Women's access to credit: does it matter for household efficiency?

Improving poor households' access to capital is a common element of rural strategies that are designed to induce growth. Economic theory predicts that by relaxing the liquidity constraints of poor households, interventions that enhance these households' access to capital will lead to a more efficient allocation of resources, increased production, and higher welfare (Singh, Squire, and Strauss 1986). A number of researchers have sought to inform this claim with empirical data by assessing the negative impact of credit constraints on farms' and households' efficiency (see for instance, Feder et al. 1990; Ray and Bhadra 1993; Carter and Olinto 2003; Foltz 2004; Petrick 2004; Chavas, Petrie, and Roth 2005; de Mel, McKenzie, and Woodruff 2007; Guirkinger, Fletschner, and Boucher 2007). (1)

By and large, these studies have used the household as the unit of analysis, an approach that can be problematic in settings where there are gender-based market imperfections and significant gender-based asymmetries in how rights, resources, and responsibilities are distributed within the household. Specifically, as has been amply documented in the development literature, rural women are more restricted in their access to and control over land (Agarwal 1994; Deere and Leon 1997, 2001) and capital (Sisto 1996; Ospina 1998; Fletschner 2006), and in the type of entrepreneurial activities in which they can engage given their household roles and social norms (Cain, Khanam, and Nahar 1979; Carney and Watts 1991; Schroeder 1996; Kevane 2004; Fletschner and Carter 2008).

Arguably, using the household as the unit of analysis may still be reasonable if, as most economists implicitly assume, households make Pareto-efficient productive decisions. However, the assumption that family members pull their resources and allocate them to their most efficient use may be incorrect. Intrahousehold dynamics tend to be complex: spouses may hold conflicting preferences, and they may not fully share their labor, assets, or information (Haddad, Hoddinott, and Alderman 1997). When spouses disagree, they may choose to not fully pool their resources even if such decisions lead to a less efficient outcome. In particular, Jones (1983) has found that families in Cameroon could have increased their overall production if the women of the household allocated more of their time to rice cultivation, the income from which typically accrues to the men; and Udry (1994) has found that households in Burkina Faso could have increased their level of production by reallocating labor and fertilizer from men's plots to women's plots.

If household members do not pool or trade resources efficiently, it would be incorrect to assume that husbands with adequate access to capital will help alleviate the gender-specific constraints their wives face in the financial market. In this case interventions that improve women's access to credit are necessary because women tend to have poorer access to resources than men and rationing mechanisms that limit women's access to capital may have substantial economic consequences for their households. In this article I explore the second part of this argument, namely that there may be efficiency gains to be had from specifically enhancing women's productive capacity within the household.

To evaluate this argument, I use data from a survey carried out in 1999, in which husbands and wives (2) from 210 rural households in Paraguay were asked to provide detailed production information and answer a series of qualitative questions designed to assess their individual access to credit. Using this information, I measured households' efficiency and assessed the impact of men's and women's credit constraints on their households' efficiency.

From a methodological perspective this article is in two important ways an improvement over most other studies that have explored how imperfections in the capital market affect efficiency. First, I use qualitative measures that capture whether or not households are credit constrained (do they have adequate access to credit?), rather than whether or not they use credit (have they taken loans?). (3) Second, and a central argument of this article, I use information on each spouse's individual access to credit instead of relying on households' access to credit.

The article presents presents empirical support for efficiency-based arguments to enhance women's access to credit. Analysis of the data confirms that households in which women reported not being able to meet their needs for credit are not producing as much as they possibly could. The costs of these constraints to society are substantial: for the average family, the woman's constraints are associated with an 11% loss in efficiency.

The article is organized as follows. I start by presenting a stylized description of farm households' economic decision making. I then explain the methodology employed to measure household efficiency, discuss the variables used to characterize efficiency, and summarize the efficiency measures. I propose a framework to assess the sources of their inefficiency. I identify spouses' credit rationing status and explain how that information can be incorporated in the econometric analysis. I discuss the factors contributing to households' inefficiency, paying special attention to the impact of women's credit constraints. Finally, I present the conclusions.

A Neo-Classical Farm Household Model

Consider a household i in which, following most standard economic models of rural household decision making, the preferences and constraints of spouses m and f are characterized as follows. Spouses have working capital [K.sup.m.sub.i] and [K.sup.f.sub.i], and they each have L units of labor time. The man dedicates all his labor to the production of marketable goods that can either be exchanged or consumed. The woman divides her time between the provision of household services [Z.sub.i] and the production of marketable goods. (4) Since the provision of household services does not require capital, [Z.sub.i] can be represented by the number of hours the woman dedicates to the provision of household services.

Spouses allocate their labor and capital (L, L - [Z.sub.i], [K.sup.m.sub.i], [K.sup.f.sub.i]) to produce farm and non-farm outputs ([F.sup.m.sub.i], [F.sup.f.sub.i], [N.sup.m.sub.i], [N.sup.f.sub.i]) according to a technology X that encompasses all the feasible input-output combinations. (5) It is important to note that following Chavas, Petrie, and Roth (2005), I use the household (i.e., farm and nonfarm production) rather than the farm as the relevant unit of analysis. Chavas, Petrie, and Roth demonstrate that evaluating efficiency at the farm level is incorrect when input markets are imperfect or when farm and nonfarm activities are based on a common technology. Both of these conditions apply to the rural setting covered in this study: fieldwork observations and survey results described in more detail in later sections of this article confirm considerable nonprice rationing in the credit market and suggest that farm and nonfarm activities are jointly produced, as nonfarm production can be a source of liquidity for farming activities.

In this one-period model households' revenue is calculated by valuing all production at market prices [p.sup.q]([F.sup.m] + [F.sup.f]) + [p.sup.n]([N.sup.m] + [N.sup.f]), loans are repaid (1 + r)([K.sup.m.sub.i] + [K.sup.f.sub.i]), and the profit, [[pi].sub.i], is used for consumption of goods [G.sub.i]. (6) Families' well-being depends on the goods available for consumption and the household services provided by the women and can be represented by a continuously differentiable and quasi-concave utility function [U.sub.i]([G.sub.i], [Z.sub.i]). Thus, family i's economic decision-making process can be modeled as

(1) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

Families differ in the extent to which they are willing to substitute household services for consumption goods, but for any given level of household services [[bar.Z].sub.i], nonsatiation of the utility function implies that household i will maximize its consumption, which is in turn equivalent to maximizing its profit conditional on [[bar.Z].sub.i]:

(2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

Equation (2) implies that for a given level of inputs (L, L - [[bar.Z].sub.i], [[bar.K].sup.m.sub.i], [[bar.K].sup.f.sub.i]), a household that is maximizing its profit must be allocating these inputs in a way that maximizes its revenues. Such a household is economically efficient: it is producing as much as it is feasible given its resources, market prices, the technology available, and the level of household services it is providing. By definition an economically efficient household is technically and allocatively efficient. It is technically efficient because for a given level of inputs, the household will produce as much output as is feasible with the available technology. It is allocatively efficient because given market prices, it will choose to produce the combination of outputs that maximizes its revenues.