The initial earnings of U.S. immigrants vary enormously by country of origin. Via three interrelated analyses, we show earnings convergence across source countries with time in the United States. Human-capital theory plausibly explains the inverse relationship between initial earnings and earnings growth rates: the good fit between data and theory suggests that variation in initial skill transferability—not variation in the “quality” of human capital—underlies variation in initial earnings. A new method of testing for emigration bias confirms that selective emigration does not cause the convergence. Functional form and sample selections embedded in most recent analyses of immigrant economic assimilation bias downwards the earnings growth of post-1965 U.S. immigrants. When both functional-form and sample-selection constraints are lifted, a dramatically different picture of the economic assimilation of U.S. immigrants emerges.
immigrant economic assimilation human capital investment country of origin immigrant earnings convergence earnings growth unbiased estimation