure of installments of other outstanding contracts. This frequently occurs in case of severe negative shocks. The presence of these shocks, in case they cannot be controlled for, falls in the error term causing rMI , rMB , and rIB to be different from zero. The hypothesis of rEA different from zero, instead, stems from the possibility that there are unmeasurable characteristics of the household, such as longer experience in carrying out some activities, which may raise investment in both working capital and fixed assets. We instead exclude there being unmeasurable characteristics common to credit and investment equations. 19 Finally, we tackle the problem of outliers by using a percentile approach. On the one hand, the distribution of working capital is—not surprisingly—skewed to the right and truncated in zero. Therefore, we gradually trim 20 the right tail, which corresponds to larger activities, since these are more likely to include outliers. 21 In particular, we carry out estimates first on the full distribution (including the zeros), then on observations below the 90th percentile, and finally on observations below the 70th percentile. On the other hand, as discussed previously in the paper, fixed assets involve both

887

positive and negative values, while zeros lay in the center of the support. Hence, possible outliers are eliminated by gradually trimming both tails of the distribution. 22 In this case, we perform regressions on the full distribution, and then on 90% and 80% of it, eliminating observations in correspondence of the tails. Differences in the portion of observations trimmed from the distributions of working capital and fixed assets are due to the non-equal number of zeros (more frequent for the latter) along their support. (a) Sources of bias As pointed out by Pitt and Khandker (1998), biases that may arise when treating program effects—particularly when dealing with cross sectional data—can be summarized into three major classes. The first originates from nonrandom placement of credit programs. Treating placement as random, when instead programs are most frequently allocated in poorer districts, can lead to a downward bias of program effects, as discussed in Pitt, Rosenzweig, and Gibbons (1993) and Heckman (1990). This problem mainly concerns MF, although a similar argument holds for banks, which may not be uniformly distributed across the sample. The second class of bias is related to unmeasured village attributes that affect both credit transactions and household behavior. Climate conditions and a high propensity to natural disasters, among others, are important characteristics affecting both these variables, especially when dealing with agriculture. The last source of bias concerns unmeasured household features that affect both credit transactions and households’ behavior (see Heckman, 1990, for a general discussion on selection bias). These are—typically time-invariant—intrinsic characteristics or personal qualities, like ability and individual aptitudes. Problems of this kind are traditionally solved using instrumental variables. In particular, the selection system originated by eligibility rules, and other factors which do not depend on borrowers’ aptitudes, are often exploited to correct for selection mechanisms. (b) Techniques and instruments Eqns. (1)–(5) are estimated by means of Two-Stage Least Squares (2SLS) with village Fixed-Effects, 23 where the latter are exploited to correct for both nonrandom allocation of credit and unmeasured village characteristics mentioned above. Instruments are instead used to further correct for selection biases that are induced by unmeasured household features. The selection mechanism for MF programs has been thoroughly investigated in the literature. The ownership of less than 0.5 acres of cultivable land, which is the eligibility rule for MF programs, is normally assumed to be exogenous with respect to households’ decision making in contexts where the land market is rather static. The absence of an active land market is in fact the rationale for the treatment of land ownership as an exogenous instrument for identifying the impact of credit in almost all the empirical work on household behavior in South Asia (see Pitt & Khandker, 1998; Rosenzweig, 1980; Binswanger & Rosenzweig, 1986; Rosenzweig & Wolpin, 1985; see also Pitt, 1999, for a discussion on de jure and de facto eligibility rules, i.e., the enforcement of the half-acre rule in practice, and for addressing the concerns raised by Morduch (1998), on land purchases by program households). Following the previous literature, we use a dummy variable (Target) that takes the value of 1 if the household owns less than 0.5 acres of land as an instrument for group lending,

888

WORLD DEVELOPMENT

while controlling for continuous measures of land owned and land cultivated in X ij . 24 We also add other exogenous measures of wealth—ownership of an inherited house and means of transport—in order to minimize the presence of other components of wealth in the error term. As can be observed in Table 5, target households are more frequently located within the set of MF borrowers (80% against 60% and 53% of households borrowing from informal lenders and banks, respectively). Turning to bank lending, the candidate instrument relates to other collateral requirements that do not directly involve households’ characteristics. By analyzing the answers to the survey one can observe that two main types of guarantees are accepted by banks, physical collateral (mainly own land and, to a lesser extent, valuables) and personal guarantees, typically provided by wealthier persons (others’ land). 25 Since the former represents an included instrument, we rely upon the latter as an exogenous one. In particular, we use the number of the household head and spouse’s relatives owning land as a proxy of the likelihood of obtaining co-signed guarantees. In our data (see again Table 5) bank borrowers have indeed the highest number of non-close 26 relatives owning land compared to other borrowers. Again, the possibility that the instrument captures household wealth is minimized by including land ownership and other inherited assets as controls. An additional issue originates from the fact that relatives owning land may represent important sources of transfers (Pitt & Khandker, 1998), which are likely to affect investment directly inducing violation of the exclusion restrictions. To avoid this possibility we control for transfers in cash, food, and other commodities re-

Table 6. Credit market participation: FIRST STAGE with village fixed-effects Group Informal lending (MFIs) lending Targeta Spouse: distance from familya No. of relatives own landa

Table 5. Descriptive statistics: instruments Variable

ceived by the household in the last year (see Table 4 for descriptive statistics). 27 The spatial distance between in-law relatives is used as an instrument for informal lending. The literature dealing with the topic of distance and inter-household financial support in South Asia suggests that there is a positive relationship between the two. Park (2006), for example, finds that credit transactions more frequently occur between relatives living either in different clusters or in distant villages, rather than between neighbors. An explanation of this phenomenon can be found in Caldwell et al. (1986), and Rosenzweig (1988), who suggest that a higher spatial distance between households reduces the correlation of incomes, mostly in the agricultural sector where climatic conditions cause a high geographical variance of returns. A longer distance should, therefore, enhance households’ advantage to engage in mutual credit transactions for income-smoothing purposes, 28 since the further the households, the lower the likelihood of being hit by the same natural event. In particular, Rosenzweig and Stark (1989) find that this kind of lending pattern mainly occurs among relatives in-law. 29 Consequently, we use the distance of the household head’ spouse from her birthplace as a measure of the flow of informal loans accruing to the household.

Mean

Std. Dev.

