Coping with weather disasters through international migration: The role of search frictions and international migrant demand

Millions of migrants leave developing countries as temporary labor migrants annually. Access to so-called “guest worker" markets can allow countries to cope with origin shocks by enabling a migration response. The key feature of temporary labor migration is that migrants need to secure an overseas contract before they leave. This process can be hindered by congestion due to excess supply of willing migrants and search frictions caused by decentralized search between potential migrants and (many) overseas recruitment agencies.

Consider someone whose livelihood was destroyed by a typhoon. As a result, they start searching for overseas employment. However, overseas contracts were already hard to find with willing migrants exceeding available jobs. Now, more are searching for overseas work due to the typhoon, further congesting the market. Will they be able to secure an overseas job? Suppose, after some time, they were able to secure work as a construction worker in Saudi Arabia. They would not accept the current offer under normal conditions. Yet, giving it up to continue searching for a better offer is uncertain and costly, especially given the urgent needs at home. Should they keep searching?

Such considerations raise multiple possibilities that impact the effectiveness of temporary international labor migration as a shock-coping tool. Binding excess supply of migrants may hinder any migration response. Migration outcomes beyond flows can be influenced by the changing option value of continued search. The migration response can further depend on contemporaneous macroeconomic conditions that determine the availability of overseas contracts through migrant labor demand.  In my job-market paper, I tackle these possibilities using migration responses to typhoons from Philippine municipalities as a case study. Using administrative migration data, I study (1) the international migration response to typhoons, (2) how typhoons impact migration outcomes such as migrant wages, occupations, and destination choices and, (3) how macroeconomic conditions that influence international migrant demand mediate the migration response.

Administrative data on new migrant contracts in the Philippines

Temporary labor migration is the most common route to migration out of the Philippines. I use unique administrative data on the universe of new migrant contracts, with ~4 million contracts in my analysis period of 2007-2016.  Critically, the administrative data includes information on contract wage, occupation, destination country, and municipality of origin of the migrant. This information allows me to move beyond migrant flows and study migrant wage responses along with substitution patterns between occupations and destination countries with geographical granularity.

Typhoons increase out-migration…

I measure the yearly typhoon exposure of a municipality by building a continuous exposure measure from meteorological data that factors in the number, intensity, and impact-area population of typhoons. I leverage the randomness in the exact timing, intensity, and paths of typhoons using a fixed-effects model. Specifically, the identifying assumption is that the deviations from a municipality’s average typhoon exposure in a given year is uncorrelated with other determinants of migration, conditional on municipality and year fixed-effects. This as-good-as-random typhoon exposure ensures that early and late treated municipalities are not systematically different and municipalities are not selecting into a treatment “dose” based on factors correlated with the migration response. The latter mitigates concerns about bias in continuous DiD designs due to treatment effect heterogeneity and selection (Callaway et al. 2021). To check whether the staggered nature of typhoon exposure leads to bias in the fixed-effects estimates,  I further employ a stacked-by-event design (Cengiz et al. 2019) on a binarized version of typhoon exposure (to capture extreme typhoons) and show the fixed-effects and stacked estimates using the binary measure track each other closely. 

With this empirical strategy, I first document that a one standard-deviation typhoon exposure increases out-migration from municipalities by 3.7% in the short-run (1-2 years after) and 4.1% in the medium-run (2-3 years after).

… but predominantly for low paying occupations and destinations

Philippine migrants leave for a wide variety of occupations and destination countries, leading to substantial dispersion in migrant wages. Typhoons impact the occupation and destination choices of migrants. The share of migrants going to lowest paying occupations (such as construction workers or domestic helpers) and destinations (such as the gulf countries) jump due to typhoons (Figure 1), even though typhoons increase the educational attainment of migrant cohorts. This is consistent with typhoons leading to lower reservation wages and potentially incentivizing migrants to search for contracts in lower paying overseas markets with higher likelihood of securing a job. Due to this downgrading, average (median) migrant cohort wages fall by 0.27% (0.43%) for each percent increase in migration in the short-run. The drop in wages substantially lower the total migrant earning response to typhoons by $28.5 million, corresponding to 22% of the average migrant earning response.

Figure 1: Migration responses to typhoons by country and occupation

The role of international migrant demand

To understand how migration responses are influenced by international migrant demand, I leverage spatial variation across municipalities in which countries they tend to send migrants to. Combining these migration shares with GDP shocks to destinations, I construct a time-varying municipality level migrant demand index. Better demand conditions lead to a substantially larger migration response, and a proportionally lower wage drop (Figure 2). Combined, total migrant earning response to typhoons are 2.3 times larger during high migrant demand (80th percentile) compared to median migrant demand. I further find that the lower wage drop is primarily driven by lower occupational downgrading when migrant demand is stronger. Intuitively, when typhoons coincide with better migrant demand, migrants are more likely to find overseas contracts, and typhoon induced migrants (who are on average more educated) are better able to secure higher paying occupations, decreasing the pressure to occupationally downgrade.

Figure 2: The variation in migration response with demand conditions

What about those left behind: the remittance response

Does the migration response lead to shock-coping benefits for those left behind? Temporary migrants commonly migrate to support household members through remittances or savings brought home. Using household survey data, I find that typhoons increase remittance flows to affected regions. The estimates suggest that the total remittance response in the analysis decade is approximately $11 billion, a substantial amount corresponding to 180% of total estimated typhoon damages by the government. How much of this remittance response is due to new labor migration in response to typhoons? The household data does not allow for estimating this directly, so I undertake a back of the envelope calculation combining migration, migrant wage, and remittance response estimates. A considerable portion, 15%-18%, of the remittance response is attributable to new migration. Further, consistent with the migration analysis, when migrant demand conditions are better and migration response is stronger, the remittance response is stronger as well: during high demand conditions (80th percentile), the remittance response is 41% larger compared to median demand conditions.

Policy implications

Overall, international labor migration can facilitate coping with shocks, particularly during periods of high international migrant demand.  Given the macroeconomic conditions in destinations are not a policy lever for origin countries, what are the policy implications of these findings, especially given expected increases in extreme weather shocks due to climate change?

• Policies that increase the availability of overseas jobs to affected communities in the wake of disasters can substantially aid in dampening impacts. For example, the government can intensify efforts to secure overseas contracts, provide incentives for private recruitment agencies to increase availability of contracts, or decrease search costs for already available overseas jobs through, for example, job fairs.
• Conversely, restrictive destination country policies can create significant negative externalities in terms of shock-coping for origin countries with strong migration ties. The consequences of lower overseas contract availability does not depend on whether it is due to macroeconomic conditions or due to migration policy.
• There may be gains from decreasing the dependence on any individual country through diversifying the “portfolio” of potential destinations.

One limitation is that we can’t draw firm conclusions about the impacts of a more permanent increase in the availability of overseas contracts from the empirical results, as the analysis is focused on migration response heterogeneity by short-run changes in migrant demand.