Jonas Hedegaard Hansen (https://sites.google.com/site/hedegaardhansen/home):
It is well documented that citizens who move are less likely to vote than citizens who are residentially stable. Furthermore, this relationship is not only because people who move differ from residentially stable citizens – moving has an independent negative effect on turnout.
We know less about the reasons for this effect though. I suggest that this difference stems from three types of costs movers face. First, convenience costs (e.g., registration to vote and time spent going to the polling station) increases when people move. Second, informational costs (e.g., obtaining information about where to vote and for whom to vote) related to voting increases when citizens change residence. Third, the social costs that are associated with voting, which cover the social rewards from voting at a polling station, and expose this act of citizenship to a broader community, also increases. By including five types of residential patterns and information about polling station assignment in the same analysis, this study brings us a step closer to understanding the relevance of the three types of costs in explaining turnout.
In my article, I have access to extraordinarily rich and reliable panel data, which includes validated voter turnout data for approximately half of the Danish population in two recent municipality elections. These are elections where voters are automatically registered to vote. On top of this, I have access to data on their exact address at any given time as well as the timing of their move. I also know their polling station assignment and have access to a large number of socio-demographic variables. Using this variety of data, it gets possible to divide the citizens into five different groups conditional on their residential status and polling station assignment and thereby test the three types of costs for the negative effect of moving on turnout. Without going into further theoretical discussions in this blog post, I highlight three central findings from the paper.
First, the negative effect of moving on turnout is primarily caused by increased social costs associated due to moving away from one’s neighborhood. When people move, the social rewards of voting decrease. The analysis shows that average marginal effect of moving on turnout is approximately -4 percentage points for both citizens who move to a new municipality and for citizens who move within the same municipality. This indicates that the extra informational costs that are related to learning a new political agenda and obtaining knowledge about new politicians when crossing a municipal border does not have a substantial impact on turnout beyond the costs of simply moving.
In addition to this, the data allows to divide the movers within a municipality into two groups: One group stays assigned to the same polling station, meaning that they have only moved a short distance. Another group moves within the municipality but to a new polling station district. The analysis shows that the effect almost does not differ between these two groups, despite of the fact that the movers who votes at the same polling station have almost no changes in convenience and informational costs. This is another indication of the importance of the close social environment and neighborhood for turnout. Disruptions of citizens’ social surroundings seem to be a key explanation for the negative effect of moving on turnout.
Second, the analysis shows that the negative effect of residential mobility on turnout is particularly large for citizens moving just before the election. For instance, the average marginal effect of moving within a month before Election Day is approximately -12 percentage points. For movers changing residence between one and three months before the election, the same figure is around -5 percentage points. These results show that large events in citizens’ everyday life can distract their attention away from politics. While this idea has been around for quite some time, there have not been any empirical analysis showing how quickly the effect wears off after people settle down.
Third, it is important to note that the study was conducted in an institutional setting with substantially smaller convenience costs compared with elections that require movers to re-register to vote. As such, my study can only provide a hint as to whether making voter registration automatic in other contexts will solve the problem of depressing turnout for movers. The analysis suggests that the negative effects of moving on turnout are not likely to disappear simply by making voter registration automatic.