Redistribution Preference Analysis

The simple Meltzer-Richard logic says the below-mean median voter should want more redistribution as inequality (the mean-to-median gap) widens, yet cross-national data show this relationship is weak or absent. Several mechanisms explain the gap: the prospect of upward mobility (POUM) leads some poor people to oppose redistribution they expect to fund later; beliefs that success reflects effort rather than luck dampen demand even where inequality is high; and the rich may have disproportionate political influence. Redistribution preference analysis tests these competing channels at the individual level to see which dampens the predicted inequality-demand link.

POUM — the prospect of upward mobility — is Benabou and Ok's idea that forward-looking individuals base redistribution preferences on expected future income, not just current income. Someone poor today but expecting to rise may oppose redistribution because they anticipate becoming a net payer, and under plausible conditions this holds even for the risk-averse. It is tested by including a measure of subjective or objective expected mobility in the individual-level regression and checking whether higher mobility expectations reduce support for redistribution among the currently poor, net of current income.

Redistribution attitudes are typically measured on ordered categorical scales — for instance, five points from strongly disagree to strongly agree. These categories are ranked but the distances between them are not meaningful numeric quantities, so treating them as a continuous variable in ordinary least squares can distort estimates and predict impossible values. Ordered logit (or probit) models the probability of falling in each category through threshold parameters, respecting the ordinal nature of the response; for cross-national data a multilevel ordered model additionally accounts for the nesting of individuals within countries.