Why does this building drive some people crazy?

Photographer: Andrew Harrer/Bloomberg

A Backdoor Way to End the Fed

Noah Smith is a Bloomberg View columnist. He was an assistant professor of finance at Stony Brook University, and he blogs at Noahpinion.
Read More.
a | A

Some Americans want to end the Federal Reserve. Most of us rightfully see that as a radical step that would involve the wholesale destruction of a major U.S. institution. We would no longer have a central bank, which would make us unique among rich countries. So the cause of abolishing the Fed would seem to be a lost one. 

There is, however, another way to neutralize the power of the Fed that would be far easier politically. This is the notion of passing a law to tie the Fed’s interest-rate setting policy to a simple arithmetic rule. If that were done, most of the giant bureaucracy of our central bank could be replaced by an algorithm that could run on an app coded in an hour by a high school student. 

For Fed opponents, that is an enticing prospect.  

So it’s no wonder that some congressional Republicans have embraced the policy proposals of Stanford economist John Taylor. Taylor has campaigned tirelessly for the Fed to be forced to set policy according to a rule, and the rule he demands is the one with his name on it. He recently got into a high-profile argument with former Fed Chairman Ben Bernanke over the subject. 

Taylor, who served in George W. Bush's administration,  introduced the Taylor Rule in 1993. The rule is mathematically extremely simple. Interest rates are a linear function of two things: 1) the deviation of gross domestic product from its target level, and 2) the deviation of inflation from its target level. In other words, if gross domestic product is growing too fast, you raise interest rates. If it’s too slow, you cut rates (as  Bernanke did in the Great Recession). If inflation is too high, you raise interest rates (as Paul Volcker did in the late 1970s and early 1980s). If inflation is too low, you cut interest rates. And the rule tells you exactly how much to raise and cut. 

If the Fed were legally constrained to follow the Taylor Rule, all it would have to do would be to set the inflation and GDP targets and stand back. Monetary policy would run itself. The vast array of macroeconomists working at the Federal Reserve Board would find themselves redundant, and both they and their academic counterparts would have to find something else to do with their time. Almost as good as ending the Fed! Taylor has insisted that his rule is not as “mechanical” as people allege, but he never explains exactly how it’s not mechanical, and it’s hard to see how it could be anything but. 

Writing the Taylor Rule into law, however, is a bad idea, John Taylor’s evangelism to the contrary. There are many reasons for this. Let me go over just a few. 

First of all, the amount that the Fed should raise or cut interest rates in response to fluctuations in GDP and inflation -- the optimal values of the coefficients in the rule -- are not clear. Taylor himself proposed coefficients of 1.5 on inflation and 0.5 on output, meaning that a one percentage point rise in inflation relative to target would cause interest rates to be raised by 1.5 percentage points, while a one percentage point fall in output relative to target would cause interest rates to be cut by 0.5 percentage points. If you think about these numbers, it’s clear that this represents a fairly tight monetary policy -- it basically means that the Fed should care more about fighting inflation than about fighting recessions. 

After Taylor’s 1993 paper, Taylor-type rules were incorporated into the dominant strand of models that macroeconomists now use to think about monetary policy -- New Keynesian models. In those models, a Taylor-type rule ends up being the best policy, provided that economic fluctuations are not too extreme. But when New Keynesian researchers calculated what the optimal rule would be, they generally found that it should be much more focused on fighting recessions, and much less worried about fighting inflation, than Taylor had initially proposed. 

Taylor, however, has stuck to his guns and insisted that his original numbers “work well.” The research backing up his assertion is, to put it mildly, very flimsy. If Congress legislates a Taylor Rule but chooses the wrong numbers, monetary policy might do more harm than good. But if it leaves the numbers up to the Fed, then monetary policy would be basically unchanged from what it is right now. 

Another problem with the Taylor Rule is that it isn't clear how it would help stabilize the financial system. Taylor himself has often asserted that it was Fed deviations from his rule that caused the global financial crisis of 2008. But as former Bank of England economist Tony Yates has testily pointed out, the models that support the use of a Taylor-type rule don’t make any attempt to account for the role the financial industry plays in the economy! In other words, the Taylor Rule wasn’t designed to prevent financial crises, so it doesn’t make any sense to think that it would help at all in this regard. 

These are not the only problems with legislating the Taylor Rule. There are many more. For example, since interest rates can’t fall below zero, the rule in some situations will tell the Fed to do the impossible. For another, the rule only works when economic fluctuations are small, and so would be of dubious value in a 2008-style crisis or a '70s-style inflation. 

So legislating a Taylor Rule for the Fed is a terrible idea. It’s just a tight-money policy with a light dusting of math. It may have received enthusiasm and support from the end-the-Fed crowd and their sympathizers in the Republican Party, but it isn’t good economics.

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.

To contact the author on this story:
Noah Smith at nsmith150@bloomberg.net

To contact the editor on this story:
James Greiff at jgreiff@bloomberg.net