Today’s Free Exchange post is entitled “Scott Sumner is Wrong.” I certainly can’t disagree with that claim, but I’m not sure I’m wrong in quite the way that M.C.K. asserts.
LAST Thursday, Jeremy Stein, a governor of the Federal Reserve Board, gave an important speech outlining the ways that monetary policy can inflate—and prevent—deeply destructive debt bubbles. (You can read my summary of his main points here. The speech was not about current policy so much as how the Fed should behave in general.) Scott Sumner, a blogger, was unimpressed by Mr Stein’s analysis, arguing that it was uninformed by history. However, the latest empirical studies support Mr Stein’s thesis that monetary policymakers who care about the long-term well-being of the citizenry should monitor private credit creation and prevent it from growing too rapidly.
. . .
Banks and other financial intermediaries usually create credit whenever they can earn what they believe is a risk-adjusted spread between their funding costs and the rates they charge their borrowers, both of which are affected, if not determined, directly by the monetary authority. Tobias Adrian and Hyun Song Shin have shown that the balance sheets of financial firms that mark their assets to market grow and shrink based on changes to the level of short-term interest rates. Meanwhile, Markus Brunnermeier and Yuliy Sannikov have shown that monetary policymakers can alter the willingness of banks to create credit by adjusting the shape of the yield curve. (I wrote a more detailed summary of this new research here.)
I read M.C.K. as assuming that short term rates are a good indicator of the stance of monetary policy. More specifically, that easy money leads to low short term rates, which leads to extra leverage. I don’t agree. For simplicity, let’s use Bernanke’s benchmarks for the stance of monetary policy; NGDP growth and inflation. More specifically, let’s assume the Fed pegs the price of a futures contract linked to a weighted average of NGDP growth and inflation. So we have a monetary policy that, by construction, is always neutral in Bernankian terms. It’s never easy and it’s never tight. But short term interest rates would still move around quite a bit. And the leverage of financial firms might well be highly correlated with the movements in short term rates. So it would look like monetary policy is having a big impact on leverage, whereas in reality (by assumption) it would be having no impact at all.
If you don’t like my assumption, change it as you please. Assume the Fed is targeting M2 growth at 4%/year, as Friedman once advocated. You get the same result, neutral money and volatile short term rates.
The bottom line is that short term interest rates are a lousy indicator of the stance of monetary policy, even though (paradoxically) on the day of a FOMC meeting a higher than expected setting of the fed funds target is almost always an easier than expected monetary policy. But over any longer period of time (when a central bank is targeting inflation or NGDP growth), short term rates will reflect conditions in global credit markets. For instance, we know that the 1% interest rates of 2003 did not represent “very easy money” because it did not lead to particularly high expectations of inflation and/or future NGDP. If low short term rates lead to a credit bubble (under NGDP targeting) the credit bubble is not caused by the Fed, it’s caused by an inflow of Asian savings, or some other (non-monetary) credit market factor.
The yield curve might be a slightly better indicator of the stance of money policy, but unfortunately it’s also an excellent indicator of changes in expected NGDP growth. A sharp slowdown in expected NGDP growth can even cause the yield curve to “invert.” If that leads to deleveraging, it might well be due to fear of recession, not the direct effect of changes in short term rates.
Mr Sumner says that central banks would do better taming the credit cycle solely with regulatory tools, although he does not specify how this would work in practice. Moreover, he asserts that monetary policy is too “blunt” to be helpful. But Mr Stein explained that monetary policy can be a useful supplement to regulatory measures precisely because those can only be applied to areas of the financial system that are being actively monitored by regulators. Unlike regulation and supervision, monetary policy “gets in all the cracks” because all financial intermediaries are exposed to the interest rates under the central bank’s control.
My first best solution (admittedly not realistic today) is to get rid of all government intervention in credit markets. No Fannie and Freddie, no FHA, no deducting interest on mortgage loans, no FDIC, etc. Laissez-faire.
My second best solution is to try to regulate to make our system look more like Canada’s. Unlike 99.99% of bloggers I think small banks are the problem and big banks (like Canada) are the solution. I also favor bans on making sub-prime loans with taxpayer-insured funds. Require a minimum of 20% down, unless the lender is not insured by FDIC. Also change laws on non-recourse loans, etc. Change tax laws so that debt is not subsidized (as compared to equity.)
