A hectic week

It's been a hectic week.

Last weekend I came to conclude that some kind of democratization does increase government consumption (see 19th October). If the dictator holds democratic elections without running for office himself, government consumption goes up. If the dictator runs for office, either winning or losing, government consumption does not change.

Last Monday I attended Torsten's lecture, realising one of his latest working papers deals with almost exactly the same question as the one I've been investigating - how democratization affects government consumption. His conclusion is that democratization followed by the parliamentary form of government increases government consumption while that follwoed by the presidential form of government does not.

Which is correct? The rest of Last Monday was spent on finding it out. The conclusion is Torsten is right and I'm wrong.

The next morning I talked to Torsten on this. He seems happy. :) But it seems to me that this is the result I was looking for. This government consumption project seems to reach its conclusion. So what should I do next?

On Thursday, I talked to Tim on this. When I started talking about my idea that turned out to be wrong, he showed his interest in whether or not the dictator runs for office. I used it as a right hand side variable (a variable explaining something else). He suggested to use it as a left hand side variable (a variable to be explained). Conversations flew and leadership survival in autocracy - when economic conditions make a difference in leadership turnover in autocracy - came out as a promising research topic. There is a huge literature on this in politicla science. But, as usual with political science, there has been no formal theory and their regressions always suffer from endogenous bias (ie. unable to tell the direction of causality). So there is room for an economist (ie. me) to investigate the issue.

Somehow, I saw some light. But the story didn't end here. (To be continued...)