Min

Max

Borrowers from MFIs Target Spouse: distance from family No. of relatives own land Dowry Marriage gifts to spouse

0.80 12.11 3.10 1,220 195

0.40 33.18 4.66 3,126 886

0 0 0 0 0

1 400 36 37,500 10,500

Borrowers from informal sectorb Target Spouse: distance from family No.of relatives own land Dowry Marriage gifts to spouse

0.60 21.72 3.18 2,198 382

0.46 55.58 4.29 4,823 1,392

0 0 0 0 0 0

1 450 21 35,500 20,500 1 478 15 21,428 40,000 1 450 27 60,000 45,000

Dowryb

a

Borrowers from banksc Target Spouse: distance from family No. of relatives own land Dowry Marriage gifts to spouse

0.53 21.02 3.95 1,461 1,671

0.51 77.41 2.43 4,093 6,820

0 0 0 0 0 0

Non-borrowersd Target Spouse: distance from family No. of relatives own land Dowry Marriage gifts to spouse

0.70 10.44 3.27 1,353 380

0.44 33.33 4.60 4,193 2,184

0 0 0 0 0 0

Marriage gifts to spouseb HH head: education Religion Area cultivated Fixed rent Sharecropper Partial R2

Note: 46 households borrowing from multiple sources. a Obs. 297. b Obs. 111. c Obs. 40. d Households not borrowing in the twelve months preceding the survey, Obs. 1,369.

76.365*** (25.221) 0.223 (0.244) 1.771 (4.571) 0.075 (0.237) 0.316 (0.411) 0.001 (0.003) 0.022 (0.040) 0.004 (0.012) 0.047* (0.027) 0.038 (0.023) 0.19

3.040 (18.164) 0.356** (0.176) 1.422 (3.292) 0.048** (0.023) 0.126 (0.296) 0.005** (0.002) 0.078*** (0.028) 0.009 (0.009) 0.011 (0.019) 0.039** (0.017) 0.15

Bank lending 0.894 (10.281) 0.161 (0.099) 3.366* (1.864) 0.003 (0.097) 0.589*** (0.168) 0.001 (0.001) 0.011 (0.016) 0.008* (0.005) 0.025** (0.011) 0.007 (0.010) 0.17

Joint significance of all parameters in simultaneous estimation equations (1)–(3): v2(119) = 301.16 (P-val = 0.003) Joint significance of instruments in simultaneous estimation equations (1)–(3): v2(15) = 34.12 (P-val = 0.000) Weak identification test: F-statistic 32.78 Anderson canon. corr. LR statistic: v2(3) 167.923 Least squares estimates; obs.: 1,798; robust standard errors in parentheses. Other regressors: age of HH head and spouse, spouse’s education, parents of the HH head and spouse living in the HH, gender of the HH head, Nr. HH members, feminine ratio, Nr. relatives alive, illness, medical expenditure, non-working days due to illness, land owned, house, transport, cash and non-cash transfers, household head conducts the same activity of his/her father. a Parameters multiplied by 1000. b Parameters multiplied by 10000. * Significant at 10%. ** Significant at 5%. *** Significant at 1%.

MICROFINANCE AND INVESTMENT: A COMPARISON WITH BANK AND INFORMAL LENDING

889

Table 7. Investment in working capital: agricultural activities: SECOND STAGE with village fixed-effects Full distribution Group lending (MFIs) Informal lending Bank lending Loan amountc Landd Transfers: cashe Father: same activityf R2 Hausman–D–Wg v2(3) Hansen-J v2(2)

90% of obs.a

70% of obs.b

2SLS

2SLS

2SLS

OLS

2011.023 (2063.347) 1064.533 (1065.704) 2391.461* (1222.182) 0.010* (0.006) 94.100** (46.785) 0.043 (0.031) 242.748** (97.160)

1253.929 (885.358) 1029.210 (1110.217) 1414.608 (1902.079) 0.003 (0.005) 33.059 (26.483) 0.066** (0.028) 183.248*** (61.284)

158.891 (131.388) 1009.266 (1617.42) 305.359** (138.447) 0.000 (0.001) 8.507 (6.707) 0.025 (0.016) 26.313*** (9.933)

19.779 (79.382) 613.294* (353.058) 172.169** (83.544) 0.000 (0.001) 1.773 (4.863) 0.011 (0.015) 19.086*** (4.620)

0.72 8.19 (P-val = 0.042) 1.37 (P-val = 0.503)

0.73 26.83 (P-val = 0.000) 0.76 (P-val = 0.681)

0.46 1.15 (P-val = 0.764) 1.14 (P-val = 0.589)

0.46

Robust standard errors in parentheses. All variables related to those reported at points (c)–(f) are rarely significant. Full estimation report is available upon request. Other regressors: age and education of HH head and spouse, religion, gender of the HH head, Nr. HH members, feminine ratio, Nr. relatives alive, illness, medical expenditure, nonworking days due to illness, total area cultivated, land tenure. * Significant at 10%. ** Significant at 5%. *** Significant at 1%. Full distribution: 1,798 obs. a Below 90th percentile, obs.: 1,618. b Below 70th percentile, obs.: 1,258. c Other measurable determinants of credit included: interest rate and collateral d Other measures of wealth included: house and transport ownership. e Non-cash transfers also included f Parents of the HH head and spouse living in the HH also included. g Computed forcing non-robust standard errors. Table 8. Investment in working capital: non-agricultural activities: SECOND STAGE with village fixed-effects Full distribution Group lending (MFIs) Informal lending Bank lending Loan amountc Landd Transfers: cashe Father: same activityf R2 Hausman–D–Wg v2(3) Hansen-J v2(2)

90% of obs.a

70% of obs.b

2SLS

2SLS

2SLS

OLS

15659.929 (12136.07) 10246.625 (16773.968) 60534.664 (73082.03) 0.043* (0.022) 745.588 (553.627) 0.390 (0.350) 282.002 (362.352) 0.18 12.85 (P-val = 0.004) 0.49 (P-val = 0.782)

224.853*** (65.176) -289.50 (318.204) 166.672 (103.161) 0.000 (0.000) 2.000 (1.437) 0.000 (0.001) 29.919*** (5.173) 0.18 6.43 (P-val = 0.092) 1.22 (P-val = 0.541)

25.236*** (5.879) 20.36 (93.32) 17.098 (10.678) 0.000 (0.000) 0.280*** (0.105) 0.000* (0.000) 3.662*** (0.471) 0.28 12.55 (P-val = 0.005) 0.07 (P-val = 0.961)

31.644*** (5.504) 76.467 (53.840) 29.703** (12.081) 0.000 (0.000) 0.125* (0.071) 0.000** (0.000) 4.643*** (0.356) 0.29

Robust standard errors in parentheses. All variables related to those reported at points (c)–(f) are rarely significant. Full estimation report is available upon request. Other regressors: age and education of HH head and spouse, religion, gender of the HH head, Nr. HH members, feminine ratio, Nr. relatives alive, illness, medical expenditure, nonworking days due to illness, total area cultivated, land tenure. * Significant at 10%. ** Significant at 5%. *** Significant at 1%. Full distribution: 1,798 obs. a Below 90th percentile, obs.: 1,618. b Below 70th percentile, obs.: 1,258. c Other measurable determinants of credit included: interest rate and collateral. d Other measures of wealth included: house and transport ownership. e Non-cash transfers also included. f Parents of the HH head and spouse living in the HH also included. g Computed forcing non-robust standard errors.

