On Friday, Tsai Ing-wen suddenly announced that, if she is elected, she will move toward a consociational model of democracy 協商式民主. This took me by surprise, since a consociational model would not seem to fit Taiwan’s conditions. I suspect Tsai doesn’t really know exactly what she is advocating, and she is interested in some aspects of the idea without understanding that way it all fits together.
Consociational democracy was introduced to the world by Arend Lijphart in the late 1960s. Most studies of democracy looked at the USA and UK and stressed the importance of conflict between parties, the alternation of power between opposed parties, and majority rule. Politics looked very different in Lijphart’s native Holland, and he sought to demonstrate that majoritarianism is not the only basis for democracy. Later in his career, he developed a broader idea of consensual democracy (which is perhaps closer to what Tsai Ing-wen is thinking about), but his first big idea was consociational democracy.
Lijphart’s big question was how democracy could survive in deeply divided societies. In the Netherlands, there were deep divisions along religious and class lines. It is not just that Catholics and Protestants were aligned against each other, but that their mutual antipathy was so strong that unfettered American style competition might lead to violence or even civil war. Likewise, the conflict between the liberals and socialists was extremely intense. (To give an example of a deeply divided society that might be more familiar to Taiwanese readers, think about the ethnic Chinese communities in Malaysia or Indonesia. These groups typically live in separate communities, speak different languages, have different school systems, practice different religions, and feel a certain sense of vulnerability because their obvious differences with the rest of society makes them a convenient target. Indeed, anti-Chinese violence occurred in the upheavals of the late 1990s.) If one group in Holland managed to win an election, had complete control of the government, and implemented its favored policy (as in the UK), the probable result would not have been that the losers collectively shrugged their shoulders and sighed that the other side had won the election and had the right to govern (as in the UK). Instead, there would have demonstrations in the streets, rioting, and perhaps civil war or the collapse of democracy. When the people are at each others’ throats, how can you have a democracy?
As Lijphart described it, the Dutch solution was based on limiting conflict at the mass level and relying on negotiations at the elite level. There are four key features to consociationalism: proportional representation, grand coalition, federalism, and minority vetoes.
Elections were run under proportional rules. If a party got 30% of the votes, it got 30% of the seats in parliament. Since political parties represented a specific group in society (ie: Catholic socialists), that meant that if the Catholic socialists made up 30% of the population, their numbers would be proportionally reflected in the parliament.
All parties in the parliament got to share in political power. In a Grand Coalition, all parties are included. There is no opposition party. Each party got cabinet posts in proportion to their seats in the parliament. So the Catholic Socialist party might get 30% of the cabinet posts. The biggest party typically got to be Prime Minister, but the Prime Minister was not particularly powerful. All major policies had to be negotiated out in the cabinet, where all parties were represented by their leaders.
This system meant that it wasn’t important to try to win an extra 3% in the election. If you had 30%, you would get 30% of the power which really isn’t much different from 27% or 33%. This is especially true if winning an extra 3% means making compromises to appeal to voters who don’t have quite the same interests. Instead, each party simply mobilized its base in each election. The Catholic Socialists would not have bothered with trying to win votes from people who normally supported the Protestant Liberals.
Consociationalism features a federal system. At the very least, it requires that local governments are very powerful relative to the central government. In deeply divided societies, a minority group might not have the power to determine public policy in the national government, but it would certainly have the power to do so in its own locale. Of course, this assumes that society is geographically segregated, and Catholic socialists don’t live together with Protestant liberals. In Holland, it was the case that Catholics and Protestants tended to live in different regions. So in the Catholic areas, local governments made choices that Catholics approved of, while the Protestants made choices that they liked where they were the majority. The federal bargain was that neither side would try to dictate policies for people in the other areas.
The last feature is minority veto. In the UK or USA, it is accepted that majorities have the right to determine government policy. In Holland, it was accepted that minorities had the right to veto any policy in their most sensitive areas. For example, if the central government tried to impose a new policy on education, Catholics might veto it, claiming that education is critical to Catholics’ sense of identity and community. This veto was exercised informally at the cabinet level. When party leaders insisted that they could not accept a result, the majority would agree not to impose it on them. The central government simply refrained from governing in these policy areas crucial to the interests of particular minorities unless everyone could reach a consensus. If there was no consensus, the choice was left up to each local government.
By now, I hope you are thinking that this model doesn’t fit Taiwan well at all. For one thing, Taiwan simply is not that deeply divided. There are distinctions between Mainlanders, Hakkas, Min-nan, and Aborigines, but they are not that deep. The groups intermarry, live in the same neighborhoods, work in the same offices, go to the same KTVs, worship at the same temples, and generally get along well. There is no sense that ethnic-based rioting is lurking right beneath the surface and could erupt at any moment.
Taiwan’s society is not sufficiently segregated for powerful local governments to serve as a compromise mechanism. Taipei City, for example, has a 60-40 partisan balance, not a 90-10 balance. If local governments were allowed to write the history section of textbooks, this wouldn’t mean that all children were learning a version of history that their parents approved of.
How would Taiwanese voters respond to a Grand Coalition? First, what is NOT a Grand Coalition? Inviting a few members of another party into the cabinet does not qualify. Chen Shui-bian made Hau Lung-bin (New Party) head of Environmental Protection. However, Hau joined the cabinet as an individual. Chen did not negotiate with the New Party and the New Party did not agree to support the actions of the government. There was no coalition at all. In a true coalition, each party negotiates on behalf of its members, and, as a condition of entering the coalition, the parties agree to support the government in the parliament. If a party wants to cease supporting the government, it has to give up its cabinet posts.
Imagine the DPP winning a majority of 60 seats and the presidency in the next legislature and then announcing that they would ask the KMT to fill the Vice Premier, Interior, Transportation, Mainland Affairs, Agricultural, Economics, and Labor Affairs spots. How would DPP supporters react to that? Wouldn’t they feel betrayed? For that matter, would KMT supporters feel queasy about taking all those posts under a DPP-led cabinet? Now imagine that no policy gets passed if any major party (DPP, KMT, PFP, and TSU) opposes it. Nothing would happen. The nominally DPP government would not be able to do anything differently. Is that what voters (hypothetically) asked for?
Grand Coalitions essentially make the elections unimportant. No matter how well or poorly a party does, every major party will get into the government and have a veto over major government policies. This is why Grand Coalitions are typically found in crisis conditions, such as during a major war. During a crisis, the normal arguments can’t be allowed to interfere with the effort to win the war, so they are put aside. The consociational model essentially suggests that the society is in a perpetual state of crisis (due to internal cleavages). Rather than letting voters decide how to solve controversial questions, the only safe way is to let elites negotiate out compromises behind closed doors.
The consociational model simply doesn’t make any sense for Taiwan. I really hope that Tsai is thinking about something different. A little more restraint by governing parties and a little more respect for the preferences of the minorities wouldn’t hurt, but there is no reason to suggest that Taiwan is on the verge of civil war and the only way to save democracy is to make elections meaningless.
 Lijphart was one of the most influential political scientists of the last four decades. He was president of the American Political Science Association, an honor reserved for the top people in the discipline. He eventually migrated to the University of California, San Diego, where he finished his career. I started my PhD program there the year after he retired, so I never directly studied under him. However, we got a heavy dose of his ideas.