<br>

A THEORY OF PARTY SYSTEM

3
EMERGENCE: ANALYTICAL STEPS

Given the insights and limitations of the models reviewed so far, a more systematic
explanation of the emergence of party systems must be based on the integration of two
theoretical steps. The first step is strictly analytical and consists in describing the
structure of incentives within which voters vote and politicians decide to run for
office. Depicting the institutional mechanisms of any representative democracy, that
is, the ways in which elections, government formation, and policy making work, will let

8 Most explanations of the choice of institutions either depict institutions as efficient solutions to
collective action and time consistency problems (Knight 1992) or simply refer to them as self-reinforcing
equilibria (Putnam 1993). The adoption of proportional representation has been related to the trade
requirements of small countries (Rogowski 1987) or to its capacity to manage political conflict in
heterogeneous societies (Katzenstein 1985; Lijphart 1977). Political parties have been purported to solve
coordination failures (Aldrich 1995; Cox 1987). Still, efficiency theories have clear difficulties in account
ing for cross-temporal and cross-national variation and, particularly, in explaining the choice of
suboptimal institutions, the pervasive existence of instances of political stagnation, and the breakdown
of political regimes. As a matter of fact, the predominance of these rather functionalist approaches is
doubly surprising given a central insight of analytical institutionalism: that, precisely because they shape
political equilibria, institutions are themselves the outcome of political strategies (Riker 1986).
<br>

508 CARLES BOIX

us predict, at a most general or abstract level, the ways in which voters and politicians act
to maximize the probability ofreaching their objectives-given the institutional rules in
which they operate. This can done by drawing upon the very extensive contemporary
literature on elections, spatial theory, and strategic coordination (driven by electoral
rules) to describe the workings of representative government.
The second step is historical. Starting from the incentive structure that shapes the
behavior of voters and parties, we need to describe and make sense of the sequence of
historical events through which politicians coordinated in parties, choosing particular
programs and electoral strategies, and voters then rallicd around them. Such an approach
should show why certain parties emerged at certain historical junctures; why others did
not or, if they did, failed at the ballot box; and, finally, why they chose particular electoral
institutions which, in turn, shaped the number and type of viable parties.
These two steps are complemerntary. Without analytical foundations, the historical
description of party systems and elections would remain an amorphous, variegated set
of facts. With some partial (and very laudable) exceptions, most of the scholarly
literature on this problem still has this character. Conversely, without any attention
to the historical sequence at work, we would be unable to account for the dynamics of
institutional change that were at the heart of the process of party formation we want to
explain. And therefore we would be unable to sort out which variables preceded which
events and how. As discussed before, this lack of attention to history (and, more precisely,
timing) handicaps most of the formal work on clectoral systems and party systems.
The findings that institutional structures covary with the number of parties are mostly
correlations that tell us little about the way in which institutions and parties unfolded
over time and how actors shaped the mechanisms of democratic representation.
Consider now in more analytical detail the way in which democratic elections
work. As summarized in Figure 21.1, in any election, some citizens first decide to
advance their candidacy (jointly with a particular electoral platform or list of policy
promises). Voters then cast their ballots and only a few candidates become elected,
that is, only a few gather a sufficient number of votes (where sufficient is defined by
the electoral law in place) to get into public office. After elections, the elected
representatives convene to form government and set policy until new elections take
place. These policies, which may well include the procedures to select candidates in
future elections (i.e. the electoral laws), purportedly have an impact on the well
being of citizens (and politicians) and should affect how the next election is fought
and how voters will decide again.

