We review the role immigration amnesties have played in US immigration policy, placing them in the context of similar programs embarked upon by other nations. The theory of amnesties suggests rent-seeking, bargaining, and costs as reasons for a country offering an amnesty, often in conjunction with increased border controls, internal enforcement and employer penalties. We model an immigration amnesty in which the destination country has a formal sector employing only legal immigrants, an informal sector employing both legal and illegal immigrants, and open unemployment. The model focuses on the productivity enhancing effects of legalization, and establishes specific conditions under which unemployment, the informal sector and the formal sectors increase/decrease in size. Building on these insights, our empirical work examines Mexican migration to the US. We study who are migrants; among migrants, who are legalized via IRCA, and who are legalized via sponsorship of family or employer. Furthermore, to measure the impact of amnesty on welfare of migrants, we estimate earnings equations of various migrants groups.