There is a movement in medicine to require that applications for licenses to sell a new drug be "evidence-based." By contrast, trained economists view their discipline as having already achieved this scientific standard. After all, they express their ideas with mathematics and arrive at quantitative estimates of implied relationships from empirical data.
But economics is not evidence-based in selecting its theoretical paradigms. Economic policy initiatives are often launched without a sufficient amount of empirical testing.
A notorious example is postwar macroeconomic policymaking under the radical Keynesians. The radicals relied on Keynes's untested theory that unemployment depended on "effective demand" in relation to the "money wage," but their policy ignored the part about wages and sought to stabilize demand at a sufficiently high level to ensure "full" employment.
Cecil Pigou and Franco Modigliani raised the objection that if demand were successfully increased, the money wage level would rise, catch up to demand and thus push employment back down to its previous level. Employment cannot be sustained above its equilibrium path by inflating effective demand.
Nevertheless, the radicals prevailed despite what economist Harry Johnson called "scorn and derision."
Postwar macroeconomic policies were dedicated to "full" employment, without any evidence that money wages would not get in the way.
In the late 1950s, neo-Keynesians finally conceded the point raised by Pigou and Modigliani. Work on wages by Will Phillips's gave them no choice. But they still insisted that steady increases in demand at a fast enough rate would keep demand one step ahead of the money wage level, so that employment could be kept as high as desired, albeit at the cost of steady inflation.
In different ways, Milton Friedman and I objected, arguing that such a policy would require an ever-rising inflation rate. Money wages will lag behind demand, I argued, only as long as the representative firm is deterred from raising wages by the misperception that wages at other firms are already lower than its own -- a disequilibrium that cannot last for long.
Like the radicals, the neo-Keynesians did not engage their challengers with empirical testing. The efficacy of high demand was a matter of faith. Yet events in the 1970s put that faith to a cruel test. When supply shocks hit the US economy, the neo-Keynesians' response was to pour on more demand, believing it would revive employment. There was little recovery -- only faster inflation.
The current era offers a parallel. Although policy has since shifted to reflect supply-side economics and real business-cycle theory, the new reigning paradigm's builders and promoters display the same antipathy to checking data for serious errors.
An earlier classroom lesson was well-founded: tax rates on labor this year that are below normal, when merged with the prospect of reversion to normal rates next year, will encourage households to squeeze more work into this year and to work less in future years. This was recently tested anew with data from Iceland and did well.
But the supply-siders jumped to the daring conclusion that a permanent cut in tax rates on labor would permanently encourage more work -- with no diminution of effectiveness. Larry Summers and I both doubted that this could be generally true. If every increase in the after-tax wage rate gave a permanent boost to the amount of labor supplied, we reasoned, steeply rising after-tax wages since the 19th century would have brought an large increase in the length of the workweek and in retirement ages. But both have fallen and in Europe unemployment has increased.