Jan Hatzius: Some Lessons from the Past Four Years
The following is based on remarks at the University Club in New York at the ceremony for the 2011 Lawrence R. Klein Award for the most accurate forecast over the prior four years to the Goldman Sachs US Economics team. The award was sponsored by the W. P. Carey School of Business at Arizona State University and Blue Chip Economic Indicators, Inc., and was presented by Dr. Lawrence H. Summers, Charles W. Eliot Professor of Economics at Harvard University.
What are the lessons that we and the economic forecasting community have learned—or should have learned—over the past four years? There are undoubtedly many, because a depressing period for the economy is also an instructive period for economists. But I want to focus on three:
First, stability can be destabilizing. There was a big decline in economic volatility in the 1980s and 1990s. In response, consumers and firms sharply increased their leverage and started to run a large financial deficit, i.e. a large excess of spending over income and of investment over saving. This supported both the demand and supply side of the economy and thereby produced very favorable economic outcomes—higher profits, higher stock market valuations, low inflation, and a strong labor market. And the favorable economic outcomes time and time again “proved wrong” the naysayers who thought that the boom would end in a bust.
The process continued until the imbalances in the financial markets, in the banking system, and in the real economy had grown so large that the economy had become extremely vulnerable to even small negative shocks. The resulting downturn has proven deep and difficult to reverse. So in that sense, the initial greater stability of the economy has proven highly destabilizing in the longer term.
The phrase “stability is destabilizing” was coined by Hyman Minsky (1919-1996), a post-Keynesian economist whose work has only become known more widely in the wake of the crisis. But I hope you will forgive me for noting that the basic story was very well laid out by Bill Dudley and Ed McKelvey in a Goldman Sachs study called “The Dark Side of the Brave New Business Cycle” in 1999, when most people were unfamiliar with Minsky's work. In my view, it remains the most important insight into the origin of the crisis.
Second, fiscal stimulus is typically undersupplied after a crisis.
“The central irony of the crisis is that while it was caused by too much confidence, borrowing, and spending, it is only resolved by increases in confidence, borrowing, and spending.” (Lawrence Summers, 2011)
I believe that this is true. But unfortunately, it is probably too much of an irony to resonate with most voters and elected officials. Especially after a period in which some fiscal stimulus has already been applied and the economy is still in bad shape, it is all too tempting to believe that government deficits are part of the problem rather than part of the solution, and that macroeconomics is a morality tale where virtuous governments are rewarded and wicked ones are punished. That view is held more strongly in some countries and cultures than in others but it has been gaining ground everywhere, including the United States.
We are concerned about the potential impact of a more restrictive fiscal policy on an economy that is still very fragile. The cutbacks at the federal, state and local level have been subtracting close to 1 percentage point from GDP growth over the past year, and are likely to subtract the same or more in 2012 depending on whether the payroll tax cut and the emergency unemployment benefits are extended.
And the potential “cliff effect” gets even higher in 2013, when the economy faces potential fiscal retrenchment from the expiring Bush tax cuts, the expiration of whatever temporary measures do get extended in 2012, and the potential automatic spending cuts if the Congressional “super committee” does not come to an agreement. So we suspect that fiscal policy will continue to be “too tight” from a cyclical perspective for the foreseeable future.
Third, monetary stimulus is also typically undersupplied after a crisis.
“[U]nder a fiat money system, a government (in practice, the central bank in cooperation with other agencies) should always be able to generate increased nominal spending and inflation, even when the short-term nominal interest rate is at zero.” (Ben Bernanke, 2002).
Again, this is true in theory. But it has not proven to be a good guide in practice. Typically, central banks do not generate a sufficient increase in nominal spending when short-term rates are near zero. Just as fiscal stimulus is "undersupplied" relative to the amount that is necessary after a financial crisis, monetary stimulus is also undersupplied.
A very clear piece of evidence is the path of US nominal GDP before and after the crisis. Even three years after the crisis, nominal GDP is still 10% below the pre-crisis trend, with no signs of any catch-up. And most of this is due to the fact that real output and employment are far below their estimated potential. That is a very big problem, not just because high unemployment is very costly in the short run, but also because sustained high unemployment raises the risk that workers will lose their skills and attachment to the labor force and ultimately become unemployable.
So why has the Fed not eased monetary policy more? I think the main reason is that central banks have a bias toward caution in an unfamiliar environment. The best explanation for this bias is the so-called “Brainard principle,” named after William Brainard of Yale University. The Brainard principle says that if a central bank is uncertain about the impact of its instrument on the economy, it should do less than if it were sure. The reason is that the policy instrument itself is an added source of uncertainty in the economy, so using it is “costly.” Inevitably, central banks are much more uncertain about the impact of unconventional monetary policy than about the impact of conventional interest rate cuts. This means that monetary stimulus at the zero bound will be too small in expectation, relative to a situation in which the central bank was sure about its impact.
So how do we overcome this bias toward excessive caution? It’s not easy to overcome because the Brainard principle is a genuine reason for why some caution is warranted. But we think one option would be for the Federal Reserve to focus not only on the setting of the instrument, i.e. the funds rate or the size of the balance sheet, but more directly on the level of nominal demand in the economy--that is, the outcome of its policies. Specifically, we think that the Fed may want to target the level of nominal GDP extrapolated from the pre-crisis trend.
This would be a big step for the Fed, and we don’t think it’s happening anytime soon. But it would be quite consistent with the Fed’s dual mandate to pursue maximum employment and price stability, because nominal GDP is simply the price level multiplied by real GDP, and real GDP in turn is extremely closely related to the unemployment rate.
That said, nominal GDP targeting would be a different interpretation of this mandate than the usual "Taylor rule" interpretation. First, it puts more weight on the employment and output side of the mandate. This is justifiable in an environment where the Fed is missing so badly on the employment side and where high unemployment has potentially large long-term costs. Second, it aims at the price level, not the inflation rate. This has the advantage of providing greater stability to future price level expectations. Third, it is much simpler and more transparent than the inevitably somewhat fuzzier Taylor rule interpretation of the dual mandate.
If the Fed did decide to do something like this, we think they should couple the shift to a nominal GDP target with a promise to use the remaining monetary policy instruments—delay in hiking the funds rate and renewed asset purchases—aggressively to achieve the target over time. Ideally, Congress would also play its part by supplementing the shift via an easier fiscal policy in the near term coupled with a credible plan for restraint in the longer term.
By telling the public that the Fed will pursue easy policies until we are much closer to pre-crisis income levels, we also think that the Fed could increase the public’s confidence in economic recovery and reduce worries in the financial markets about a rapid return to tighter policies if growth does pick up. There is a lot of talk about the central role of “confidence” in economic recovery. The problem is that it is difficult to know what “confidence” means, how one can measure it, and what policymakers can do to improve it. But a nominal income target might help by stabilizing and coordinating expectations around economic recovery, so we think it deserves serious consideration.
Read related research: Jan Hatzius: The Case for a Nominal GDP Level Target