A drop in cohesion predicts ousting of cabinet

With Antti Pajala we have built a database of roll-call votes in Finland for the last decade and a half or so. One of the things we focused on was examining the cohesion of the Eduskunta (Finnish parliament). Perhaps the most interesting thing we found was that if in the year preceding the elections, the cohesion of either the cabinet (and sometimes also of the opposition) would drop, the structure of the cabinet would be replaced with a different group of parties.

finnish-cohesion2.png

Notice the drops when the cabinet composition changed in 1995, 2002 and 2006. No change took place in 1999, however, and there was no change in cohesion either: so our approach predicts it correctly every time. My causal theory is that as MP’s start disliking each other, they will no longer be as motivated to seek consensus and the dislike will also reflect in the next election, but Piero says that dislike is in turn caused by the cabinet losing popularity in the public and followed by the emergence of competing factions.

Our working paper (only written in Finnish, however) is here. It would be interesting to see if similar phenomenon shows up in other parliaments. It’s always cool to be able to forecast elections a year in advance.