International migration can be viewed as a natural experiment, placing a worker into a different economic environment while holding his human capital endowment fixed. Migration data therefore provide an opportunity to learn not only about the economic forces underlying migration, but also about the nature of human capital. Existing migration theories have not taken advantage of this opportunity. Instead of developing a framework capable of accounting for a broad range of observations, the literature has studied individual observations, such as migrant selectivity or assimilation rates, largely in isolation. A variety of separate theories has been proposed, based on assumptions that are not always mutually consistent. This paper makes a step towards a unified theory of migration. We develop a set of stylized facts characterizing immigrant earnings histories and show how a model of human capital can provide a coherent explanation for these observations. In contrast to most of the literature, we do not assume that unobserved primitives, such as preferences or technologies, differ across countries. In particular, cross-country wage differentials arise endogenously as a feature of labor market equilibrium. Since explaining immigrant earnings is of primary interest, this is an important innovation.
