Attributing Policy Influence under Coalition Governance
By David Fortunato, University of California, San Diego, Nick C. N. Lin, Academia Sinica, Randolph T. Stevenson, Rice University and Mathias Wessel Tromborg, Aarhus University
In the overwhelming majority of democracies, it is very rare that one party has majority control of government, allowing it to draft, pass, and implement policy on its own. Most often, compromise must be struck between legislatures and presidents, or, in parliamentary systems, across different parties within governing coalitions. These compromises are often complex, or opaque, and there are very rarely outright “winners” and “losers” in the negotiation process. This presents voters with an important problem to solve: when multiple actors have influence over policymaking, how can we know how much credit or blame each party deserves for policy outcomes?
In our recent article published in the APSR, we offer a solution to this problem. We argue that voters can use a simple heuristic model to infer how much influence parties had in the policy-making process. A heuristic model is a method of making educated guesses in which a person takes some simple piece of information they have on hand and turn it into an inference by applying a simple rule. For example, if I were to ask you which city is largest: Frankfurt, Stuttgart, or Leipzig; you probably automatically guessed the correct answer (Frankfurt), even though you’re not a German demographer. What you’ve likely done is applied the “recognition heuristic” to solve the problem – you subconsciously intuited that the most familiar sounding city is largest, and in this case, if you guessed Frankfurt, you’d be correct. The recognition heuristic is a powerful for guessing which city has the largest population, because, as cities grow, they tend to have more of the characteristics that makes them well known. Frankfurt has a large airport, beautiful museums, a bustling finance sector, professional sports teams, etc. Similarly, this heuristic is excellent for guessing the likely winner of tennis matches, because tennis players become more well known as they get better and better.
The heuristic model that we suggest voters can use to infer policy-making responsibility across many parties relies on voters knowing what role the party plays in government – does it provide the prime minister, is it partner in coalition to the prime minister, or is the party in opposition? – and its relative size in parliament – about how many seats does it have? If voters infer that prime ministerial parties have more influence than their partners in government, that those partner parties have more influence than opposition parties, and that, all else equal, larger parties have more influence than smaller parties, they can probably make reasonably accurate (though certainly no perfect) attributions of policy-making influence. Why is this? Because, just as a city’s name recognition is correlated with its population, parties’ size and role really are the best predictors (as far as we know) of their policy influence. This means that people who use this heuristic model are unlikely to make big “mistakes” or “get it wrong” and so the model should be “stable” – it should be used widely and consistently across the electorate.
How do find out if voters are using this model? We ask real people in Denmark, Germany, Italy, Netherlands, and the United Kingdom how much policy influence they believe all of the parties in their legislature have and we also ask for their perceptions of the sizes and roles of all of the parties. This is very important because it allows us to find evidence for our model. That is, because size and role really are correlated with policy influence, if every had the same (accurate) perceptions of parties’ size and role – or, if we just assumed everyone had the same (accurate) perceptions of size and role, rather than asking (and political scientists do this all the time) – then we wouldn’t be able to distinguish between heuristic users and people who just knew how much influence parties had; the heuristic model and perfect information would be observationally equivalent. If, however, some people mistakenly believe a party provides the prime minister, when in reality it is partner to the prime minister, and also believe it to be more influential as a result, or, if some people mistakenly believe a large party is smaller than it actually is and also believe it to be less influential as a result, then we can say with confidence that people are using the heuristic model.
Our design allows for just this kind of comparison. We compare people’s perceptions of a party’s role and size to their evaluations of that party’s policy influence. As it turns out, people who believe a party provides the prime minister perceive it is being more influential than people who believe that same party is partner to the prime minister, and, people who believe a party is small also perceive it is less influential than people who believe that same party is large. There is really strong evidence that people use this heuristic model to infer policy responsibility.
We believe that this is good news, because it suggests that regular people have adapted a set of powerful tools to hold their democratically elected leaders accountable for their performance. Not everyone has perfect information, but the overwhelming majority of people in our study use their information in really reasonable ways to solve a problem that many political scientists believed was prohibitively difficult, or maybe even impossible.