890

WORLD DEVELOPMENT

Statistics reported in Table 5 provide evidence of a positive relationship between the distance of spouses from their families and participation to the credit market. Correlation between the two variables is higher for both households borrowing from informal sources and banks, compared to those participating to MF programs or not borrowing. The variance, though, is lower for informal transactions, which should be indicative of the fact that the instrument is more suitable for this type of credit. It might be possible, however, that distance between households was part of a marital agreement serving to mitigate income risk (Rosenzweig & Stark, 1989). Typically in rural contexts such agreements are carried out by parents of grooms and brides. Therefore, problems may arise if, at the time of the survey, these lived sufficiently close to the household as to take part to investment decisions. In this case, in fact, parents’ unmeasurable characteristics may interact with both credit transactions—through marriage arrangements—and investment choices. In order to mitigate this effect we include a set of dummy variables controlling for both the presence of parents in the household and the household head’s engagement in his/her father’s activity (Table 4). 30 Dowries are also contemplated among instruments for both informal and bank lending. In general, a dowry is intended as the money, goods, or estate that a woman brings to her husband in marriage. To the purposes of this paper, however, it is useful considering also other forms of payments that took place at the time of marriage of the household head and spouse. The same culture, for example, may simultaneously practice dowries and other transfers, such as gifts to the bride. 31 Typically, dowries and other gifts are not decided

by grooms and brides themselves, but are instead settled by their families. It is a matter of culture that these are passed down the family line as something personal that others cannot easily claim (Anderson, 2007) and it is very infrequent that they are sold to buy productive inputs. This provides an argument in favor of the exogeneity of such variables since their effect on investment is not likely to be direct but may instead pass through credit. In fact, as Anderson points out, both dowries and marriage gifts to the spouse are often pledged as collateral against loans from moneylenders (see also Aleem, 1990; Bhattacharyya, 2005). It is also frequent that women are financed by formal institutions, normally cooperative banks—which are part of the banking sector in our sample—acting as pawnbrokers and accepting valuables as collateral. 32 From statistics in Table 5 it emerges that more generous dowries are associated to a higher probability of accessing informal loans, while marriage gifts received by the spouse are considerably higher for bank borrowers. As in the case of the other instruments, marriage transfers can be considered exogenous to investment, provided that suitable controls are included to avoid violation of exclusion restrictions. In particular, we refer to measures of wealth, transfers, and other variables capturing relatives’ interference with household investment decisions, as discussed above. All instruments and controls mentioned in this section are listed in Table 12 in the Appendix. In conclusion, it is worth stressing that despite the inclusion of several controls, there is sill possibility that the instruments may have some direct impact on investment. For this reason, at the end of the following section we provide robustness checking through regressions of investment on instruments

Table 9. Investment in fixed assets after twelve months: agricultural activities: SECOND STAGE with village fixed-effects Full distribution

Group lending (MFIs) Informal lending Bank lending Loan amountc Landd Transfers: cashe Father: same activityf R2 Hausman–D–Wg v2(3) Hansen-J v2(2)

90% of obs.a

80% of obs.b

2SLS

2SLS

2SLS

OLS

2032.27 (1846.737) 6783.825** (3173.856) 14314.559 (23610.896) 0.003 (0.010) 28.707 (39.401) 0.063 (0.047) 356.047** (174.058)

1804.435 (1634.776) 4730.050* (2808.193) 2597.857 (3192.838) 0.002 (0.009) 22.656 (34.874) 0.055 (0.041) 281.847* (153.987)

1776.115 (1977.570) 3473.591* (1884.595) 1595.794 (2373.350) 0.002 (0.005) 36.623* (20.986) 0.028 (0.026) 300.609*** (90.656)

1293.399 (1139.227) 1489.511* (887.498) 2044.014 (2971.352) 0.023 (0.043) 32.227** (16.184) 0.012 (0.009) 149.093*** (57.774)

0.10 0.40 (P-val = 0.940) 1.46 (P-val=0.4817

0.12 8.92 (P-val = 0.030) 0.29 (P-val = 0.862)

0.13 13.00 (P-val = 0.004) 0.26 (P-val = 0.874)

0.13

Robust standard errors in parentheses. All variables related to those reported at points (c)–(f) are rarely significant. Full estimation report is available upon request. Other regressors: age and education of HH head and spouse, religion, gender of the HH head, Nr. HH members, feminine ratio, Nr. relatives alive, illness, medical expenditure, nonworking days due to illness, total area cultivated, land tenure. * Significant at 10%. ** Significant at 5%. *** Significant at 1%. Full distribution: 1,798 obs. a Ten per cent of the tails dropped, obs.: 1,618. b Twenty per cent of the tails dropped, obs.: 1,438. c Other measurable determinants of credit included: interest rate and collateral. d Other measures of wealth included: house and transport ownership. e Non-cash transfers also included. f Parents of the HH head and spouse living in the HH also included. g Computed forcing non-robust standard errors.

MICROFINANCE AND INVESTMENT: A COMPARISON WITH BANK AND INFORMAL LENDING

and controls. In any case, however, since instrumentation is still open to question, a certain degree of caution is used in the following section while interpreting the results. 4. RESULTS In this section we present estimates of Eqns. (1)–(5). Results of the first stage Eqns. (1)–(3) are reported in Table 6. Tables 7–10, instead, refer to the second stage Eqn. (4) and (5) for investment in working capital and in fixed assets, respectively. OLS parameters are also reported for the main regressions for comparison. (a) Credit market participation Results reported in Table 6 show that instruments are significant in determining credit access. As expected, participation in MF programs considerably increases with the eligibility status measured by the ownership of less than half an acre of land, a result that is in line with the previous empirical literature. Next, having more landed relatives gives a higher probability of accessing bank lending. Hence it seems that when a project is planned, households are more likely to resort to MFIs or banks in order to obtain funds, and they enter these markets based on land status and land collateral availability. Instead, the non-significance of both the eligibility threshold and the number of relatives owning land in the informal market equation indicates that this type of credit behaves differently from the other two.

