What if the economy decided the Dutch election?

Dutch election forecast : What if the economy decided the Dutch election?

Philippe Mongrain (Université de Montréal), Ruth Dassonneville (Université de Montréal) and Michael S. Lewis-Beck (University of Iowa)


While most observers are covering the Dutch election as a litmus test for populism in Europe, focusing on issues such as the European Union, immigration and the PVV-leader Geert Wilders, these issues might actually be a lot less important than the traditional bread-and-butter issue of the economy.

If the election is not (only) about populism and Wilders, what would it be about? Building on a rich literature in the field of electoral research, it would not be surprising if the economy had an important impact on Dutch voters’ choices and on the outcomes of today’s election. In many democracies, we find that economic conditions are systematically correlated with how the incumbent government performs on Election Day. That is, governments tend to be punished for worsening economic conditions while they are rewarded in times of economic growth.

We have verified to what extent electoral results in the Netherlands are generally associated with economic indicators. Despite the fact that the Dutch party system is strongly fragmentized, we do find a clear link between economic indicators such as GDP or unemployment rates and the performance of the incumbent.

Using these insights, we have built a simple and parsimonious model to predict electoral results in the Netherlands. We have developed a model to predict the vote share of the Prime Minister’s party (in this case the VVD of Prime Minister Rutte). Our key indicators are the incumbent’s vote share in the previous election, a measure of GDP growth in the year before the election and an indicator of the number of months the incumbent has been in office. This last variable takes into account the possibility of a cost of ruling-effect.

This simple model performs remarkably well. Applying it to previous elections in the Netherlands, the model predicts the Prime Minister’s party’s vote share with an average error of 3.7 percentage points.

If we apply this model to the 2017 elections, what are our predictions? We estimate that the VVD of Prime Minister Rutte will obtain 22.3% of the votes.

This prediction is higher than what the latest polls suggest will be the outcome of the Dutch elections. The latest polls suggest the VVD will obtain only 17% of the votes. Obviously, polls come with error, and so does our very parsimonious forecasting model (recall the average error of our model in previous elections is 3.7 percentage points).

Both of types of predictions thus have to be taken with a grain of salt, but based on our forecasting model we would suggest that the VVD is likely to do quite a bit better than what polls suggest. If Dutch voters are taking into account the state of the economy today, Mark Rutte’s party will outperform its standing in the polls.

For more info, see Dutch election forecast-final

This content has been updated on 15 March 2017 at 21 h 55 min.