Presentation given at the Migrating out of Poverty Conference, 2017.

1. Migration as coping strategy for
natural shock recovery: Evidence
from Hurricane Mitch in Nicaragua
Sara Giunti
University of Trento
Migrating out of Poverty: From Evidence to Policy conference
London – March 28th, 2017

2. Introduction
 Natural disasters and climate-related events are nowadays considered
one of the main risk factors in developing countries, especially for
natural resource dependent households.
 The effectiveness of usual risk-management institutions is limited by the
fact that weather shocks are spatially covariant.
 New Economics of Labour Migration (NELM) literature considers
migration as a risk-diversification strategy: a household level collective
decision driven by mutual insurance purposes.

3. Motivation
 A large body of literature has tried to determine whether exposure to
environmental risks influences mobility decisions, ending up with
counteracting conclusions and leaving space for further contributions
(Piguet et al., 2011; Gray and Mueller, 2012; Gray and Bilsborrow, 2013).
 Although climatic shocks have been proved to lead to a substantial
increase in remittance flows (Clarke and Wallsten, 2004; Bettin and Zazzaro,
2016), the extent to which these income flows contribute to long term
recovery from natural disasters is still an underexplored issue.

4. Research Questions
 Does exposure to unexpected sudden-onset climatic shocks act as
push factor for out-migration flows? For whom and under which
circumstances?
 Do remittances constitute an effective coping strategy for long-
run recovery?
Role of migration as a strategy to cope with the consequences of natural
shock damages, examining the case of Hurricane Mitch hitting
Nicaragua on October, 1998.

5. Data
 Nicaraguan Living Standard Measurement Studies 1998 and 2001
(panel subsample around 15000 individuals and 2600 households):
 First wave: April - August 1998
 Second wave: April - August 2001
 The 1998 wave is exploited as baseline to collect ex ante household
information.
 The 2001 wave is used to identify individuals migrated abroad during
the time in between and to obtain information on household welfare
performances.

6. Shock exposure
Source: Tropical Rainfall Measuring Mission (TRMM) database

7. Empirical strategy
 RQ 1: Probit regression models.
 Separated estimations conducted for individuals migrated in the 12 months after the
Hurricane.
 Multinomial logit model is estimated distinguishing between two destinations:
regional migration within Central American countries and international migration
towards US and Canada.
Separated model for the subsample of agricultural households.
 RQ 2: Multivariate models are estimated assuming consumption
or asset outcomes as dependent variables.

8. Identification
 IV estimations to deal with selection into migration issues. Two
instruments:
 Rate of residents abroad at the department level in 1971;
 Rate of households receiving remittances at the department level
1993.
 Endogeneity issue: households more exposed to this type of climate-
related events may be more likely to adopt migration as an ex ante
insurance strategy. The unforeseeable magnitude of this exogenous
shock is exploited to deal with this concern.

9. Results (I)

10. Results (II)

11. Results (III)

12. Conclusions
 Rainfall levels do not act as a migration push factor as a whole. Only
members of agricultural households increased their probability to
move abroad with the exposure to high rainfall shocks.
 International migration turns out to be a preferred strategy to deal
with natural shock exposure in case of lack of other regular sources
of liquidity, but when household asset endowments allow to cover
migration costs.
 Remittance flows from abroad are an effective strategy to sustain
long term rebound. Particularly relevant is the positive impact of
receiving remittances on agricultural household asset base.