891

In particular, as previously found in the literature (see above), the parameter measuring the effect of distance from the spouse’s original birthplace is positive and significant in the equation of informal lending, providing support to the fact that in-law relatives located far from the household are likely to extend a large portion of credit and smooth income and investment. Dowries also enter the equation of informal lending with a positive sign. Conditional on other measures of household’s wealth, this possibly suggests that marriage valuables are used as collateral in informal credit transactions. Furthermore, the significance of marriage gifts to the spouse may also be indicative of the fact that women pledge their own assets in order to obtain cooperative credit. Econometric tests may provide some idea as to whether these instruments are relevant. First, Hausman–Durbin–Wu statistics reported in Tables 7–10 often indicate that there is endogeneity in the relationships we are estimating conditional on the controls. We choose to use 2SLS estimation techniques also when tests reject the hypothesis that OLS and 2SLS do not provide substantially different output. This is the safest approach one could adopt, since treating a behavior as endogenous when it is in fact exogenous still yields consistent, although less efficient, estimates (Pitt, 1999). The Hansen-J test for overidentifying restrictions virtually every time suggests that the instruments are valid. The Kleibergen–Paap Fstatistic reported in Table 6 for the first stage estimates also indicates that instruments are likely to be non-weak. In conclusion, land related variables, including tenure (fixed rent), and the village fixed effects are the most significant factors explaining participation to group lending and bank

Table 10. Investment in fixed assets after twelve months: non-agricultural activities: SECOND STAGE with village fixed-effects Full distribution

Group lending (MFIs) Informal lending Bank lending Loan amountc Landd Transfers: cashe Father: same activityf R2 Hausman–D–Wg v2(3) Hansen-J v2(2)

90% of obs.a

80% of obs.b

2SLS

2SLS

2SLS

OLS

24507.309* (14153.903) 17027.124 (33295.945) 13735.815 (16388.338) 0.251 (0.240) 353.584 (320.630) 0.168 (0.581) 4241.267*** (1012.939)

5800.231 (4689.378) 3399.026 (2201.167) 2523.209 (15690.349) 0.015 (0.012) 159.536 (277.821) 0.007 (0.035) 3080.130*** (711.609)

3513.601* (2130.521) 3654.011 (3555.580) 2795.600 (3832.159) 0.013 (0.011) 86.242 (55.935) 0.015 (0.011) 1389.622*** (193.635)

4846.695** (2088.774) 1897.052 (2010.292) 3913.304 (4035.651) 0.055 (0.038) 58.683 (39.785) 0.006 (0.010) 1148.477*** (111.843)

0.18 9.41 (P-val = 0.024) 0.12 (P-val = 0.937)

0.19 8.60 (P-val = 0.035) 0.05 (P-val = 0.973)

0.20 15.64 (P-val = 0.001) 0.20 (P-val = 0.900)

0.20

Robust standard errors in parentheses. All variables related to those reported at points (c)–(f) are rarely significant. Full estimation report is available upon request. Other regressors: age and education of HH head and spouse, religion, gender of the HH head, Nr. HH members, feminine ratio, Nr. relatives alive, illness, medical expenditure, nonworking days due to illness, total area cultivated, land tenure. * Significant at 10%. ** Significant at 5%. *** Significant at 1%. Full distribution: 1,798 obs. a Ten per cent of the tails dropped, obs.: 1,618. b Twenty per cent of the tails dropped, obs.: 1,438. c Other measurable determinants of credit included: interest rate and collateral. d Other measures of wealth included: house and transport ownership. e Non-cash transfers also included. f Parents of the HH head and spouse living in the HH also included. g Computed forcing non-robust standard errors.

892

WORLD DEVELOPMENT

lending. Education, religion, and the sharecropper status of the household head (on the latter see Braverman & Stiglitz, 1982) seem instead to be good drivers of the access to informal lending, as it turns out from parameters reported in Table 6. (b) Investment Estimates of the impact of the three different types of credit on investment in working capital are reported in Tables 7 (agricultural activities) and 8 (non-agricultural activities), 33 whereas estimates of the impact of credit on investment in fixed assets are reported in Tables 9 (agricultural activities) and 10 (non-agricultural activities). Moving from the left to the right columns of Tables 7 and 8 we trim higher levels of expenditure corresponding to the left tail of the working capital distribution. According to the discussion in the previous section, this should manage the problem of outliers. Moreover, dropping the left tail of the distribution also allows isolating the effect of credit on the smallest activities—presumably associated to the poorest households—in the last column on the right. Similarly, moving from the left to the right columns of Tables 9 and 10 implies concentrating on observations involving less volatile investment activities located in the center of the distribution of fixed assets, that is, in the area around zero. This may also isolate

poorer households, since their asset buying/dismissal activity is likely to be smoother as compared to that of richer families. For each regression we report parameters associated to the main variables. 34 In particular, we display results concerning predicted participation to each type of lending transaction, which should account for non-measurable determinants of contracts, as discussed above. We also report the parameters associated to the amount borrowed and to the main variables which are critical to the instrumentation argument. 35 Remaining controls are listed in the bottom of each table. As for working capital, we find that farmers’ investment is positively associated to bank credit (Table 7), while MF is likely to be more beneficial for non-farmers (Table 8). In particular, estimated parameters for bank lending are significant for farmers who reach either very high levels of investment or relatively low ones. We interpret the former evidence as being a collateral/wealth effect, namely larger activities are managed by richer households who are also endowed with more valuable collateral, while the latter possibly reveals the practice of pledging valuables in cooperative banks, as extensively discussed above. Parameters associated to MF become instead more significant for observations corresponding to mid-low non-agricultural expenditure levels, that is, after dropping 10% of the distribution corresponding to the largest activities.

Table 11. Significance of instruments in the second stage equations Investment in working capital Without controlsa Target Spouse: distance from family No. of relatives own land Dowry Marriage gifts to spouse With controlsb Target Spouse: distance from family No. of relatives own land Dowry Marriage gifts to spouse Land Transfers: cash Father: same activity

Investment in fixed assets

Agricultural activities

Non-agricultural act.

Agricultural activities

Non-agricultural act.