I agree that monetary policy “gets in all the cracks” but I disagree with the implications he draws:
Mr Sumner might argue that monetary policy “gets in all the cracks” because it affects the level of nominal output. But the cutting-edge research makes it clear that financial firms operate according to a unique set of incentives different from those of the broader economy. Monetary policy affects those incentives, which I described above, more directly than it affects the incentives of the nonfinancial sector. Regular people and firms in the “real” sector simply do not care about small changes to the level of short-term interest rates to the same extent as commercial banks, investment banks, insurers, the repo market, and all of the other intermediaries responsible for creating money and credit.
File this under “never reason from a price change.” If the change in short term rates reflects credit market conditions (which is usually the case) then I agree that most people don’t care. But if it reflects a change in monetary policy then it really does get in all the cracks—Main Street is affected as much or more than Wall Street. The tight money of late 2008 (not picked up by the short term rate indicator, BTW) was devastating to Main Street.
Thus, the evidence suggests that the Fed and other central banks can in fact use monetary policy tools to restrain credit growth without crushing the economy—if they want to.*
That’s not how I read the evidence from 1929.
My bigger problem with this entire line of analysis is that it’s all (implicitly) based on a giant misconception, that the severe recession was caused by financial turmoil, not tight money. If the Fed had kept NGDP plugging along at a 5% rate we would have had some very mediocre years—call it stagflation if you wish. Maybe 1% RGDP and 4% inflation until the excesses were worked off. But the high unemployment and greatly intensified debt crisis need not have happened.
My view is obviously the minority view, held by neither policymakers nor my fellow academics. But it does follow from what we’ve been teaching our students in recent decades; that the Fed has both the ability and the duty to keep expected NGDP growth plugging along at a decent rate. Until we correctly diagnose the real problem, we will not be able to come up with solutions. Unfortunately, the newly resurgent credit view reflects a basic misconception about what went wrong.
But I’ll give M.C.K. and Stein a lot of credit in one respect. If my view is wrong and the now standard credit view of 2008 is correct, then there might be a role for monetary policy in this area. The people who should really be embarrassed here are the professors who continue to teach Sumnerian macroeconomics out of their Mishkin textbooks (Fed never out of ammo—low rates aren’t easy money), but don’t believe a word of what they teach.
PS. In their paper, Adrian and Shin call for monetary policy to be more forward looking than traditional interest rate targeting rules. I agree, and believe a policy of targeting NGDP one or two years out would do a lot to prevent the build up of credit excesses, or excessive deleveraging on the other side of the crisis.
PPS. Karl Smith asks two questions. Here are my answers:
1. Maybe; not in a mechanical sense, but perhaps in a signaling sense. Yes, but it’s probably ineffective.
2. Yes, as much as required to keep the central bank’s subjective forecast of NGDP growth on target.
PPPS. This debate reminds me a bit about the debate over utilitarianism. Some anti-utilitarians believe the world would be a happier place if we abandoned utilitarianism. Some NGDP opponents believe NGDP would grow at a more stable rate if central banks targeted something other than NGDP.
PPPPS. Noah Smith recently pointed to the fact that Japan grew at a decent rate from 2000-07 without anything close to 5% NGDP growth:
On the other hand, as we saw above, Japan remained in or very near deflation for the entire period, and ever since (meaning that NGDP didn’t grow anywhere near the 5% that Scott Sumner and others claim is optimal).
Just to be clear, I’ve never claimed Japan needs 5% NGDP growth to have healthy RGDP growth. Indeed I don’t even think that is true for the US. In the long run monetary policy doesn’t affect RGDP growth (very much), and in the short run what matters is the NGDP growth rate relative to expectations from a few years back. I do believe that very low trend NGDP growth rates will slightly reduce the level of RGDP, but that was “priced in” by 2000, and so would not have been expected to impact post-2000 RGDP growth rates in Japan.
Japan did have relatively poor RGDP growth after 1991, but the big problem in Japan since 2000 is excessively low levels of RGDP, relative to the US. Low productivity. For people of my generation, who recall the incredible dynamism of the Japanese economy in earlier decades, the big shock is that they have actually grown slightly slower than the US since the bubble burst in 1990. Not much slower, but I would have expected further convergence.
As an aside, I expect Zimbabwe and North Korea to grow faster than the US over the next 20 years, although I’m not at all positively inclined toward their economic policies.
HT: Travis V.