Beaulieu and Hyde 2
isolation.
1
This paper shows that an examination of the relationship between election boycotts
and international election monitoring provides a lens through which we can view the strategic
interaction during democratization between incumbent leaders, opposition parties, and the
international community, thus increasing our understanding of elections in developing countries.
The frequency of internationally monitored elections has mushroomed since the end of
the Cold War. In the 1970’s internationally monitored elections were virtually nonexistent.
Today, it is unlikely that a leader of a democratizing country will choose to have an election
without inviting a delegation of international election monitors to judge the validity of the
process. In 2000, 67.5% of elections in developing democracies were observed by one or more
international delegations.
2
This trend has brought the international spotlight onto many elections
in the democratizing world, thus influencing the payoffs to parties that choose to boycott
elections.
Election boycotts on the part of political parties are not necessarily a post-cold war
phenomenon, but they have happened with increasing frequency in the 1990’s. For example,
whereas opposition parties boycotted 8% of elections worldwide in 1990, by 2000 22.5% of all
elections worldwide were boycotted. In all of these boycotts, the parties involved did not merely
abstain from participation. Rather, they announced explicitly and publicly that they were not
participating in the election in question, and urged voters to follow their lead. Generally
speaking, political parties stage boycotts to improve their chances of attaining power in the
1 Both empirical phenomena have received scant attention on their own. For more information on political party
boycotts, please see the forthcoming dissertation by Emily Beaulieu. For more information on the rise of
international election observation, see the forthcoming dissertation by Susan Hyde.
2 This percentage varies depending on what is counted as an election. Many countries have separate legislative and
executive elections that nevertheless take place in the same calendar year. The decision by incumbents to invite
observers, and the decision by observers to attend, is a separate process for each temporally distinct election.
However, for the purposes of merging observation data with boycott data, the dataset used in this paper is country-
year. This method of counting depresses the percentage of all elections that are monitored by about 12% in the year
2000.
Convention