Saturday, July 9, 2011

Econometrics Without Borders?

We're all familiar with the "Doctors Without Borders" organization, and the valuable international medical work that it performs. Perhaps you didn't know that there's also a group called "Statistics Without Borders"? To quote from their web site, their mission statement is as follows:

"Statistics Without Borders (SWB) is an apolitical organization under the auspices of the American Statistical Association, comprised entirely of volunteers, that provides pro bono statistical consulting and assistance to organizations and government agencies in support of these organizations' not-for-profit efforts to deal with international health issues (broadly defined)."

It's great to see The American Statistical Association, which I've been a member of for 38 years, supporting this type of venture.

Now, a new initiative, provisionally called "Data Without Borders" (DWB), has been established by data scientist, Jake Porway. You can read about it on Porway's web site, of course, and also in a recent post on The Guardian's DataBlog here. Briefly, the aim is to match important data from not-for-profit organizations with experts in the analysis of data. According to a recent item in the Royal Statistical Society's newsletter, RSSeNEWS, there were over 300 expressions of interest, internationally, within the first 24 hours of DWB being announced.

So, don't let anyone ever tell you that being a "quant" who works with data can't be socially meaningful. Even econometricians can make a difference if we want to!

© 2011, David E. Giles


  1. Dave, I was bemused by your comment that even econometricians can make a difference. As a person whose knowledge of econometrics has not advanced much since about 1971, it seems to me that econometricians do tend to make a lot of difference. Econometrics seems to have become a lot more useful over the last few decades.

    It seems to me that econometrics is even becoming slightly intimidating. When I read papers in which econometric results manage to stand up to a battery of tests including things like Bayesian averaging over classical estimators (BACE) my confidence in the results is increased. Yet I am not sure that I should be so impressed. I would probably be less impressed by the results of a BACE test if they were contrary to my prior knowledge of how the world works. It hasn’t happened yet – and perhaps it isn’t likely to happen in areas where my prior knowledge is based on tested theory. Yet I feel slightly uncomfortable that the distinction between testing theory and constructing theory is being lost. Do I have reason to be concerned?

  2. Winton - Thanks for the comment. I agree that econometrics has become a lot more useful in recent years. Apart from anything else, it's good to have access to richer, more timely, and more relevant data-sets than before, and advances in computing haven't hurt either!

    On your other point, personally I think you do have some reason for concern. Maybe I'm old-fashioned, but take a Popperian view. There needs to be a (theory-based) hypothesis that we then test and either reject or do not reject (yet).