Are poor people credit-constrained or myopic? Evidence from a South African panel

Abstract

Credit constraints are an almost ubiquitous assumption in development economics. Yet direct evidence for credit constraints is limited, and many observations consistent with credit constraints are equally compatible with myopic (non-forward-looking) consumption or precautionary saving. Using household panel data and a source of widely anticipated income in South Africa, this paper tests and rejects the standard consumption model with perfect capital markets. Then, myopic consumption and precautionary saving are tested as alternative explanations for the observed jumps in expenditure. The standard model with credit constraints cannot be rejected in favour of myopic consumption or precautionary saving.

Experimental Evidence on Credit Constraints

“…The permanent income hypothesis states that if credit constraints do not bind, then consumption should not respond to changes in expected income. Empirically, expected changes in income are associated with changes in consumption in developing countries (Angeletos et al, 2001;Banerjee and Duflo, 2004;Berg, 2013;Browning and Collado, 2001;Edmonds, 2006;Galiani, Gertler, and Bando, 2016;Hsieh, 2003;and Lindskog, 2013). A test for credit constraints based on the sensitivity of consumption to expected income faces two main challenges.…”

“…They all conclude that policy interventions that take into account the decisionmaking process of the poor can lead to welfare improvements. Berg (2013) noted that credit constraints cannot bind when the agent faces an anticipated decrease in income because the technology required is savings and not credit. He ruled out bounded rationality by noting the lack of symmetry in consumption response to changes in expected income.…”

“…However, evidence on the existence of such market imperfection is scarce (Berg, 2013;Karlan and Morduch, 2010). Testing for the existence of credit constraints is difficult.…”

“…It states that consumption should not respond to expected income changes. On the other hand, consumption is sensitive to consumption (Angeletos et al, 2001;Banerjee and Duflo, 2004;Berg, 2013;Browning and Collado, 2001). Some researchers interpret this empirical fact as evidence of credit constraints.…”

“…Berg (2013) observed that consumption for credit-constrained individuals should respond to increases in expected income but not to decreases. He observed changes in consumption in households that became eligible for a pension program and households that lost their eligibility for a child grant in South Africa.…”