A couple of days ago, Simon Jack of BBC Radio 4’s Today Programme interviewed me about unconventional economic indicators. We chatted about the cranes index, written up by Chris Giles in the FT as a marker of regional imbalance in the UK economy, about hemlines, lipstick, champagne sales. One new to me, unearthed by the researcher, was Alan Greenspan’s supposed interest in sales of men’s underpants. I thought it was in Greenspan’s book, [amazon_link id=”0141029919″ target=”_blank” ]The Age of Turbulence[/amazon_link], where he does talk about his interest in detailed economic statistics, but it turns out the source is a 2008 NPR interview about the book.
[amazon_image id=”0713999829″ link=”true” target=”_blank” size=”medium” ]The Age of Turbulence: Adventures in a New World[/amazon_image]
My discussion on the radio was mildly frivolous, but the light-heartedness covers a serious point about macroeconomics, namely how studiously unempirical it is. This might seem a contrarian statement, given how frequently macroeconomists bandy about debt-GDP ratios, GDP growth rates, inflation and unemployment rates. But in fact Greenspan was something of an exception with his obsession with the statistics underpinning the aggregates. Most macroeconomists trade blows with the same aggregate figures drawn from the same online databases, and their differences are disagreements about the interpretation of the figures in the light of their prior beliefs about a ‘true’ model of the economy. They demonstrate confirmation bias in finding aggregate figures to support their views.
One of the problems with macroeconomics, therefore, is how little attention its practitioners pay to either understanding the construction and intellectual framework underpinning the aggregate statistics they do use (none of them being natural entities, all analytic constructs – see my forthcoming GDP book), or to collecting new statistics. So I’m with Alan Greenspan on this point, and think sales of underpants could be more revealing than the conventional figures.