The paper presents the "prospect of upward mobility" (POUM) hypothesis, which suggests that even relatively poor people oppose high rates of redistribution because they anticipate that they or their children may move up the income ladder. This hypothesis explains why most democracies do not engage in large-scale expropriation and highly progressive redistribution. The authors formalize the POUM hypothesis, showing that agents oppose lasting redistributions if tomorrow’s expected income is increasing and concave in today’s income. The more concave the transition function and the longer the policy horizon, the lower the demand for redistribution. The analysis is illustrated with an example calibrated to the U.S., where 3/4 of families are poorer than average, yet a 2/3 majority has expected future incomes above the mean, and therefore desires low tax rates for all future generations. The paper also analyzes empirical mobility matrices from the PSID and finds that the POUM effect is indeed a significant feature of the data. The authors show that the POUM hypothesis is compatible with rational expectations, as long as agents have realistic expectations about their future income prospects. The paper also discusses the role of concavity and skewness in shaping the long-run distribution of income and provides two analytical examples. The first is a Markov process with a steady-state distribution where 77% of the population is below the mean income, yet 67% of voters have expected incomes above the mean. The second is a log-linear, log-normal process with complete closed-form solutions. The paper concludes that the POUM hypothesis is a robust explanation of political behavior, as it accounts for the fact that poor voters may support low tax rates due to the slim prospects of upward mobility.The paper presents the "prospect of upward mobility" (POUM) hypothesis, which suggests that even relatively poor people oppose high rates of redistribution because they anticipate that they or their children may move up the income ladder. This hypothesis explains why most democracies do not engage in large-scale expropriation and highly progressive redistribution. The authors formalize the POUM hypothesis, showing that agents oppose lasting redistributions if tomorrow’s expected income is increasing and concave in today’s income. The more concave the transition function and the longer the policy horizon, the lower the demand for redistribution. The analysis is illustrated with an example calibrated to the U.S., where 3/4 of families are poorer than average, yet a 2/3 majority has expected future incomes above the mean, and therefore desires low tax rates for all future generations. The paper also analyzes empirical mobility matrices from the PSID and finds that the POUM effect is indeed a significant feature of the data. The authors show that the POUM hypothesis is compatible with rational expectations, as long as agents have realistic expectations about their future income prospects. The paper also discusses the role of concavity and skewness in shaping the long-run distribution of income and provides two analytical examples. The first is a Markov process with a steady-state distribution where 77% of the population is below the mean income, yet 67% of voters have expected incomes above the mean. The second is a log-linear, log-normal process with complete closed-form solutions. The paper concludes that the POUM hypothesis is a robust explanation of political behavior, as it accounts for the fact that poor voters may support low tax rates due to the slim prospects of upward mobility.