18
even though their short-term interests may diverge, parties at times are able to act successfully in
a coordinated fashion as a cartel (see Katz and Mair 1995, Blyth 2002) or in fashioning pre-
electoral alliances, joint policy platforms and coalition agreements (Strøm and Müller, 1999b).
Even so, when attempting to apply these models to understand the actual dynamics of systems of
coalition government, crucial attention must be paid to the competing interests of individual
actors that the model assumes aggregate into a single strategic player.
Moves and Game Structure
The models also assume an order to the game: first Out moves, then 1, then 2. This order
was not chosen arbitrarily – it was chosen because the models are ‘heresthetics’ models that
focus on opposition strategies in coalition games. In the unidimensional games, changing the
order of play does not largely change the equilibrium outcomes or the comparative statics that
suggest that parties that value government highly and face low costs in moving positions on
issues are more successful at entering (and staying in) governing coalitions.
However, the multidimensional heresthetics model gives Out a significant advantage.
The option of introducing a new dimension is not available to 1 and 2. However, this is
primarily done for analytical tractability – if we allow each party to raise issues, we begin to deal
with 3-dimensional and 4-dimensional space which becomes too messy to handle in the
relatively straightforward manner in which it is approached. However, the fundamental logic of
the multidimensional heresthetics game should hold even if subsequent responses by 1 and 2 are
to introduce new dimensions. Each party will attempt to introduce issue dimensions that
privilege themself: dimensions on which they are likely to pick up voters or improve their
relative distance to potential coalition partners. By allowing all actors to introduce issue
Convention