Mexico has reported worker's remittances to equal $16.6 billion in 2004, which constitutes nearly 2.5 percent of Mexico's GDP, exceeding the inflows from direct foreign investment and aid. We develop a model of remittances based on a net income concept. The model is used to generate a series of testable hypotheses. We test these hypotheses using what we term a type II generalized ordered probit model based on survey data for Mexican Migrants. Our results are generally consistent with standard utility maximization theory, and more specifically are consistent with a net income hypothesis. we find, for example, that migrant income is a strong positive determinant of remittance levels except for the lowest remittance category. We also find that migrants remit more when they have more family members in Mexico and fewer in the U.S., when they own land and real estate in Mexico, and when they plan on returning to Mexico relatively soon.