In the early 1980s, Puerto Rico experienced a conversion from the Food Stamp Program (FSP) to the Nutrition Assistance Program (NAP) and a severe recession. To compare relative undernutrition in Puerto Rico in 1977 and 1984 stochastic dominance methods and statistical inference procedures are applied to nutrient distributions. To separate the effect of the recession from the effect of converting food coupons to cash benefits, a multiple regression model of nutrient consumption is also estimated. In the dominance comparisons, every improvement in nutrition status between 1977 and 1984 is statistically significant at the lowest quintile, while no case of worsening undernutrition is significant at the lowest quintile. The multiple regression model provides some evidence that the conversion from food coupons to cash might have reduced nutrition status in the lowest quintile. But a larger proportion of food assistance was allocated to low-income households under the NAP than under the FSP. No evidence of significant income effects on undernutrition is found.