172.003 (138.708) 1.250 (0.998) 9.392 (9.065) 0.021* (0.012) 0.003 0.004)

714.126 (929.490) 0.256 (1.241) 5.750 (36.150) 0.014 (0.049) 0.045 0.043)

37.315 (175.668) 1.444 (1.243) 7.153 (21.783) 0.019 (0.015) 0.015 0.023)

2,304.669* (1,204.774) 3.823 (8.001) -113.739 (98.975) 0.018 (0.066) 0.042 0.105)

76.409 (127.120) 1.091 (1.002) 10.378 (8.826) 0.013 (0.009) 0.007 (0.014) 138.116*** (48.565) 0.058** (0.027) 122.369** (50.554)

1,187.496 (1,019.296) 0.214 (1.097) 5.215 (36.930) 0.014 (0.037) 0.028 (0.028) 643.806 (482.257) 0.227 (0.206) 205.704 (207.620)

41.614 (173.201) 1.130 (1.251) 9.025 (21.755) 0.018 (0.015) 0.015 (0.023) 25.603 (33.233) 0.049 (0.035) 193.435** (93.468)

1,765.158 (1,294.166) 4.576 (8.003) 116.819 (97.971) 0.025 (0.060) 0.013 (0.100) 199.921 (239.393) 0.109 (0.379) 2,648.067*** (415.189)

Observations: 1798; robust standard errors in parentheses; fixed-effects regressions. * Significant at 10%. ** Significant at 5%. *** Significant at 1%. a All the variables used in second stage (Tables 7–10) except credit, land, house, transport, parents of HH head and spouse living in the HH, and transfers have been included in the regressions. b All the variables used in second stage (Tables 7–10) except credit have been included in the regressions.

MICROFINANCE AND INVESTMENT: A COMPARISON WITH BANK AND INFORMAL LENDING

Fixed assets almost behave like operating costs. In this case, informal lending rather than bank lending is associated to a higher investment in farming activities (Table 9), while MF still performs better in the non-farming sector (Table 10). It is interesting to note that within each sector the significant parameters associated to fixed assets are larger than those of working capital, particularly with regard to MF. Furthermore, in the case of fixed assets there is no clear pattern that links a more intensive use of credit with the dimension of borrowers’ activity. Measurable components of credit, such as the amount borrowed, interest rates, and collateral requirements, are rarely significant. The amount seems to matter for very large expenditure only, whereas there is no clear-cut pattern for the incidence of the other credit conditions. Control variables that have been exploited to curb the possible violation of exclusion restrictions have the expected signs and are often significant. Among other variables associated with investment— which are not reported in the tables—there is evidence of a strong relationship between agricultural expenditure and other features connected to farming, such as cultivated land and land tenure. Finally, the traditional lack of incentives in sharecropping seems evident in agriculture, but only for larger activities, otherwise sharecroppers behave well. (c) Robustness to exclusion restrictions In order to check the robustness of instruments to exclusion restrictions we perform regressions without—previously instrumented—credit variables included in the specification but with instruments substituted in their place. In the upper part of Table 11, parameters refer to the specification not including controls for household wealth, transfers of various types, and variables connected to the possibility that the head of the household and spouse’s parents may interfere with investment activities. The lower part of the table shows parameters of the regressions including these controls (only the most significant ones are reported). Without the presence of controls, the instruments which in principle appear to suffer more from violation of exclusion restrictions seem to be both the eligibility rule for MF programs and dowries. The former may have several explanations. In principle, one should expect either a positive or negative sign of the parameter. For example, target households may invest less because they are poorer or the landless are more likely to invest in activities other than agriculture. Our interpretation of the results is that the latter effect prevails over the former. As for dowries, we interpret their significant positive relationship with investment as a wealth effect, since higher dowries may correspond to both richer households and richer extended families providing more transfers to the household. However, including all the controls, and particularly all the measures of assets and transfers—lower part of the table—the parameters associated with the eligibility rule and dowries become not significant, suggesting that part of the significance seen above is likely to pass through the credit channel. 5. CONCLUSIONS A considerable number of works dealing with the impact of MF programs have been written so far. Many assess the success of these programs on several household behaviors such as increasing consumption, labor supply, and children school enrollment. This paper, in particular, concentrates on invest-

893

ment in productive activities since this is likely to represent an important way toward capital accumulation and growth for households in less developed countries. Using data from a World Bank survey documenting loans from different types of lenders in Bangladesh, we investigate which kind of lending—micro-lending, informal lending, and bank lending—is better suited to promote investment in either farming or non-farming activities. Distinguishing between two sectors and comparing three types of credit is thus important for the purpose of verifying whether MF contracts are equally suitable for all sectors of the economy. In fact, standardized lending agreements, which are typical of MF programs, have been recently criticized as not being suited to match the needs of the agricultural production. In particular, borrowers seem forced into a situation where they have to produce rapidly in order to meet a maturity date that has perhaps been fixed too close to the date when the loan was granted. This, besides being an empirically controversial way to foster financial education (see, e.g., Field & Pande, 2008; McIntosh, 2008) may push farmers, whose production cycle is longer than in other activities, toward more flexible—but sometimes more expensive—credit channels, such as the informal one. Results from the empirical analysis show that there is no superior credit agreement in terms of efficiency, since a positive relationship between all types of credit and investment is observed. However, there are interesting differences which are likely to support the hypothesis of a low suitability of MF programs for some sectors. In particular, we find that households whose members belong to a group lending program invest more in non-agricultural inputs compared both to households who are borrowing from other sources and non-borrowers. The situation is reversed in agriculture, where it emerges that only households borrowing from informal sources and banks invest more. A key point of the paper is that the emerging evidence is not ascribable to measurable features of contracts such as the amount borrowed, interest rates, and collateral requirements. This makes it more likely that the explanation of our results rests in the repayment system required by different lenders. The observed pattern, however, could also be the outcome of the MFIs’ bias in favor of non-farming activities since the risk of default related to climate conditions is lower and the production cycle is considerably shorter than in agriculture, so that repayments are made effortlessly on a regular basis. The eligibility criterion based on land ownership may represent another factor excluding a number of farmers—who are more likely to be endowed with land— from the MF network. Also in this case, however, it is possible to interpret things in the opposite way, since it may be that collateral availability allows farmers to access credit channels different from MF. Other features, such as “joint” compared to “individual” liability systems, “peer” versus “lender” monitoring technologies, and sanctions, can be considered some of the factors that account for the evidence we observe. Non-credit services (McKernan, 2002), such as financial education and healthprograms associated to credit, can also help stimulating distressed households’ willingness to care about the future, and by this, foster investment. However, we do not strictly rely on these elements to justify our results since, as discussed in Section 1, there is apparently no specific reason why these should favor non-agricultural activities more than farming. Finally, while tackling the issues of endogeneity between credit variables and investment, we obtain interesting information on credit market selection mechanisms. We find that

894

WORLD DEVELOPMENT

the principle of targeting landless households seems enforced by MFIs, while the access to bank lending depends on collateral—land and other valuables—endowment. Besides the possible use of dowries as collateral in some type of infor-

mal credit transactions, the informal channel seems working for households linked by long-distance in-law relationships who benefit from lending as a means to reduce income covariance.

