The article, “A Natural Experiment in Proposal Power and Electoral Success,” by Peter J. Loewen, Royce Koop, Jaime Settle, and James H. Fowler, appears in the January 2014 issue of the AJPS. Here, Peter Loewen provides a summary:
Common sense would suggest that politicians should be rewarded for the work they do in a legislature. After all, in the real world we hope that employees get rewarded for doing their jobs. Political scientists are not sure this is the case with politicians, however. After all, voters don’t follow politics closely, especially the finer details of the legislative process. If voters don’t watch their politicians closely, how can they know who to reward and who to punish? There is some evidence to the contrary, however, namely that more active politicians are in fact more richly-rewarded come election time.
Imagine that we did observe all the legislative action politicians took, and we then correlated more action with a greater vote share. This could be evidence of voters rewarding politicians for taking action. But the relationship could also be caused by something else. For example, more talented politicians might be able to take legislative action and earn votes through other means, at the same time. Our paper uses data from a “natural experiment” among the Canadian House of Commons to solve this problem.
In the House of Commons, the right of backbench legislators to introduce legislation is assigned randomly, by means of a lottery. This makes it much like in a laboratory experiment. Because there is no relationship between the right to introduce legislation and a legislator’s ability, we can estimate the independent effect of proposal power (and presumably actual proposals) on electoral success. We find a modest effect – about 3 percentage points in the next election – among members of the governing party. In other words, voters provide some reward for legislative action. Further data analysis suggests that this is because members with the power to propose legislation can use it to cultivate a greater personal vote.