The labor market "quality" of immigrants is a subject of debate among immigration researchers, and a major public policy concern. However, traditional methods of measuring human capital are particularly difficult to apply to recently arrived immigrants. Many factors that have a negative effect on entry earnings also increase either the incentive or the opportunity for faster human capital investment and earning growth. In addition, many country-of-origin acquired skills that are not immediately valued in the U.S. labor market are useful to the acquisition of U.S. skills. Thus entry earnings are not a good measure of the stock of immigrant human capital. This article presents a model of immigrant human capital investment and, using 1970- 1990 census data, presents strong evidence of a systematic and important inverse relationship between initial immigrant earnings and subsequent earnings growth. This result – which persists even after accounting for differences in the immigration flows from different countries, sampling error, and the effects of emigration – is fundamentally different from both earlier cross-sectional estimates and more recent pooled models that constrain cohort growth rates to be equal. Although our model provides theoretical support for an inverse relationship only when source country human capital is held constant, faster earnings growth for low-entry-earnings immigrants is found empirically even when age and education are not controlled for. The immigrant human capital investment model presented here explores general principles that may apply to other labor market transitions that involve skill transferability – including occupational change and labor market reentry.