circumstances; and explained government termination. The research on government
coalitions has produced several theories and models that anticipate what sort of
government will form and how long it will last under varying circumstances. Early
coalition theories examined what motivated politicians when considering the range of
possible government coalitions (von Neumann and Morgenstern 1953; Gamson 1961;
Riker 1962; Dodd 1972; DeSwaan 1973). Later research incorporated the institutional
features that alter parliamentary outcomes (Laver and Schofield 1990; Huber 1996;
Mershon 1996, 2002; Grofman and Van Roozendal 1997; Martin and Stevenson 2001).
One implicit assumption underlying these theories is that the parliament -- understood to
be the popularly elected lower assembly -- chooses the prime minister and cabinet. When
political scientists build models of coalition politics, they generally impose a unicameral
assumption when estimating how long a coalition will last and what type of government
will form, despite nearly one-third of all countries having a bicameral legislature
(Tsebelis and Money 1997: 1).
According to Druckman and Thies (2002), the upper assembly in bicameral
systems has the potential to influence coalition politics. Although government majorities
rarely depend on upper chamber support for coalition survival, the upper house does
influence policy outcomes. Therefore, potential coalition partners in bicameral systems
may consider capturing upper and lower house majorities to facilitate passage of
legislation. As a result, governments with concurrent support should survive longer than
those with sole support in the lower assembly. Druckman and Thies (2002) test these
hypotheses in ten Western European countries and find little evidence of bicameralism
when coalition builders form a government. However, governments with concurrent
majorities survive significantly longer those without upper chamber support.
3
Convention