NOTES 1. There is evidence that inviting farmers to plant hybrids, for instance, is another measure adopted by some MFIs in this regard. 2. Madajewicz (2003a and b) and McKernan (2002), while analyzing the impact of MF on business profits, have also been dealing with investment and credit, but they either ignored expenditure in fixed assets or disregarded other credit sources. 3. It is no surprise that we find this recommendation in the Grameen Bank program, which states: “We shall grow vegetables all the year round. We shall eat plenty of them and sell the surplus.” 4. In the absence of this evidence, instead, we would not be in the position to infer that MF is not suitable for these borrowers. For example, they may not need loans because of higher self-financing capabilities. 5. The Bangladesh Rural Development Board operates by organizing small and marginal farmers, asset-less men and women in cooperative societies and informal groups (http://www.brdb.gov.bd/general_Info.htm). 6. Although the problem of missing data is particularly severe, regressions have also been performed on the three-rounds panel using household fixed-effects. In terms of significance of the parameters results do not bring additional information compared to the cross-sectional setup. 7. A question on the difference between the ownership of fixed assets is actually present in the survey, but these data cannot be used due to the considerable amount of missing information. Moreover, this measure is available for farmers only, and over a period that is not clearly specified. 8. The period goes from January 1991 to December 1991. 9. Due to the fact that the production cycle is likely to be shorter than one year in non-agricultural activities, we observe that non-farmers frequently buy raw materials during a one-year period. Hence, we consider operating costs for one average cycle only. The assumption lies in the fact that if credit is available, it is used to immediately raise input demand and not saved for further cycles, since proceeds generated by the loan are more likely to be used for that purpose (i.e., rolled over subsequent cycles). This eliminates overlapping and possible overestimates of variable input expenditure. 10. Note that, in any case, if we were aggregating investment the estimated parameters of the aggregate equations would result in the sum of the parameters separately estimated in two different equations, one working capital and one for fixed assets. We omit this redundant step. Results are available on request. 11. Data not enable a clear separation between assets dismissal and missing answers (some households declared a positive value of assets in the first round of the survey and then did not answer this question in the following rounds). This problem is partially managed through the percentile analysis carried out in the empirical section. 12. Anti-usury ceilings were 16% at the time of the survey. A handful of reports of slightly higher rates in MF may include fees.

13. Such high interest rates (up to 240% in our sample) may either reflect moneylender’s usurious behavior (see, e.g., Basu, 1984; Rahman, 1979; Blitz & Long, 1965, and Bhaduri, 1977) or be a result of high monitoring costs as argued by Aleem (1990). See also Timberg and Aiyar (1984) for a discussion. 14. The actual duration (including delays) cannot be computed since most of the loans we consider are still outstanding at the time the household is surveyed. 15. In Table 1, average investment in fixed assets apparently looks lower than working capital expenditure. This is due to the fact that a substantially low number of self-employed households invest in fixed assets, while all of them make use of working capital, conditional on the fact of belonging to a specific sector. By looking at individual households recording a value of investment in fixed assets which is greater than zero the average expenditure is considerably higher. 16. Note that we cannot use the ratio of investment to the amount of borrowed funds as a dependent variable since this measure is available only for entrepreneurs that are actually borrowing. However, truncating the sample at positive values of borrowing would cause a selection bias (Heckman, 1976). 17. This variable cannot represent an instrument for MF since besides affecting access to MF, females may also have a different attitude toward risk and investment. There is, however, no consensus in the literature on this point (see, e.g., Schubert, Brown, Gysler, & Brachinger, 1999, for a discussion). 18. For the specification of the system (1)–(5) see Park (1974) and Khazzoom (1976). 19. These assumptions are also a consequence of testing the hypothesis of no correlation between the residuals of the system (1)–(5). Tests suggest that the error terms are correlated between credit equations, and particularly between group lending and bank lending. The correlation between credit and investment is instead negligible, while it is weak between Eqns. (4) and (5). 20. On Least Trimmed Squares techniques see, for example, Rousseeuw (1997) and Bassett (1991). For an overview of the properties of Least Trimmed Squares in linear and nonlinear regression see Cı`zek and Vı`sek (2000), and Stromberg (1993). 21. Zeros cannot be trimmed all at once without incurring a selectionbias (Heckman, 1976). 22. Tails are dropped in such a way to keep a stable ratio of non-zero observations in each tail. 23. Due to the possible bias embedded in the first stage Linear Probability model (see, e.g., Greene, 2000), we also performed the first stage using a Multinomial Logit model. The predicted probabilities of borrowing from each type of lender have then been plugged in the secondstage Eqns. (4) and (5). Estimates do not provide substantial changes as compared to 2SLS. In particular, correcting for negative predicted values

MICROFINANCE AND INVESTMENT: A COMPARISON WITH BANK AND INFORMAL LENDING of the probability of accessing each credit market in the 2SLS estimation procedure returns second-stage parameters which are closer to those obtained by the use of the Multinomial Logit. 24. According to the discussion in the previous section it is plausible that controlling for land removes a considerable part of wealth from the error term. This should enhance the argument in favor of the validity of the instrument. 25. Besides agricultural land (required in 29% of bank loans), we find that respondents declare that banks also accept guarantees from the land registration book, such as barga certificates (see Bhattacharyya, 2005, for more details) or other property documents (35% of cases). Other buildings are required as collateral in only 4% of the cases. 26. Note that we do not account for close relatives—parents, siblings, sons, and daughters living outside the household—since these are more likely to provide direct contribution to household investment and, therefore, violate exclusion restrictions. 27. Past transfers may not account for the incentive effect stemming from being assured future assistance from wealthy relatives, so that the latter remains embedded in the instrument. Hence, the problem of violating the exclusion requirement may not be completely ruled out unless there is positive correlation between past and potential transfers. This, however, seems a reasonable hypothesis since a household relies on the amount of aid received in the past in order to calculate the probability of a relative’s helping in the future. Therefore, our choice of controlling for past transfers appears sufficient to preserve the quality of the instrument. 28. In particular, Park finds that credit is mainly used for investment smoothing purposes—especially to buy cattle, which is the main buffer against hardships—while the transfers in nature occurring among neighbors are instead used for consumption.

