Thursday, July 21, 2011

Rules-based Keynesian Policy?

John Taylor, who is one of the most prominent academic critics of administration and Fed policy over the past several years, grapples with the label "anti-Keynesian" that was pinned on him by The EconomistHe writes:
In a follow-up to the Economist article, David Altig, with basic agreement from Paul Krugman, argued that it was a misnomer because I developed and used macro models (now commonly called New Keynesian) with price and wage rigidities in which the government purchases multiplier is positive (though usually less than one), or because the Taylor rule includes real variables in addition to the inflation rate. In my view, rigidities exist in the real world and to describe accurately how the world works you need to incorporate such rigidities in your models, which of course Keynes emphasized. But you also need to include forward-looking expectations, incentives, and growth effects—which Keynes usually ignored.

In my view the essence of the Keynesian approach to macro policy is the use by government officials of discretionary countercyclical actions and interventions to prevent or mitigate recessions or to speed up recoveries. Since I have long been critical of the use of discretionary policy in this way, I think the Economist is correct so say that I am anti-Keynesian in this sense of the word. Indeed, the models that I have built support the use of policy rules, such as the Taylor rule for monetary policy or the automatic stabilizers for fiscal policy, which are the polar opposite of Keynesian discretion. As a practical prescription for improving the economy, the empirical evidence is clear in my view that discretionary Keynesian policy does not work and the experience of the past three years confirms this view. 
"Keynesian" means different things to different people - at its broadest, it means accepting that there are frictions in the economy which mean that aggregate demand matters and policy can have real effects.  This is in contrast to the pure classical view, in which Say's law holds, demand is irrelevant, and output depends on technology and preferences.  In the version of Keynesian economics in our undergraduate textbooks - the IS-LM/AS-AD framework - the frictions are nominal rigidities and the Keynesian model deals with "short run" fluctuations around a "long run" equilibrium determined by the classical model.  In this setting, both monetary and fiscal policy matter (by shifting the LM and IS curves, respectively), though early Keynesians emphasized fiscal policy and "monetarists" (most prominently Milton Friedman), gave primacy to monetary policy.  The version of Keynesian economics in our graduate textbooks and academic journals - "New Keynesian" - combines dynamic optimization with sticky prices, and explicitly addresses the lack of "forward looking expectations" in the traditional textbook version.  Furthermore, some argue that both the IS-LM and New Keynesian incarnations really miss the point and gloss over more fundamental irrationality and instability Keynes saw in the capitalist system.

As Taylor describes his views of the economy (and from what I know of his academic work), it seems consistent with mainstream New Keynesian economics (though his version has been less favorable to fiscal policy than some others).   His criticism of recent fiscal and monetary policy grows out of another longstanding conundrum in macroeconomics, "rules versus discretion."  He is not claiming that countercyclical fiscal and monetary policy are fundamentally impossible, which is what I would say is the true "anti-Keynesian" view.  Rather, he is arguing that discretionary policy may do more harm than good, and policy should be based on stable, predictable rules. 

A primary argument for rules is that discretionary "fine tuning" is impractical based on "long and variable" lags associated with (i) recognizing the state of the state of the economy, (ii) designing and implementing a policy and the (iii) the policy's impact reaching the economy.  Often lurking behind this argument is a political philosophy that is skeptical of government (no coincidence that Milton Friedman was the most famous proponent of rules - Brad DeLong recently argued this is how he resolved the contradiction between an economics that said monetary policy can be effective with a libertarian political philosophy).

Taylor is careful to say that he opposes "discretionary Keynesian policy" - I think "anti-discretion" might be a better characterization of his critique than "anti-Keynesian."  Of course, that only matters if it is possible to be "anti-discretion" without being "anti-Keynesian."  I think it is.

I don't share the political philosophy, but the experience of the last several years has underscored the practical difficulties of discretionary policy.  The early-2009 Obama administration with large congressional majority is about as close to government by center-left mainstream Keynesian technocrats as the American political system is likely to ever give us.  In retrospect, it is clear they misjudged the scope and duration of the downturn and were not able make adjustments as that became apparent.

Monday morning quarterbacking in April, I suggested that the stimulus should have been designed in a "state-contingent" fashion to remain in place until the recovery reached certain benchmarks.  It is a small step from there to a "rules based" countercyclical fiscal policy - policies like aid to state governments, extended unemployment benefits, payroll tax cuts and even increased infrastructure spending could be designed to kick in and ramp down automatically based on the state of the economy (e.g., with triggers based on the unemployment rate).   To me, that's very "Keynesian", but also "rules-based", and its easy to imagine that might have worked better than the actual policies that were put in place. 

2 comments:

The Arthurian said...

It's good to be consistent, but not when you're wrong. The Taylor rule has an inflation component and a growth component, but no component that considers accumulating debt. Thus policymakers could depend on the rule right up to the moment it started saying the interest rate should go negative.

On the other hand, if one is to adjust and improve the rule as needed... Isn't that discretion?

Bill C said...

Yes - rules have their own set of limitations. They are fashionable academically, but Greenspan, for one, didn't like the idea. I didn't touch on monetary policy in the post (which was long enough already), but clearly the Taylor rule needs some extension to cover the zero lower bound (although Taylor himself believes the fed funds rate should be positive). Also, arguably we need more rules based financial policy - but "no bailouts ever" just isn't a credible rule...