Candidates Voters Government Choice of policy

run for office formation (& electoral rules)
Vote

ig. 21.1 The temporal sequence of representative democracies

<br>

EMERGENCE OF PARTIES AND PARTY SYSTEMS 509

some policy outcomes which

Given this temporal structure, where elections lead to
may have some welfare effects on the electorate, the decision of voters to cast their ballot
voters generally
for a given candidate depends in turn on two things. In the first place,
vote for the candidate who makes promises (or who, in case she is running for
are relatively closer to their ideal
re-election, brings in certain accomplishments) that
Yet, in the
policies or goals than the positions adopted by the remaining politicians.
a on them
second place, the decision of voters to support politician is conditional
considering her a viable candidate, that is, on believing that she has reasonable chances
voters vote
of winning (and probably of affecting the choice of policy). More precisely,
for a candidate if they (individually) expect that she will muster enough support from
other voters to be elected to office. Thus, voters may decide to eschew the candidate they
like most but believe very unlikely to win and instead concentrate their support
on a less
a even less preferred
preferred but electorally "safer" politician (to defeat third and
candidate).
Given the mechanisms that define the behavior of voters, we
can turn to see how
ways they will
politicians will behave-that is, under what conditions and in what
are attractive
decide to run for office. First, candidates will tend to pick policies that
to a sufficient section of the electorate (to go on and win the seat they are contesting).
Second, politicians will only contest an election (and therefore parties will only be
launched) if they can appear as viable candidates, that is, only if voters will take them
as serious alternatives worth voting for.
From the structure of incentives of voters (electing welfare-maximizing represen
tatives) and politicians (being recognized as useful and viable candidates and then
being elected), there follow three main components or building blocks
on which we
should construct a theory of party system formation: the preferences of voters; their
information and beliefs (about what other voters will do and the chances of politi
are
cians to get elected); and, finally, the electoral institutions through which votes
aggregated and candidates elected.
a
First, since elections discipline politicians to act on behalf of the electorate,
an account of the distribution of
theory of party system formation must depart from
the voters' preferences, that is, of the electoral space in which parties compete. Parties
cannot deviate too much from the demands of their voters because they risk being
heavily punished in the electoral arena. Still, knowing the profile of the electorate
would seldom be enough to predict the nature of the party system in any given
country for reasons I discussed before: only on rare occasions do voters have full
information about the policies of parties; and they may have positions in multiple
dimensions that may make them susceptible to many different types of political
appeals. All in all, however, the types and numerical distribution of voters act as the
main determinants of how parties position themselves in the electoral marke.
Second, since the expectations voters have about the electoral chances of different
candidates may be as central as (or sometimes even more central than) the interests
of voters in determining their vote, we need to pay attention to the mechanisms that
determine those beliefs and therefore that shape both the strategic coordination of
politicians and the instrumental vote of voters. Two points are in order here.
<br>

510 CARLES BOIX

The existence of expectations (and the corresponding strategic behavior that

follows from them) confers some considerable advantage on those politicians that
entered at the beginning of the sequence of elections relative to later entrants. That is,
in an empty electoral arena, where there still is little information (provided by past
elections) about who may be the best candidate around which to coordinate to defeat
the least preferred politician, strategic considerations may play a minor role in voters'
minds. By contrast, once some parties beconme established as the main electoral
contenders, voters are suddenly much more constrained in their behavior. The
main parties constantly appeal to their own electoral viability, as already proven in
previous elections, in opposition to new, perhaps more preferred but as yet untested
candidates, to maintain the allegiance of voters.
This electoral advantage, which comes from having some recognizable label, some
organization, and some reputation at winning elections, has an additional and very
important consequence. It gives parties the capacity and time to adjust policy promises
and particular candidates to shifts in the electorate or in the preferences of electorates so
that they can remain strongly competitive in the electoral arena. In short, the combin
ation of being an early entrant and exploiting the strategic calculations of voters explains
why we see so much persistence in party labels over time, even under periods of
considerable ideological and social flux. Thus, for example, in spite of several wars, the
transfornmation of the economy away from traditional smokestack industries, and
substantial changes in social values, Democrats and Republicans have split the American
electorate for over a century. Similarly, Conservatives and Labour have alternated in
power in London for about eighty years. It has been by exploiting the strategic coord
ination of voters and gradually adjusting policies over time that these parties have
systematically deterred the entry and consolidation of many third political forces.
This discussion on the stability of party systems leads us in turn to a second point,
namely the conditions that precipitate the breakdown of the prevailing party system. At
first sight, new parties seem to emerge whenever the electoral market changes abruptly
either because substantial numbers of new voters participate in the polls (due to a change
in franchise rules) or when the interests of significant portions of the electorate change
(for example, after massive migrations to the city or sudden political realignments
precipitated by war). Yet, by and large, the formation and consolidation of new parties
takes place only when their candidates are able to break the expectations" advantage
that the existing parties tap into to sustain their leading position. This can only happen if
the entering parties enjoy sufficient organizational strength to mobilize their electors and
move them away from the old equilibrium (in which they voted or are going to vote for
one of the old parties because it was viable) to a new equilibrium (in which they support
the new party). As the historical section discusses in more detail, the irruption of socialist
parties in Europe followed the decision of trade unions to shift their support away from
(left-leaning) liberal politicians to social democratic candidates.
Finally, since the decision of voters over whom to vote for (and, in fact, of
politicians to form a party) is strongly affected by the institutional rules that specify
the translation of votes into seats, we need to examine the nature and selection of the
electoral laws to understand the formation of party systems. In the international
<br>

EMERGENCE OF PARTIES AND PARTY SYSTE MS 511

system, actors (the states) do not choose the rules of the game -in fact, there are
hardly any, except for some informal cooperation within a general condition of
anarchy. In markets, firms only partially (if at all) define the rules of competition.
By contrast, in representative democracies, politicians set the very rules according to
which they will be selected. And they generally draft them to maximize their electoral
chances. This simple fact forces us to tackle the causes underlying the choice of the
rules of the game. In other words, a theory of party formation would be incomplete if
we did not pay attention to the incentives that make politicians maintain the legal
status quo (something that happens most of the time) or alter it (something that
oCcurs very rarely and with substantial consequences).