ECON 429 Lecture Notes - Lecture 17: Instrumental Variable
Document Summary
Migration and networks: from the munshi 2003 paper. Goal: to identify the causal effect of networks on employment (likelihood and quality) Potential problem: networks may be endogenous, e. g. migrants go to localities that are experiencing an economic boom. Economic growth attracts many migrants and makes it easier to find jobs (or good jobs). Solution: find a variable that affects emigration (network) without affecting the host country labor market conditions (instrumental variable). M03 uses rainfall shocks in the home village. Mx village i increases the employment opportunities at home (agriculture is the main source of employment in the mmp villages) but has no effect on wages or employment in us locality j. Scarce rainfall=big network=better job opportunities for migrants in the us (this last link is the one we want to test). Results: bigger network (i. e. scarce rainfall) increases both p(j = 1| m = 1) and p(jg | j =