Engaging today's political economy
with truth and reason

sponsored by

Rule by Numbers

22 Nov 2014

I am here at the Southern Economic Association in Atlanta, Ga, where it is very nice, pleasantly warm weather, and where numbers and statistics flow like water.  The first session I attended was on measures of economic freedom, and interesting topic on its face.  And the papers were interesting, but excruciatingly painful, looking at factor analysis, regressions, and all those items political science and economics students learn but many seem to hate–and even some scholars in those fields hate them.  Well, I now hate them again too, though it might be more accurate to say they annoy me, but not for the reasons you might think.

I am not anti-empiricist.  Nor do I hate math.  In fact I very much enjoy it, as I was a Physics major with lots of math, and had statistics and methods in graduate school.  So it isn’t a visceral dislike.  It seems to be, if I can frame the right words, that we are asking our numbers to do too much.  It was enlightening that one of the presenters was even asked to co-write a paper almost solely because, as he himself put it, he had no position at all regarding the research.  He was just “letting the numbers speak for themselves,” totally unbiased, with seemingly no presuppositions or biases.

What do I make of that?  Perhaps I am naive, though I don’t think so, but the “numbers people” seem to want to have us believe in a sort of “Baconian fallacy.”  That is, we can, they almost say, just go look for correlations and let them fall where they may.  We gather statistics and “run” the data and see what emerges.  But then we don’t know how to explain it.  In fact, that is probably my main objection.  We have become so narrow and technocratic that no one (not literally of course) knows how data and its manipulation is to be interpreted properly.  And I am solidly on the side that argues that data does have a proper and an improper interpretation.

Again, there is value, even great value, in empirical work.  But it really doesn’t do that much good unless it can be used appropriately–and used at all–to help us live better.  There are people who do this well.  I have met some.  And there are those who can’t help us except in a limited way.  Finally, when it comes to interpreting data, including after correlations have been drawn, we can do no better than to have a Christian worldview as a lens of interpretation.  While that sounds like mixing oil and water, remember that empiricism is nothing but a different way of knowing, which itself ought to be subject to an ultimate source of knowing for its “boundary conditions,” its limits.  Otherwise, empiricism develops its own hubris, as indeed it has in the last 250 years.

So on to other sessions today.  It gives me pleasure to say that my own paper tomorrow contains no numbers except for section headings.  OK, false humility.  Sorry, but I did enjoy it.