Hal R. Varian

In my quest to keep upto date with the big ideas in econometrics and finance I happened upon Hal R. Varian. Hal is both a professor at UC Berkley and also Chief Economist at Google, a couple of impressive positions held simultaneously positions held simultaneously.

There is a trove of Hal’s essays and papers available here, I particularly enjoyed his comment on big data in economics which essentially espouses economists to not only analyse the electronic data that pretty much any firm is amassing but also to perform experiments within the bounds of the collection system, giving rise to similar advances as have been gained from similar methods applied to e-marketing (see Patrick McKenzie’s essays for more).

Another paper of note was Predicting the Present with Bayesian Strucural Time Series (or see the slides for an abbreviated view) in which is described a system for economic nowcasting using a system where terms correlated to an economic variable e.g. claims for unemployment, are discovered with Google Correlate and then Google Trends data for those terms is analysed and ranked in terms of predictability using a Markov chain Monte Carlo (MCMC) algorithm - interesting!