Sri Lankan households a decade after the Indian Ocean tsunami

• DOI: http://doi.org/10.1111/rode.12586
• Date: 01 May 2019
• Author: Diana De Alwis,Ilan Noy
• Published date: 01 May 2019

1000

|

wileyonlinelibrary.com/journal/rode Rev Dev Econ. 2019;23:1000–1026.

© 2019 John Wiley & Sons Ltd

DOI: 10.1111/rode.12586

REGULAR ARTICLE

Sri Lankan households a decade after the Indian

Ocean tsunami

DianaDe Alwis

|

IlanNoy

Victoria University of Wellington,

Wellington, New Zealand

Correspondence

Diana De Alwis, School of Economics, and

Finance, Victoria University of Wellington,

Rutherford House, Pipitea Campus,

Wellington 6140, New Zealand.

Email: Diana.dealwis@vuw.ac.nz

Abstract

This study estimates the causal effect of the 2004 Indian Ocean

tsunami on household consumption and income in Sri Lanka 8

years after the event, using a difference- in- differences method-

ology and extensive household survey data. The analysis finds

a strong association between the area- wide tsunami disaster

shock and increases in household income and consumption in

the long term. The increase in consumption is much smaller

than the observed increase in income; while the increase in in-

come is mostly observed in nonagricultural wage income (and

a decline in agricultural income). We also find that households

in high- income regions and lower- damage districts experi-

enced a much better recovery, in terms of income, than those

in poorer regions or those districts that experienced more de-

struction. Deviating from the common observation on short-

term adverse impacts of catastrophic disasters in low- and

middle- income countries, these results are suggestive of a po-

tential for long- lasting and more successful recovery scenarios.

Still, Sri Lanka received a very large amount of external as-

sistance post- tsunami—an amount that may not be replicable

elsewhere. It is likely that this massive inflow of assistance,

further helped by the end of the armed conflict in 2009, has

contributed significantly to this relatively successful recovery.

JEL CLASSIFICATION

Q54

KEYWORDS

disaster, household survey, long-run impact, Sri Lanka, tsunami

|

1001

DE ALWIS AnD nOY

1

|

INTRODUCTION

The 2004 Indian Ocean earthquake elevated the ocean floor by at least 3 meters generating a very

powerful tsunami that killed about 226,000 and displaced more than two million people in around a

dozen countries. The large number of casualties and extensive property damage had obvious short-

term adverse impacts on economic activity. Relatively less is known about long- term post- catastrophe

economic dynamics.1 Here, we are interested in households’ vulnerability to the long- term impacts of

the 2004 disaster; specifically, we aim to identify the impact of the tsunami on Sri Lankan households

in the 8 years that followed.

In Sri Lanka, the event was completely unexpected and, from a statistical perspective, undoubtedly

exogenous. Still, households’ socioeconomic characteristics, their exposure and vulnerability to the

hazard itself, their resilience and access to tools and mechanisms to manage the disaster's aftermath,

their preferences, their decisions when the circumstances around them changed, and their choices

during the post- event reconstruction all eventually determined the disaster's long- term consequences

(Hallegatte et al., 2014; McCarthy & Smith, 2009; World Bank, 2013). This necessitates detailed

household- level data to more precisely identify the tsunami's impacts.

With data from very detailed household survey waves, we evaluate the long- term household- level

consequences of this event. The average effect of the tsunami on household income and consumption

in the affected districts is examined in a difference- in- differences setting using cross- sectional house-

hold data from multiple survey years between 1995 and 2012. Since households were not directly

asked whether they were impacted by the tsunami, we identify those affected using their reported

location. As we do not identify those directly affected separately, we ultimately quantify the aggre-

gation of the indirect economic impacts of the event through several channels. These include the

changing behavior as a result of changes in perceived risk of future occurrence or because of increases

in ex- ante preparedness measures, indirect effects on neighbors that were spared the direct impact but

experienced local spillovers, and general equilibrium effects (for example, operating through labor

markets). Overall, we found that affected households experienced a rise in income, and a much smaller

rise in spending, as a consequence of the tsunami. We are further able to document the sources of these

increases.

Our most important contribution is the focus on longer- term impacts, as almost the whole lit-

erature that examines the impact of disasters on household welfare examines the impact of events

only for a short term after the event (at most 2 to 3 years, and frequently much less). The tsunami,

as it was experienced in Sri Lanka in 2004 is also unusual as a case study for two other reasons.

First, and most obvious, it was a catastrophic event, with very high mortality and damage. Most

research focuses on more local and smaller events, and it is conceivable that the extent of damage is

important for recovery trajectories. Second, and what appears to be the more important distinction,

the event generated an unusually intense reporting in the international media.2 At least partially as

a consequence of that, Sri Lanka received a very large amount of foreign assistance. In this case,

the amount of foreign assistance was most likely larger than the value of damaged assets. Our

estimates of the impacts of this event can thus demonstrate what may be considered the positive

consequences of a well- funded recovery even after a catastrophic event. In most cases, the avail-

able funding for recovery is far less than the extent of damage (see Becerra, Cavallo, & Noy, 2014,

2015). As such, our results show what can be achieved in post- disaster recoveries if sufficient funds

were to be made available for low- income countries to recover from disasters through financing

instruments such as more generous disaster contingency aid grants and loans, sovereign insurance

(such as those available to small island states in the Caribbean and the Pacific), or higher flows of

remittances.