The Epistemic Sequester: Budget Cuts Kill an Important Statistical Program
After already slashing R&D funding, the Sequester is about to deliver another kick in the teeth to American competitiveness: it’s going to sharply reduce our ability to measure it. This one comes courtesy of the Bureau of Labor Statistics, which announced last month that the sequestration has forced it to eliminate its International Labor Comparisons (ILC) program, a neat little database that adjusts foreign data to a common framework, allowing you to compare the traded sector health and competiveness of the United States against that of other countries.
This may not sound like much, but in the nerdy world of competitive analysis economics, it’s huge. No one else provides this data to the same extent as ILC. The OECD does a bit,[i] but their data are rife with warnings about the perils of cross-country comparison among their indicators. Moreover, the OECD has little-to-no data on the big boys such as China and India, which renders its data useless for any “big picture” comparisons of our competitive health. Other organizations, such as the UN Industrial Development Organization, provide limited competitiveness data that is vastly incomplete.
In contrast, the ILC program gives us complete and comparable time series for extremely useful indicators including manufacturing output, hours, compensation and productivity, as well as labor force, employment, price and industry statistics. Their studies of output, compensation and productivity in China and India are second-to-none. On top of this, they provide a handy little dashboard that you download to your computer and which provides you with competitiveness statistics on demand.
ILC is not a set of abstract, pie-in-the-sky statistics. Rather, these statistics are crucial to current economic policy decision-making in a world of globalized trade. These are the day-to-day statistics that everyday people wish they had on those rare occasions when anything from the lack of jobs to the decline of manufacturing to free trade comes up at the dinner table. Try it. Have a couple wines and go argue trade policy with a neighbor. Inevitably, China will come up, and more specifically, labor costs in China will come up. If you really get into it, maybe even Chinese productivity will come up. Who’s right? Where is this information? I’ll tell you where it is: the BLS ILC program. If these are relevant statistics to dinner table discussions around America, then these are relevant statistics to economic policymakers in Washington.
But, wait, there’s more! And it goes a bit deeper. In the bubble of Washington, where economic thought is endlessly dominated by neoclassical ideology, it is exceedingly difficult to penetrate the bubble with policy ideas that don’t already fit into the neoclassicists’ theories of how the American economy works in a globalized world. This applies to our statistical system too: since neoclassicists dominate Washington, they also tend to be the ones managing our economic statistics, deciding what to measure, how often to measure it, etc. Not surprisingly, they tend to favor statistics which measure parameters that are already incorporated into their neoclassical models. Those that do not fit are dismissed. Economic policy in Washington then becomes a self-sustaining cycle of epistemic closure: neoclassical thought supports neoclassical statistics, and neoclassical statistics support neoclassical thought. But more on that later. For now, the tragedy of the ILC program is that it was one of the very few bastions we had in the Washington bubble that housed traded sector statistics deemed unimportant by neoclassical theory. ILC has helped us, and others, to confidently suggest that there might be something other than productivity, redistribution or a financial crisis causing our endless economic malaise. And now it’s gone.
For now, if there’s a chance to save or restore ILC (and that’s a big “if”), you can sign a petition here. And, of course, you can write to your congressman.
[i] For example, see their STAN database, which is pretty useful, although very limited in scope.