895

29. This large flow of lending transactions seems consistent with statistics reporting that approximately a very high portion of rural-to-rural migration in East Asia is represented by women who move for marriage reasons. For example, according to the 2001 Indian Population Census, 156 millions females against 2 millions males move from their birthplace for marriage reasons. 30. Furthermore, controlling for land ownership is again crucial since land has been recognized as being one the major determinants of marriage agreements. This is driven by an assortative matching argument, since landed (richer) families tend to mutually insure more frequently (Rosenzweig & Stark, 1989). 31. Bride prices and Groom prices are other forms of transfers which we do not consider since they accrue to brides’ and grooms’ families. For more details see Anderson (2007). 32. A detailed description of this practice is provided, for example, by Bouman and Houtman (1988) who document the experience of the People’s Bank in India. 33. Note that if we were aggregating investment, the estimated parameters of the aggregate equations would result in the sum of the parameters separately estimated in two different equations, one for farmers and one for non-farmers. We omit this redundant step. Results are available on request. 34. Full results available on request. 35. In particular, we report the coefficient of land as a representative measure of wealth, while we omit house and transport ownership since their coefficients are rarely significant. We do the same for transfers in cash, neglecting transfers in food and other goods. For the same reasons, only the presence of the household head father in the house is displayed while leaving out his/her mother and the spouse’s parents.

REFERENCES Ahlin, C., & Jiang, N. (2008). Can micro-credit bring development? Journal of Development Economics, 86(1), 1–21. Aleem, I. (1990). Imperfect information, screening and the costs of informal lending: A study of a rural credit market in Pakistan. World Bank Economic Review, 4(3), 329–349. Anderson, S. (2007). The economics of dowry and brideprice. Journal of Economic Perspectives, 21(4), 151–174. Armendariz, B. (1999). On the design of a credit agreement with peer monitoring. Journal of Development Economics, 60(1), 79–104. Armendariz, B., & Morduch, J. (2005). The economics of microfinance. Cambridge, MA: MIT Press. Ba˚ge, L. (2004). Savings: the forgotten half of microfinance, Statement delivered by the president of the International Fund for Agricultural Development. . Banerjee, A. V., Besley, T., & Guinnane, T. (1994). Thy neighbor’s keeper: The design of a credit cooperative with theory and a test. The Quarterly Journal of Economics, 109(2), 491–515. Barua, D., & Dowla, A. (2006). The poor always pay back: The Grameen II story. Bloomfield, CT: Kumarian Press Inc.. Bassett, G. W. (1991). Equivariant, monotonic, 50% breakdown estimators. American Statistician, 45(2), 135–137. Bhattacharyya, S. (2005). Interest rates, collateral and (de-)interlinkage: A micro-study of rural credit in West Bengal. Cambridge Journal of Economics, 29(3), 439–462. Basu, A., Blavy R., & Yulek M. (2004). Microfinance in Africa: Experience and lessons from selected African Countries. IMF Working Paper 04/ 174.

Basu, K. (1984). Implicit interest rates, usury and isolation in backward agriculture. Cambridge Journal of Economics, 8(2), 145–159. Bouman, F. J. A., & Houtman, R. (1988). Pawnbroking as an instrument of rural banking in the third world. Economic Development and Cultural Change, 37(1), 69–89. Besley, T., & Coate, S. (1995). Group lending, repayment incentives and social collateral. Journal of Development Economics, 46(1), 1–18. Bhaduri, A. (1977). On the formation of usurious interest rates in backward agriculture. Cambridge Journal of Economics, 1(4), 341– 352. Binswanger, H. P., & Rosenzweig, M. R. (1986). Behavioural and material determinants of production relations in agriculture. Journal of Development Studies, 22(2), 503–539. Blitz, R. C., & Long, M. F. (1965). The economics of usury regulation. Journal of Political Economy, 73(6), 608–619. Brandsma, J., & Chaouali R. (1998). Making microfinance work in the Middle East and North Africa. World Bank Discussion Paper, Washington DC. Braverman, A., & Stiglitz, J. E. (1982). Sharecropping and the interlinking of agrarian markets. American Economic Review, 72(4), 695–715. Caldwell, J. C., Reddy, P. H., & Caldwell, P. (1986). Periodic high risk as a cause of fertility decline in a changing rural environment: Survival strategies in the 1980–1983 south Indian drought. Economic Development and Cultural Change, 34(4), 677–701. Christen, R. P., & Pearce, D. (2005). Emerging lessons in agricultural microfinance: Selected case studies. CGAP Occasional Paper, 11, Washington, D.C.

896

WORLD DEVELOPMENT

Cı`zek, P., & Vı`sek, J. A. (2000). Least trimmed squares. In W. Hardle, Z. Hlavka, & S. Klinke (Eds.). XploRe application guide (pp. 49–64). Heidelberg: Springer. Eswaran, M., & Kotwal, A. (1985). A theory of contractual structure in agriculture. American Economic Review, 75(3), 352–367. Field, E., & Pande, R. (2008). Repayment frequency and default in microfinance: Evidence from India. Journal of European Economic Association Paper and Proceedings, 6(2–3), 501–509. Ghatak, M., & Guinanne, T. N. (1999). The economics of lending with joint liability: Theory and practice. Journal of Development Economics, 60, 195–228. Greene, W. (2000). Econometric analysis (4th ed.). New York: Prentice Hall. Heckman, J. (1976). The common structure of statistical models of truncation, sample selection, and limited dependent variables and a simple estimator for such models. Annals of Economic and Social Measurement, 5(4), 120–137. Heckman, J. (1990). Varieties of selection bias. American Economic Review, 80(2), 313–318. Hermes, N., Lensink, R., & Mehrteab, H. T. (2005). Peer monitoring, social ties and moral hazard in group lending programs: Evidence from Eritrea. World Development, 33(1), 149–169. Hermes, N., & Lensink, R. (2007). The empirics of microfinance: What do we know?. The Economic Journal, 117(1), F1–F10. Jain, S., & Mansuri, G. (2003). A little at a time: The use of regularly scheduled repayments in microfinance programs. Journal of Development Economics, 72(1), 253–279. Khazzoom, J. D. (1976). An indirect least squares estimator for overidentified equations. Econometrica, 44(4), 741–750. Lawrance, E. C. (1991). Poverty and the rate of time preference: Evidence from panel data. Journal of Political Economy, 99(1), 54–77. Llanto, G. M. (2007). Overcoming obstacles to agricultural microfinance: Looking at broader issues. Asian Journal of Agriculture and Development, 4(2), 23–40. Madajewicz, M. (2003a). Does the credit contract matter? The impact of lending programs on poverty in Bangladesh. Working Paper. Columbia University, New York. Madajewicz, M. (2003b). Capital for the poor: The effect of wealth on the optimal credit contract. Working Paper. Columbia University, New York. McIntosh, C. (2008). Estimating treatment effects from spatial policy experiments: An application to Ugandan microfinance. Review of Economics and Statistics, 90(1), 15–28. McKernan, S. M. (2002). The impact of microcredit programs on selfemployment profits: Do noncredit program aspects matter?. The Review of Economics and Statistics, 84(1), 93–115. Meyer, R. L. (2002). The demand for flexible microfinance products. Lessons from Bangladesh. Journal of International Development, 13(1– 18), 351–368. Morduch, J. (1998). Does microfinance really help the poor? Evidence from Flagship Programs in Bangladesh. Harvard University, Cambridge, MA. Murray, I. (2001). Cultivating client loyalty. MicroBanking Bulletin, 6(16), 20–24. Park, C. (2006). Risk pooling between households and risk-coping measures in developing countries: Evidence from rural Bangladesh. Economic Development and Cultural Change, 54(2), 423–457. Park, S. (1974). On indirect least squares estimation of a simultaneous equation system. Canadian Journal of Statistics, 2(1), 75–82. Paxton, J., Graham, D., & Thraen, C. (2000). Modeling group loan repayment behavior: New insights from Burkina Faso. Economic Development and Cultural Change, 48(3), 639–655.

