The Election Monitor's Curse

Abstract

Election monitoring has become a key instrument of democracy promotion. Election monitors routinely expect to deter fraud and prevent post-election violence, but in reality post-election violence often increases when monitors do expose fraud. We argue that monitors can make all elections less fraudulent and more peaceful on average, but only by causing more violence in fraudulent elections. Due to this curse, strategic election monitors can make a positive impact on elections only if their objectives are aligned in a very specific fashion. Monitors who do not aim to prevent violence can be effective only if they are unbiased, whereas monitors who do aim to prevent violence can be effective only if they are moderately biased against the government. Consequently, election monitors with mis-aligned objectives will fail to prevent violence, whereas monitors with well-aligned objectives will be blamed for causing violence.

Keywords: elections, fraud, violence, monitoring, authoritarianism, time-inconsistency