Fifteen Eighty Four

Japan is a democracy, yet electoral competition is utterly dominated by a single party.  For sixty-six of the past seventy years, the Liberal Democratic Party (LDP) has controlled Japan’s government.  Since its formation in 1955, the LDP has failed to win a plurality of seats only once.  In every other election to Japan’s most powerful parliamentary house, the LDP has consistently outperformed its rivals, often securing more than twice as many seats as its closest competitor.  My book asks, how has this situation come about, and how is it sustained?

Simple explanations will not do: it is not voter preferences or an association with economic growth that helps the LDP stay victorious.  Instead, the LDP is linked to factionalism, corruption, entrenched male patriarchy, and pandering to special interests.  Its socially conservative positions are increasingly out of step with a majority of Japanese voters.  While the party once oversaw Japan’s economic miracle, it has failed to reverse decades of economic stagnation and mounting demographic pressures.  Recent work underscores this paradox: despite having some of the least popular policy positions among all parties in three recent elections, the LDP secured more than 55% of seats in each – a remarkable feat in a multi-party system.

In this book, I argue that the LDP’s grip on power enables its members to leverage central government resources for electoral gain.  While the claim that the LDP engages in this practice is not new, the mechanics of how it does so – to whom it delivers resources, when, and according to what rationale – remain underexplored.  My book:

• Offers a theory for how politicians in a dominant party can allocate government resources to get the most electoral bang for their buck, meaning the greatest return (in votes) for resources delivered.

• Argues that this “tournament” strategy is viable only under a particular configuration of political institutions.  Specifically, politicians need their electoral districts to be composed of multiple voter groups, at which levels of electoral support are discernible, and government resources targetable.

• Uses thirty-five years’ worth of new fiscal, demographic, and voting behavior – carefully matched to thousands of Japanese municipalities – to empirically test my claims.

In what follows, I unpack each contribution further.

1. Group-based clientelism

Over the past two decades, clientelism – the practice of tying the distribution of benefits to voter behavior – has become a major focus of research.  Typically, clientelism is considered an exchange between politicians and individual voters, and studies focus on developing democracies, where a small payment from a politician has the potential to change voter calculus as to who to vote for.

My book challenges this perspective by showing that politicians sometimes have stronger incentives to engage in exchanges with groups, rather than individuals – trading collective benefits for levels of electoral support.  Prior research has largely dismissed group-based clientelism due to collective action problems within groups.  I argue that politicians can design benefit allocations in ways that overcome these problems, making such exchanges not only possible, but highly effective.

My theory of group-based clientelism challenges the assumption that clientelism cannot be a fixture of competition in wealthier democracies and questions the prevailing view that political institutions have little impact on a politician’s decision to pursue clientelism.  It calls on scholars of distributive politics to devote serious attention to the possibility that collective benefits are being used in ways that tether voters to incumbents and restrict their freedom to vote for whom they choose.

1. Tournaments

When it comes to how collective benefits are harnessed for electoral gain, I turn to recent theory on “tournaments”.  I explain that when politicians in a dominant party compete in electoral districts composed of distinct voter groups – at which levels of electoral support can be discerned and resources targeted – they can turn those groups against one another to compete for resources.

In such a competition, groups are ranked in accordance with the level of electoral support they provided the politician in the most recent election.  By offering a disproportionately large prize to the highest-ranked group, politicians can induce competition, which drives up their vote share.  Among groups that are highly supportive of the politician, rankings can hinge on very small numbers of votes.  Voters in those groups are highly incentivized to set aside their policy preferences, and think only about the degree of influence their individual vote wields over the size of their group’s prize.  Put simply, no one wants to be the reason their group misses out on a large chunk of resources it could have otherwise won.

When politicians are relying on this electoral strategy, we should observe distinct patterns of resource allocations and turnout rates across voter groups.  To examine these hypotheses, I turn to the case of Japan, 1980-2014.

1. New Explanation for LDP Dominance

My book explains that characteristics of Japan’s national and sub-national organization mean that the vast majority of Japanese politicians have competed in electoral districts where tournaments between voter groups (municipalities) are possible.

Using a comprehensive new dataset covering the universe of Japanese municipalities over a thirty-five-year period, I show that resource allocations in this period bore the hallmarks of a tournament.  Within electoral districts, the resources municipalities received was a function of their position in a rank order of municipalities in the same district – constructed on the basis of support for the LDP politician in the most recent election.  Moreover, municipalities received progressively larger sums of money as they climbed in rank, consistent with the idea that politicians calibrate the size of the prize municipalities are fighting over to be larger at higher ranks.

Across electoral districts, resource allocations are influenced by another variable: the relative sizes of municipalities in the district.  The theory expects this to be the case: it is much harder to induce competition between unequally-sized groups.  As a result, politicians in those districts have to offer more (in terms of resources), but are likely to receive less (in terms of support).

In testing this implication, I shed light on a longstanding puzzle: why, if LDP politicians are leveraging government resources for electoral gain (as we are led to believe in many studies), do we observe districts where the party did worse (in terms of votes), getting more (in terms of resources)? I explain that a negative correlation between resources and support across electoral districts is in fact an implication of resources being exchanged for support within electoral districts.

My book also shows that turnout rates are a function of the rank municipalities expect to attain in these tournaments.  Within electoral districts, municipalities concentrating larger shares of votes cast on a single LDP winner have higher turnout, while those concentrating larger shares of votes cast on a single winner from the opposition, a single LDP loser, or all LDP winners (prior to 1994), do not.  I also show that when elections become less competitive, turnout declines, but the decline is less pronounced in municipalities that are highly-supportive of their LDP politician.  In these municipalities, voters turn out to influence their group’s prize, not who wins.

My book explains that Japan’s electoral reform (1994) and municipal mergers (2002-6) led to a redrawing of the boundaries of electoral districts.  I show that this increased the share of districts where tournaments are not possible. Whereas tournaments were once possible in 91% of districts, by 2014 – the last year of my study – this had dropped to 52%.

I show that tournament-possible districts have higher support for LDP politicians and higher turnout.  The LDP candidates in those districts spend more time advertising their ability to secure central government money for constituents in their election campaigns.  There is also evidence that LDP voters in districts where tournaments are being conducted are less happy with the programmatic aspects of the party’s platform.

In sum, the takeaway is that many LDP politicians win elections by making the receipt of valued resources contingent on repeated displays of electoral loyalty.  This raises the costs of voting for the opposition.  In a subset of districts, however, LDP politicians cannot conduct tournaments.  This has nothing to do with the type of voters in those districts.  It is solely a function of how the boundaries of those districts are drawn, relative to municipalities.  Opposition parties do better in these districts, and competition in these districts is more likely to revolve around the programmatic aspects of party platforms.

If the number of districts where tournaments cannot be conducted continues to increase, the LDP will find it harder and harder to win elections.  In the meantime, the party has to manage the conflicting pressures emanating from its two sets of politicians: those for whom continued access to targetable resources is key, and those for whom policies popular with the median voter represent their best chance of clinching victory.