1. Limits to Arbitrage: Understanding How Hedge Funds Fail by Getmansky and Lo. The agent-based approach was interesting and yielded interesting (to me) insights. The paper is very rough and neither Andrew Lo nor Mila Getmansky have updated it. It occured to me that perhaps the agent-based approach is still a little ahead of its time despite all the talk about the failure of models.
... We propose that given positions are well diversified and notclosely correlated, leverage by itself does not lead to the collapse of a fund.Correlated positions in the absence of leverage might lead to a loss, but are notsubject to collateral collapse. However, the superimposition of both leverage andinduced high correlation between assets can lead to a collapse.
2. Evaluating the Impact of Development Projects on Poverty: A Handbook for Practitioners by Judy Baker. This is pretty much what it sounds like and the various evaluations undertaken by the World Bank in the annex was to me the most interesting part. Having been part of randomization projects (one clinical and one early childhood) I can attest to its complications and difficulties in terms of implementation and the annex shed some light on this. Various cost estimates were also illuminating - these ran much lower than the ones I was in and perhaps it was because these were in less developed countries. The stories one can tell from these projects could probably fill volumes and would definitely make one pause to wonder whether randomization really is the gold standard? I certainly do.