I am attending this conference organized by Marco Battaglini for the first time. Stream of conference blogging (and hence perhaps completely incorrect!).

Eric Weese, Efficiencies of Scale in Local Public Good Production,

If Wilmette and Winnetka become one municipality, they can combine their firefighter services and enjoy economies of scale. Each city council might vote and if both agree, they can become Willetka. How big are such efficiencies of scale? It turns out that this kind of merger scenario arose in Japan and Weese studies. The Japanese central government sought to reduce transfers to local municipalities by encouraging mergers. Before the central government intervened, there was little incentive to merge as it would lead to less transfers. You can restrict the merger model using this phase. This is work in progress so there were some preliminary estimates of the size of economies of scale. Weese can compare his estimates to central government estimates and finds they are reasonably good. The empirical methodology might be useful in other coalition formation scenarios. There are no transfers between municipalities before merger hence there is nontransferable utility. The estimation procedure accounts for this issue.

Steve Matthews, Achievable Outcomes of Dynamic Contribution Games

More general than public good contribution games but easiest to motivate the paper that way. So, suppose players are contributing to a public good. They can contribute as much as they want and as many times as they want. The game has an infinite horizon. But once they have contributed, players can only add to the contribution and not take any money away, so there is some commitment. All achievable contributions lie in the “undercore”, hence there is no Folk Theorem. Gradual contribution can help to resolve the free rider problem as each player becomes “pivotal” in public good provision. This intuition obtains from :threshold” pubic goods with zero-one production. But what about the no threshold case? Suppose an efficient contribution profile is asymmetric and involves contributions by some players but not others. Then, with gradual contribution, at some point, some player prefers not to contribute as he is contributing to other player’s utility more than his own. This prevents implementability of efficient outcomes and hence also shows he Folk Theorem does not obtain. This just one example. There are results for discounting and no discounting for general and specific settings.

Stelios Michalopoulos, Divide and Rule or Rule the Divided

How do precolonial vs post colonial institutions affect economic development? Many ethnic groups end up in different countries after European colonization. Across country within ethnicity variation is studied. And then within country across ethnicity variation. But the GDP data is crap. But there is great satellite data on light density. Lights are public goods and public good provision is highly correlated with income. Surprisingly, the same ethnic group in different countries do not display much variation. But pre-colonial ethnic institutions are the only robust correlate of economic development.

I’ll see how much energy I have to blog on day 2 given I am presenting.