Pitt, M. M., Rosenzweig, M. R., & Gibbons, D. M. (1993). The determinants and consequences of the placement of government programs in Indonesia. World Bank Economic Review, 7(3), 319– 348. Pitt, M. M., & Khandker, S. (1998). The impact of group-based credit programs on poor households in Bangladesh: Does the gender of participants matter? Journal of Political Economy, 106(5), 958– 996. Pitt, M. M. (1999). Reply to Jonatan Murdoch’s “ Does microfinance really help the poor? Evidence from Flagship Programs in Bangladesh”. Brown University, Providence, RI. Rahman, A. (1979). Usury capital and credit relations in Bangladesh agriculture. Bangladesh Development Studies, 12(2), 1–46. Rosenzweig, M. R. (1980). Neoclassical theory and the optimizing peasant: an econometric analysis of market family labor supply in a developing country. Quarterly Journal of Economics, 94(1), 31–55. Rosenzweig, M. R., & Stark, O. (1989). Consumption smoothing, migration, and marriage: Evidence from rural India. The Journal of Political Economy, 97(4), 905–926. Rosenzweig, M. R. (1988). Risk, implicit contracts and the family in rural areas of low-income countries. Economic Journal, 98(393), 1148–1170. Rosenzweig, M. R., & Wolpin, K. I. (1985). Specific experience, household structure and intergenerational transfers: Farm family land and labor arrangements in developing countries. Quarterly Journal of Economics, 100, 961–987. Rousseeuw, P. J. (1997). Introduction to positive-breakdown methods. In G. S. Maddala, & C. R. Rao (Eds.), Handbook of statistics (pp. 101–121). Elsevier: Amsterdam. Schubert, R., Brown, M., Gysler, M., & Brachinger, H. W. (1999). Gender and economic transactions. financial decision-making: Are women really more risk-averse?. The American Economic Review, 89(2), 381–385. Sharma, M., & Zeller, M. (1997). Repayment performance in group-based credit programs in Bangladesh: An empirical analysis. World Development, 25(10), 1731–1742. Stiglitz, J. E. (1990). Peer monitoring and credit markets. The World Bank Economic Review, 4(3), 351–366. Timberg, T. A., & Aiyar, C. V. (1984). Informal credit markets in India. Economic Development and Cultural Change, 33(1), 43–59. Stromberg, A. J. (1993). High breakdown estimation of nonlinear regression parameters. Journal of American Statistical Association, 88(421), 237–244. Varian, H. (1990). Monitoring agents with other agents. Journal of Institutional and Theoretical Economics, 146(2), 153–174. Wydick, B., Karp, H., & Hilliker, S. (2007). Social networks, neighborhood effects, and credit access. In Paper presented at the 1st Centre for International Banking, Insurance and Finance Conference on “Microfinance: What Do We Know?”, Groningen, The Netherlands. Wydick, B. (1999). Can social cohesion be harnessed to repair market failures? Evidence from group lending in Guatemala. Economic Journal, 109(457), 463–475.

APPENDIX A See Table 12.

MICROFINANCE AND INVESTMENT: A COMPARISON WITH BANK AND INFORMAL LENDING

897

Table 12. List of the variables used in the regressions Variable description Investment

Credit market participation Other characteristics of loans

Characteristics of the household

Relationship network

Illness/Injuries

Name

Contents

Investment in working capital (agricultural activities) Investment in working capital (nonagricultural activities) Investment in fixed assets (agricultural activities) Investment in fixed assets (nonagricultural activities) Group lending (MFIs) Informal lending Bank lending Amount Interest rate Collateral required HH head: age HH spouse: age HH head: education HH spouse: education HH head is male Religion Feminine ratio No. of persons in HH Father: same activity

Expenditure in seeds, fertilizers, pesticides during one production cycle (taka) Expenditure in raw materials, fuel, equipment maintenance, etc. during one production cycle (taka) Increase/decrease of the stock of agricultural assets in a one-year period (taka) Increase/decrease of the stock of non-agricultural assets in a one-year period (taka) Binary variables (1= the household borrows from each one of the three sources) Loan principal (taka) Percentage Binary (1 = yes) Age of the household head and spouse (years) at the time of the survey

Head: father lives in HH Spouse: father lives in HH Head: mother lives in HH Spouse: mother lives in HH No. of parents alive N. siblings alive N. other relatives alive Injury last year Medical expenditures last year No. of days not working last year

Measures of wealth

Transfers

Farming related variables

Instruments

Land House value Transport value Transfers: cash Transfers: other Transfers: food Total area cultivated Tenure: fixed rent Tenure: sharecropper Target Spouse: distance from family No. of relatives own land Dowry Marriage gifts to spouse

Years of schooling achieved by the household head and spouse at the time of the survey Gender of the household head. Binary (1 = male) Religion of the household head. Binary (1 = Islam, 0 otherwise) Ratio of females to males in the household Number of household members Binary (1 = household head carries out the same activity of his/her father) Binary variables stating whether the head of the household and spouse’s parents reside in the household (1 = yes)

Number of the head of the household and spouse’s relatives who are alive The household head suffered any injury or disease last year Binary (1 = yes) Amount spent last year for medical expenditures (taka) Number of days that the household head has lost during the last year due to injury Land ownership (acres). Value (taka) of the house and means of transport owned by the household Amount of transfers in cash, food or other commodities received by the household from relatives, friends, and neighbors during the last year (taka) Acres, included those received for sharecropping Binary (1 = yes) Eligibility rule for MFIs. Binary (1 = the household owns less than half an acre of land) Distance (in miles) between the household and the spouse’s birthplace Number of the head of the household and spouse’s relatives owning land Amount of money, goods, or estate given by the woman’s family at marriage Amount of money, goods, or estate received by the woman at marriage

Available online at www.sciencedirect.com