A Better Model Won’t Fix The Fed

Can the Federal Reserve build a better mousetrap? Wall Street Journal columnist Joseph Sternberg hopes so. Sternberg rightly indicts the Fed for failing to predict and contain the worst inflation in 40 years. The problem is “the central bank’s model of the economy — the set of spreadsheets, as it were, that are supposed to help Fed staff economists and leaders understand developments on Main Street.” If the Fed had a better model, the argument goes, it would make better policy.

While no model is perfect, we should acknowledge some models are better than others. Nevertheless, I worry Sternberg has conceded the fundamental issue — whether better theory and more accurate data can help solve the problem of discretionary macroeconomic management. In reality, the problem is insoluble. As Peter Boettke, Daniel Smith, and I argue in Money and the Rule of Law, the Fed won’t stabilize the economy so long as it relies on discretion instead of rules.

Sternberg lists the flaws of the Fed’s model: it doesn’t do a good job of accounting for fiscal policy, it overstates the effects of government spending on productivity improvements, and relies on naive Philips curve mechanisms to capture the effects of labor market slack on inflation. These are, indeed, problems. But suppose the Fed fixed these problems. Would policy predictably improve?

Perhaps. But I’m skeptical. We shouldn’t assume that Fed officials will adopt and follow the policy recommended by a better model. 

Discretionary monetary policy permits a central bank to act as it sees fit, even if such actions don’t make sense in light of the available forecasts. Discretion also affects the models central banks pick. As Andrew Haldane, who previously served as chief economist at the Bank of England, recently observed, “many policymakers produced their economic forecasts by working backwards from their preferred stance.” A better model wouldn’t help us if its results are inconvenient for Fed officials, since they can invoke their discretion to choose a different model.

The better model Sternberg seeks would not only have to be adopted and followed. It would also have to remain the better model. Discretionary monetary policy makes that less likely, too. 

Expectations matter a great deal for monetary policy. Ideally, the central bank will set nominal (current-dollar-valued) expectations and then deliver on them. Discretion makes it harder to set nominal expectations because the public knows the central bank may choose not to follow through. Consequently, the parameters of the world economists attempt to model are less stable under discretion than it would be were the central bank to commit to an effective monetary rule. A model that takes those parameters as given may perform well for a short while. But, when expectations change, the given parameters of the model will no longer reflect reality. Then, Fed policy will become either too hot or too cold, even though the previously better model is telling monetary policymakers it’s just right. Fed officials will look like overconfident generals who discover, too late, that they can’t fight this war like the last one. Millions of Americans suffer because of it.

The Fed doesn’t need a better model. The Fed needs a rule. Congress must force the Fed to hit a specific target variable and establish an accountability mechanism if it fails. Politicians shouldn’t control day-to-day monetary policy. But they should, and constitutionally they must, instruct monetary policymakers on their appropriate goals. There are several good options. One is an inflation target — force the Fed to stabilize the dollar’s purchasing power. Another is a nominal income (current-dollar GDP) growth target — force the Fed to stabilize aggregate demand. Both would improve monetary policy by creating a nominal anchor and preventing central bankers’ discretionary money mischief.

Ultimately, monetary policy requires changing the money supply only to the extent necessary to offset changes in money demand. Otherwise, there will be an excess supply of or demand for money at the prevailing price level, and market prices will fail to deliver accurate signals about relative resource scarcities. We should structure our monetary institutions to systematically eliminate these monetary disequilibria to ensure money is a helpful facilitator of exchange.

Instead of predictably maintaining monetary neutrality, the Fed is trying to convince markets about the future level of its policy instrument. This is totally counterproductive. An instrument, such as the fed funds rate, should be at whatever level brings market expectations in line with the rule. There’s no need to make markets believe that the future fed funds rate will be at a particular level. Markets themselves will undertake the requisite adjustments so long as the rule is credible. The whole point of expectations-responsive monetary policy is to remove the discretionary and technocratic elements from central banking. Disappointingly but unsurprisingly, the Fed is doing the opposite: doubling down on discretion and technocracy.

Sternberg is correct that the Fed confronts a credibility crisis of its own making. The Fed’s reputation “now depends on Mr. Powell’s success in suppressing that inflation.” So long as Powell’s Fed continues picking the wrong models, the wrong tools, and the wrong goals, it deserves to lose its lofty status. The solution is not merely to introduce better models, better tools, and better goals. It is to eliminate the Fed’s ability to choose between the available models, tools, and goals. Discretion won’t cut it. We need a rule-bound central bank.