Lawrence H. Summers, the Charles W. Eliot university professor at Harvard, is a former treasury secretary and director of the National Economic Council in the White House. He is writing occasional posts, to be featured on Wonkblog, about issues of national and international economics and policymaking.
My friends Mike Spence and Kevin Warsh, writing in the Wall Street Journal on Wednesday, have produced what seems to me the single most confused analysis of U.S. monetary policy that I have read this year. Unless I am missing something -- which is certainly possible -- they make a variety of assertions that are usually exposed as fallacy in introductory economics classes. (Brad DeLong has expressed related views).
My problem is not with their policy conclusion, though I do not share their highly negative view of quantitative easing (QE). There are many harshly critical analyses of QE, such as those of Martin Feldstein, which are entirely coherent and consistent with the macroeconomics of the last 50 years. My differences are based on judgements about empirical magnitudes and relative risks -- not questions of basic logic.
Spence and Warsh do not focus on the inflation risks, financial stability risks or distributional risks from overly expansionary monetary policy. Instead, they assert a proposition that I have not encountered in 40 years as a professional economist: that overly easy monetary policy reduces business investment. Indeed, they blame the weakness of business investment during the current recovery on the Fed.
This line of argument is to say the least surprising.
Every major macroeconomics textbook -- whether Keynesian monetarist or classical in orientation -- teaches students that investment increases as real interest rates decline. This is motivated in a variety of quite compelling ways. It is noted that lower rates raise the present value of the returns from investment and so make them more attractive. It is observed that lower rates mean lower borrowing costs or lower costs of drawing-down liquid asset holdings, making the purchase of capital goods more attractive. For these reasons, millions of students have been taught that Hicks famous IS curve slopes downward. Hundreds of empirical studies have found that investment responds to capital cost as theory predicts, though the magnitude is open to debate.
What arguments do Spence and Warsh offer for their heterodox conclusion? They note rightly that monetary policy has been easy and investment has been weak in the current recovery. This is a little like discovering a positive correlation between oncologists and cancer and asserting that this proves oncologists cause cancer. One would expect in a weak recovery that investment would be weak and monetary policy easy. Correlation does not prove causation.
They then argue that low rates somehow promote corporate stock buybacks, and this is an alternative to real new investment. I confess that I cannot follow the logic here. I would have supposed that the choice between real investments and share repurchases would depend on their relative price. If, as Spence and Warsh assert, QE has raised stock prices, this should tilt the balance toward real investment.
Likewise, I would have thought that by making the return to holding cash less attractive, easy money would tend to drive firms into making real investments. And, for the many firms that do not have huge hoards, it would bring down borrowing costs.
Perhaps Spence and Warsh are on to something that I am missing. I'm curious whether they can point to any peer reviewed economic research, or indeed any statistical work, that backs up their views. I am certainly open to any new evidence or new argument after all that has happened in recent years that easy money reduces business investment. And there is plenty of room for debate over policy.
For now, though, I would put the Spence-Warsh doctrine that easy money reduces investment in a class of propositions backed by neither logic nor evidence.