DYNAMICS OF PARTY AGGREGATION
IN THIS CHAPTER and the next, we summarize data from hundreds of national elections and tens of thousands of district elections in our four countries to make the case that changes in party aggregation relate systematically to changes in political authority across different levels of government. We take a broad view of party systems in this chapter, focusing on changes over time in the degree to which party systems are national and comparing party systems across countries. In the next chapter we devote closer attention to our specific cases of Canada, Great Britain, India, and the United States.
For party aggregation to occur, voters and candidates have to be partisan in two distinct ways. First, voters must show a preference for choosing candidates with party labels over independent candidates, and candidates have to prefer running under party labels over running as independents. Second, voters and candidates must show partisanship of a more specific kind by demonstrating a preference or loyalty toward particular party labels. In order for voters to have the latter kind of partisanship, they must exhibit the first.
The first version of partisanship—a preference for party-affiliated candidates over independents—is a necessary condition for party aggregation, and, as demonstrated in chapter 4, with few exceptions candidates and voters in the four countries use party labels either to run for office or to vote for candidates. Supporting modern party systems is a shared belief by voters that candidates with party labels are more likely to win and will be more effective in office than those without such labels. Party labels that have meaning across vast reaches of the country—such as the Congress Party in India, the Liberal Party in Canada, the Conservative and Labour parties in Great Britain, and the Democratic and Republican parties in the United States—exist because voters believe that coordination at election time around a party label with voters in other regions, among other things, will make the processes of forming intragovernmental coalitions easier, will ease the difficulties of managing large and farflung national bureaucracies, and will enable voters in future elections to know which collective group of politicians was responsible for the national policies previously adopted. In federal systems, voters may reason that common party labels not only help politicians from different