Why do governments manipulate elections at varying levels? This paper argues that variation in electoral manipulation can in part be explained by the fundamental trade-off between an incumbent's desire to signal his popularity and his risk of losing elections. Incumbents with a secure office do not need to convince the opposition of their wide support and avoid electoral risks by manipulating elections. Insecure incumbents, however, want to generate a strong signal of their popularity and adopt a riskier strategy of electoral uncertainty. Measuring office insecurity through the incidence of economic crises, coup threats, and protests, I document robust evidence supporting this proposition in a panel of autocracies and transitional democracies using a variety of measures and statistical designs. Politically insecure incumbents are more likely to hold elections, to tolerate electoral competition, and ultimately to lose in the very competition they permit.
