This paper examines how individual support for welfare spending in the United States is influenced by both financial self-interest and interpersonal preferences. Using data from the General Social Survey (GSS), the study finds that individuals' preferences for income redistribution are shaped by two key factors: a negative exposure effect and racial group loyalty. The negative exposure effect refers to a decrease in support for welfare as the welfare recipiency rate in one's community rises. Racial group loyalty, on the other hand, indicates that individuals are more likely to support welfare spending when a higher proportion of local recipients belong to their own racial group. These findings suggest that racial diversity in a community can lead to lower levels of welfare benefits, as individuals are less supportive of welfare when they perceive a higher number of recipients in their area. The study uses survey data to measure self-reported attitudes toward welfare spending and compares these with voting behavior on a welfare cuts ballot proposition in California. It finds that individuals who support more welfare spending are more likely to vote against welfare cuts, indicating that self-reported preferences can be used to infer underlying attitudes. The paper also includes a detailed empirical strategy, using regression models to estimate the effects of welfare recipiency rates at different geographic levels (tract, MSA, and state) on individual support for welfare spending. The results show that both the negative exposure effect and racial group loyalty are significant at multiple geographic levels, and that these effects are robust to various specification checks. The study also investigates the validity of the self-reported measure of welfare support by comparing it with voting behavior. It finds that self-reported preferences are a useful measure of underlying preferences, as they closely correspond to voting outcomes. The paper concludes that interpersonal preferences, particularly the negative exposure effect and racial group loyalty, help explain why racially heterogeneous states have relatively lower levels of welfare benefits. These findings contribute to the understanding of the forces driving redistribution and highlight the importance of interpersonal preferences in shaping public opinion on welfare spending.This paper examines how individual support for welfare spending in the United States is influenced by both financial self-interest and interpersonal preferences. Using data from the General Social Survey (GSS), the study finds that individuals' preferences for income redistribution are shaped by two key factors: a negative exposure effect and racial group loyalty. The negative exposure effect refers to a decrease in support for welfare as the welfare recipiency rate in one's community rises. Racial group loyalty, on the other hand, indicates that individuals are more likely to support welfare spending when a higher proportion of local recipients belong to their own racial group. These findings suggest that racial diversity in a community can lead to lower levels of welfare benefits, as individuals are less supportive of welfare when they perceive a higher number of recipients in their area. The study uses survey data to measure self-reported attitudes toward welfare spending and compares these with voting behavior on a welfare cuts ballot proposition in California. It finds that individuals who support more welfare spending are more likely to vote against welfare cuts, indicating that self-reported preferences can be used to infer underlying attitudes. The paper also includes a detailed empirical strategy, using regression models to estimate the effects of welfare recipiency rates at different geographic levels (tract, MSA, and state) on individual support for welfare spending. The results show that both the negative exposure effect and racial group loyalty are significant at multiple geographic levels, and that these effects are robust to various specification checks. The study also investigates the validity of the self-reported measure of welfare support by comparing it with voting behavior. It finds that self-reported preferences are a useful measure of underlying preferences, as they closely correspond to voting outcomes. The paper concludes that interpersonal preferences, particularly the negative exposure effect and racial group loyalty, help explain why racially heterogeneous states have relatively lower levels of welfare benefits. These findings contribute to the understanding of the forces driving redistribution and highlight the importance of interpersonal preferences in shaping public opinion on welfare spending.