Climate Change and Migration: Two Case Studies on Mexico
The main objective of this thesis is to empirically assess the impact of climate factors on international and internal migration in Mexico. Climate models have predicted that Mexico is facing long term drying trends and increased frequency and magnitude of weather related extremes in future decades, two climate factors to which migration is sensitive. Given the historical trends of Mexico-U.S. migration and rural-urban migration within Mexico, I evaluate the extent to which precipitation factors have impacted international and internal migration in the decade of the 1990s. Using migration data extracted from the 1995 and 2000 Mexican censuses and disaster data from the DesInventar database, I assess the impacts of disasters derived from precipitation events on interstate and international migration after controlling for other economic and demographic factors. In the case of international migration, I use a multi-level mixed-effects probit model to measure the extent to which frequency of precipitation based disasters in a state influences the probability that a household in that state has members who recently left abroad. I find that in the period 1990-1994, frequency of disasters is positively associated with the probability that a household has sent migrants abroad in that period. I find further that in the period 1995-1999, disaster frequency is negatively associated with the probability that a household has sent migrants abroad in the general case, but positively associated with the probability that a rural household in a dry state has sent migrants abroad during that period. In the case of internal migration, I construct an expanded gravity model using a Poisson pseudo-maximum likelihood estimator to estimate the impact of the frequency of precipitation based disasters at the origin state on the size of interstate migration flows. I find that higher frequency of disasters at the origin state in the years 1985-1989 and 1990-1994 is associated with higher numbers of migrants for the years 1990-1994 and 1995-1999, respectively. That is, disaster frequency appears to affect migration after a lag period. My findings are consistent with previous research suggesting that climate drivers affect migration in Mexico. Further, as climate stressors are expected to increase over time, we can expect an increased potential for climate change to trigger migration. These findings have implications for policymakers when addressing management of climate risk and future migration flows.