Do elections do what we think they should? If not, what does that suggest about how to organize democracies?
I’m thinking of this because of an aside in a recent Vox post about Barack Obama and Chris Rock in which Ezra Klein wonders about “one of the more interesting counterfactuals of the Obama era: What if Obama had taken office a bit later?” — that is, after the recession was a little further along.
Let’s think of two different scenarios.
Suppose the recession, which began in late 2007, had started three to six months earlier, and the crash of mid-September 2008 had happened in March or April 2008. In this timeline, Democrats do even better in November 2008, winding up with 60 or more seats in the Senate even before Arlen Spector flips from Republican to Democrat.
With unemployment already sky high when Obama takes office (and with bailouts more firmly associated with George W. Bush), Obama’s honeymoon dissipates much more slowly in this scenario. Pundits give Obama high marks for dealing with Congress as his proposals sail through. The Republicans’ electoral recovery in 2010 is much more modest, possibly even failing to win back the House majority. Democrats easily win unified government in 2012.
Now imagine instead that the recession didn’t get started until spring 2008, and the crash happens not in September of that year, but after the 2008 election. In this timeline, Obama still wins the presidency against Senator John McCain, though more narrowly, and Democrats fail to approach 60 Senate seats.
Pundits bash Democrats for being unable to get things passed in Congress. Senator Harry Reid, doomed to lose in 2010, is unable to defeat Republican filibusters. Obama’s approval falls to less than 40 percent by the end of his first year, and never recovers before he is defeated in 2012.
The differences between these two timelines are potentially enormous when it comes to policy. In the first, health care reform passes fairly easily, and it’s possible that a climate bill accompanies it. Democrats might also have passed voting reform, legislation to help unions and perhaps even D.C. statehood. In the second scenario, none of that happens, and the Affordable Care Act is as much of a legislative fiasco as President Bill Clinton’s plan was in the 1990s.
What drives all this consequential change is something random: the way economic cycles coincide with political cycles. And it’s the non-economic policies which would change. The odds are that the policies most relevant to the economy are relatively unaffected. No, it isn’t certain that Bush would have bailed out the auto industry, but chances are he would have. Some large stimulus bill probably happens in each timeline, too.
Voter preferences on individual government policies don’t really change on the different timeliness. The same people who were for federal health care would have wanted it no matter when the crash occurred.
The real action in these election scenarios is being driven by voters who mainly care about the overall economy, and how their behavior changes depending on what condition it is in. Yet this behavior would cause huge shifts on policies for almost everything else.
All of this demonstrates how troublesome the “responsible party” way of thinking about democracy can be. This is the idea that parties should have strong (and opposing) agendas, and that government institutions should allow the winning party to carry out those policies.
It’s supposed to be democratic because it offers voters a real choice on issues. But real voters don’t make that sort of real choice — and especially not the real voters who determine election results, at least when the parties are more or less balanced.
In short: Elections don’t and can’t reflect voter preferences about public policies.
This doesn’t mean we should despair about democracy or about elections as one of its central institutions. For one thing, no matter how blunt a weapon elections are, it’s good to give the governing party a strong incentive to produce outcomes that voters like.
Elections are also crucial for building strong representational relationships between politicians and their constituents. Politicians make promises to voters. Then they govern with their interpretation of those promises in mind — and how they will explain their actions to voters in the next election, when they also make new promises and the cycle begins again. But if we see them as registering the preferences of voters on the issues, and regard them as definitive, then we weaken democracy because we put too much weight on the randomness of those results.
Better to make policies based on bargaining and deal-making, in which election outcomes matter (they strengthen or weaken various bargainers) but are less central. For that, the U.S. system of separated institutions sharing powers, with staggered elections and different though overlapping constituencies, is effective in reducing the importance of a single election or group of elections. So were the less ideological, less coherent political parties of the past.
This means we should be skeptical of claims that a winning party “should” be able to enact what it wants, or that a party “should” have a clear platform and then stick with it.
The virtues of that kind of democracy are overstated and its vices are overlooked.