We study the seasonality of labor migration from Armenia to Russia based on a unique
panel data set on Armenian workers. As seasonal migration is a repeated decision, the panel
nature of our data set allows us to specify a dynamic switching regression model with
endogenous switching. In line with the Harris and Todaro model and the New Economics of
Labor Migration theory our evidence pinpoints migration experience, the expected income
gap between the home country and the host country, as well as low job opportunities in
Armenia and the possibility to diversify income risks as the most important factors
determining seasonal migration. The hypothetical income gap is about 300 percent, both for
migrants had they stayed and non-migrants had they migrated. This suggests that the
sorting into migration is also influenced by specific migration costs, which are related to the
distribution among tertiary educated migrants and non-migrants.