Enterprise recovery following natural disasters

From SKAIPedia
Jump to navigation Jump to search


De Mel, Suresh, David McKenzie, and Christopher Woodruff. 2012. "Enterprise recovery following natural disasters." The Economic Journal, 122(559): 64-91. | Link


The authors randomly assign grants to microenterprises in Sri Lanka, some of which were affected by the December 2004 Indian Ocean tsunami. They estimate that the grants helped microenterprises recover faster.

Experimental design

The authors recruited 608 microenterprise owners in Sri Lanka into a panel survey. 205 enterprises suffered damage to business assets as a result of the tsunami. 208 enterprises did not suffer direct damage to business assets but were located sufficiently close to the tsunami-affected areas to experience significant market disruptions. 195 enterprises did not experience either direct damage or market disruptions.

The authors conducted 11 quarterly surveys beginning in April of 2005 and ending in April of 2007. Two additional follow-up surveys were conducted in October 2007 and April 2008.

Grants were framed as compensation for participation in the panel survey and were randomly provided to 117 microenterprises. 78 grants were for 10,000 Sri Lankan rupees (LKR) or about US$100, and 39 grants were for 20,000 LKR. Half of the grants were administered as cash; the other half were in-kind, with the owner choosing the items to be purchased. 83 grants were provided in May of 2005 and 34 were provided in November of 2005.

The authors compare microenterprises directly damaged by the tsunami to indirectly affected microenterprises, which are generally located in the same geographic area, and to unaffected microenterprises, which are generally located further inland. Directly damaged firms are on average larger and more profitable than either of the comparison groups, prompting authors to either use firm fixed effects in their estimates or condition on pre-tsunami characteristics.


Average losses among microenterprises suffering damage were 331,800 LKR, of which household assets comprised 242,100 LKR. Conditional on reporting damage, one third of microenterprises lost all of their assets and half lost more than 85 percent of their assets.

Over 94 percent of microenterprises reported receiving grants for repairs, and 3.5 percent reported receiving loans. The average grant was LKR 33,200 and the average loan was LKR 49,600. The correlation between damage and aid received was small and statistically insignificant.

Intent-to-treat estimates show that an extra LKR 100 in grants increases monthly profits by LKR 5.3 (s.e. of 2.1), monthly sales by LKR 37.7 (s.e. of 11.8), and capital stock by LKR 89.4 (s.e. of 19.2). There is no significant change in owner hours worked. There is suggestive evidence that the impact of grants is rising with the proportion of assets damaged for profit and capital stock and falling with the proportion of assets damaged for profit, but these interaction terms are not statistically different from zero.

There is suggestive evidence of diminishing returns with respect to grant size (return to LKR 20,000 grant is less than twice the return to the LKR 10,000 grant), of a larger return to the in-kind grant compared to the cash grant in the affected areas, and of a smaller return to the in-kind grant compared to the cash grant in the affected areas, but none of the differences are statistically significant.