Local droughts and income risk among Thai households

• Published date: 01 November 2021
• Author: C. Rashaad Shabab
• Date: 01 November 2021
• DOI: http://doi.org/10.1111/rode.12812

2084

|

Rev Dev Econ. 2021;25:2084–2112.

wileyonlinelibrary.com/journal/rode

1

|

INTRODUCTION

In low- and middle- income countries access to credit markets may be incomplete, especially for poor

or rural households. These households may therefore have an incentive to insure their consumption

levels by obtaining their income from less- volatile sources or by diversifying their income sources. In

Received: 4 October 2019

|

Revised: 8 April 2021

|

Accepted: 22 June 2021

DOI: 10.1111/rode.12812

REGULAR ARTICLE

Local droughts and income risk among Thai

households

C. RashaadShabab

This is an open access article under the terms of the Creat ive Commo ns Attri bution License, which permits use, distribution and reproduction

in any medium, provided the original work is properly cited.

© 2021 The Authors. Review of Development Economics published by John Wiley & Sons Ltd.

Department of Economics, University of

Sussex, Brighton, United Kingdom

Correspondence

C. Rashaad Shabab, Department of

Economics, University of Sussex, Jubilee

Building, Brighton BN1 9SL, United

Kingdom.

Email: c.shabab@sussex.ac.uk

Funding information

Emily Louise Wells Fellowship of Vassar

College

Abstract

This paper investigates the extent to which households in

rural Thailand across the income distribution are able to

mitigate income risks in the face of shocks. It uses especially

high- quality household income and consumption data span-

ning 64 Thai villages over 15 years. The paper identifies

income shocks by village- level variations during drought

conditions. It finds that richer households are better able to

mitigate income risk than poorer households, in contrast to

some studies of the South Asian subcontinent. These possi-

bilities for managing income risk are shown to be correlated

with the type of contract the head of household is likely to

be employed in, the share of salaries in total household in-

come, the education level of the head, the relative youth of

the heads of richer households, and location effects.

KEYWORDS

inequality, poverty, risk, Southeast Asia, Thailand

JEL CLASSIFICATION

O12; O15; D15; D31; I32

|

2085

SHABAB

an influential pair of papers, Morduch (1994, 1995) used the term “income smoothing” to describe

this phenomenon and contrasted it with “consumption smoothing” that is commonly observed in

richer countries. A number of papers have documented low- risk income strategies among the rela-

tively poor in rural communities of low- income countries. Morduch (1995) presents evidence that

asset- poor households in rural India, whose consumption is most vulnerable to income shocks, de-

vote a greater proportion of their land to safer but lower- yielding traditional varieties of crops than

richer households. Rosenzweig and Binswanger (1993) demonstrated that rural Indian households

in lower- wealth quartiles used production techniques that were less susceptible to rainfall variation,

even though on average these techniques were less productive. Kochar (1999) found that when faced

with crop failure, such households protected their consumption levels by diverting labor from farm

employment to off- farm employment, thereby reducing income variability by diversification rather

than smoothing consumption directly through borrowing or dissaving. More recently, using data from

Mexico, Gutierrez (2014) extends the analysis of income smoothing to implicit contracts that reallo-

cate risk from formal salaried workers to their employers in middle- income countr ies.

However, another strand of the literature has observed that poorer households may be constrained

in their ability to enter low- risk income- generating activities. Dercon and Krishnan (1996) find t hat

in rural Ethiopia and Tanzania, poorer households lack the lumpy assets required to enter high- return,

low- risk activities (e.g., cattle rearing or shopkeeping). They also find that low levels of education

restrict the ability of relatively poor households to gain low- risk salar ied employment. Dercon (2002)

surveys the various constraints to effective risk management faced by poor households.

If this last evidence is general, then although richer households are less likely to be liquidity con-

strained and therefore less in need of insuring themselves through smooth income, they may have priv-

ileged access to low- risk income streams and therefore be more able to manage income risk. It is then

an empirical question as to whether richer households are more likely to utilize low- risk incomes to

satisfy their insurance needs than their poorer counterparts. The present paper examines this empirical

question of the distributional impact of low- risk income oppor tunities using a long running panel of

high- quality household surveys from rural Thailand (see Townsend,2011) spanning 64 villages over

15years from 1997 to 2011.

To distinguish between relatively rich and relatively poor households, I compute observed “perma-

nent income” for each household. From Friedman (1957), I take the average over time of real, equiv-

alized consumption for each household in the 15years for which I have data as a proxy for permanent

income. I show that consumption volatility as measured by the standard deviation of real, equivalized

consumption is largely constant over the distribution of permanent income, but a similarly constructed

measure of income volatility declines systematically with permanent income. These findings are not

consistent with models of income risk that predict that low- income households will rely more heavily

on low- volatility income (as in Morduch,1994). Rather, they lend support to Dercon and Krishnan’s

(1996) hypothesis that poorer households may be excluded from low- risk income opportunities.

The paper adopts a formalized model of the income- generating process of households where in-

come comprises a long- term component and a transitory, stochastic component (as in any number of

studies of consumption smoothing, including Friedman,1957; Hall,1978). In formal empirical mod-

eling, I follow Paxson (1992) and further decompose transient income into a village- wide component

and a household- specific component. I use information gathered by the Townsend Thai Project from

key informant interviews with village headmen to compute the proportion of households in each vil-

lage affected by a drought in 14 of the survey years and use this as a source of exogenous variation in

transient income of all households surveyed within that village cluster.

The empirical section of this paper tests for differences in the extent to which the income streams

of individual households are insured against the cluster- level prevalence of drought by the level of