Social Security and Democracy

with Ricard Gil and Xavier Sala-i-Martin


Many political economic models use and emphasize the process of voting in their explanation of the growth of Social Security, government spending, and other public policies. Their use and emphasis implicitly assumes that, without voting, Social Security spending would grow less, and/or that other public policies would be different, if voting were not used to make public decisions. But is there an empirical connection between democracy and Social Security program size or design? Using some new international data sets to produce both country-panel econometric estimates as well as case studies of South American and southern European countries, we find that Social Security policy varies according to economic and demographic factors, but that very different political histories can result in the same Social Security policy. We find little partial effect of democracy on the size of Social Security budgets, on how those budgets are allocated, or how economic and demographic factors affect Social Security. If there is any observed difference, democracies spend less of their GDP on Social Security than do economically and demographically similar nondemocracies, and are more likely to cap their payroll tax. Democracies and nondemocracies are equally likely to have benefit formulas inducing retirement and, conditional on GDP per capita, equally likely to induce retirement with a retirement test vs. an earnings test.

© copyright 2001 by Casey B. Mulligan, Ricard Gil, and Xavier Sala-i-Martin.