Data Analysis Ahead of Theory

Almost everybody agrees that we need to rethink macro. Of course, there is very little agreement about what that rethinking will involve. Predictably, there is a strong desire on part of the "mainstream" to cordon off the debate, to assert that it is all in there in the mainstream papers, we just need to rediscover them. I propose a different approach. A moratorium on macro theory, at least for now.

A little over 30 years ago, Ed Prescott, proclaimed that business cycle theory was ahead of measurement. Although he said it because the data kept rejecting his pet theories, he was probably right about macro data, which was uninformative. Yet, today we have reams of micro and disaggregated data that are being used to inform the macro debate. Mian and Sufi of House of Debt fame are pioneers in this regard and many others have since followed suit. Not just micro data, researchers are able to use macro and macrofinancial data more effectively. It is theory that is still in the dark ages. So, let us focus on more data analysis and try to build a consensus on a new set of stylized facts. Then perhaps we can think of building theories.

In that spirit, I am going to compile a list of stylized facts based on empirical work of the past several years.

1. Demand has Persistent Effects

Since Samuelson's neoclassical synthesis, mainstream economics has generally operated on the principle that demand matters in the short-term but not in the long term. Theories have been updated to DSGE compliant NK models, but the principle of demand being a short-term phenomenon has remained constant. Post Keynesians have always disagreed with the MIT Keynesians in this respect. Yet recent empirical work almost unambiguously shows that recessions have persistent effects on unemployment and even on the long-term trajectory of output.

2. Fiscal Policy is Effective in Recessions

Until the Great Recession, the standard macropolicy consensus was that fiscal policy had no role in macro demand management, that monetary policy alone could steer the economy to nirvana. Most believed in what is called monetary policy dominance--that is monetary authorities would take into account fiscal actions and offset them. Outside of a few unreconstructed market monetarists, nobody now believes that because the data is simply overwhelming. Several studies show that a) fiscal policy is quite effective in recessions, and not just recessions at the zero bound, b) monetary policy does not come close to offsetting fiscal policy, c) fiscal consolidations have persistent negative effects on output whereas fiscal easing in recessions actually improves debt.

3. Private Debt Matters, Public Debt Not So Much

Private sector debt in mainstream theory was seen as an epiphenomenon. Macroeconomics has generally worked under the assumption that the financial dimension of the economy has no bearing on the working of the real economy. Balance sheets merely reflect the real economy. Buildup in private debt is of no consequence because defaults would just transfer wealth from creditors to debtors. Bernanke, Krugman, and others in their dismissal of Minsky's validity for understanding debt deflations have argued that private debt could not possibly matter on a scale large enough to cause depressions. On the other hand, fear of public debt as some sort of a burden was well-entrenched. Recent empirical work have turned both of these findings on their head.

Macrofinancial studies have consistently shown that a large buildup in household debt is often a precursor of financial crises and poor future economic performance. Meanwhile, Mian and Sufi have documented the channels through which debt and leverage work their way.

On the other hand, the dubious work of Rogoff and Reinhart combined with the fact that actual economic and financial markets outcomes have belied the public debt worry-worts has called into question the conventional wisdom about sovereign debt and default.

4. How Does Monetary Policy Work?

The general consensus seems to be that money is neutral in the long run and that monetary policy works by lowering the cost of capital (or expectations). The great inflation of the 1970s further reinforced the then emerging consensus that the job of the central bank is to deliver low and stable inflation and that everything else would take care of itself. Yet, the apparent failure of central banks worldwide to hit their inflation targets and the supposed failure of the Phillips Curve is causing some soul-searching. Meanwhile, empirical works suggest that inflation targeting does not work the way it is assumed to work, that the costs of inflation vastly exaggerated, and that the cost-of-capital channel of monetary policy may we weak.

There is increasing evidence that the expectations formation process is complex and informational rigidities abound. Not surprisingly, inflation targeting does not appear to anchor expectations in the ways widely assumed in New Keynesian models. Nakamura and Steinsson, in an excellent paper, find that the costs of inflation assumed by the standard policy models are elusive in practice. They also find evidence of non-neutrality in high-frequency data. Meanwhile, survey data suggest that business capital spending plans are not sensitive to the cost of capital—another presumed benefit of low and stable inflation. Moreover, a slew of papers finds that corporate capital spending does not respond to Q in ways predicted by conventional theory.

Concluding Thoughts

We may well be reemerging from the dark age of macroeconomics—a phrase memorably coined by Paul Krugman in 2009. The Peterson Institute for International Economics recently held a conference “Rethinking Macroeconomic Policy” that had an agenda questioning almost all aspects of the received wisdom of recent decades. The IMF too has been reexamining the consensus views and increasingly arguing for expanding the policy arsenal. Macroeconomists are discovering—rather rediscovering—that monetary policy alone cannot deliver stability and that fiscal policy and macroprudential policies have important roles to play. In treating finance as an epiphenomenon and money as a mere veil, Dark Age macro contributed to the Great Recession and its aftermath. Hopefully, we can now put that chapter behind us.

However, it is imperative that we keep the data horse in front of the theory cart. In this regard, I would like to point out the difference between ancient Greeks and ancient Indians in their approach to astronomy and astronomical measurements.

Needless to say, the Indians were far more accurate. In fact, their accuracy was not matched in the West until the 18th century.


Popular posts from this blog

Government Investment--Key to America's Success

Socializing Risk is Bedrock of Capitalism