Dear Mr. Krugman,
Since last October I have been worried that nominal GDP growth would fall far short of the level consistent with full employment. Last fall I forcefully presented this argument to a number of economists (and was fortunate that Greg Mankiw and Robert Barro were willing to spend more than an hour listening to my views.) I suggested that despite the near-zero interest rates, an unconventional monetary policy could still be highly effective.
Because you are currently the most influential progressive voice on economic issues, and because you are an expert on liquidity traps, and because you have been skeptical about the effectiveness of monetary policy in the current environment, I decided to write you in the hope that you will reconsider your views on monetary policy. Not reconsider your model of “expectations traps,” but rather consider whether things have gotten so bad that the risks of a highly unconventional monetary policy are now outweighed by the risks of not adopting such a policy.
Before getting into specifics, I should add that although I have a reputation as a “right wing economist,” I believe my that proposal is very much in the interests of those who favor the broader policy goals of President Obama. The American public is not as patient as the Japanese public. If we stagger through 4 years of Japanese-style deflation (or even zero inflation), it is very unlikely that Obama will be re-elected. And the American electorate today is also very different from the electorate in 1935 (when FDR was concerned with Huey Long), middle class voters with falling 401k balances would not turn to someone to the left of Obama.
I think that we all agree that faster nominal GDP growth would be desirable. You have argued that the stimulus plan is too small, both private forecasters and the various financial markets now seem pessimistic, and even the Fed expects nominal GDP growth to fall well short of their target for the next several years (and they’ve been notably too optimistic throughout this crisis.) So the only question now is: Can monetary policy be effective in an environment with zero interest rates and a damaged financial system? For several different reasons, I believe it can.
1. Historical examples: As you know FDR was able to turn deflation into substantial inflation almost immediately after taking office, through his policy of leaving the gold standard and sharply devaluing the dollar. I would be the first to admit that the specific policy of devaluation is inappropriate in the current environment, as the rest of the world also faces a severe demand shortfall. But this example shows that rapid inflation can be achieved through unconventional monetary policies in an environment with near zero interest rates and a severely damaged banking system.
2. Easy Reforms: Eventually I will get to quantitative easing, but there are some even easier steps that could make the problem much more manageable, without incurring the risks of highly unconventional policies. One easy step would be to stop paying interest on reserves. These interest payments increase the demand for reserves, and are thus deflationary (as were the reserve requirement increases of 1936-37.) Of course this would make T-bill yields immediately fall to zero, and banks would still probably hoard substantial amounts of reserves. But then why not go one step further and charge an interest penalty on excess reserves? That would end the current problem of banks treating reserves and T-bills as near perfect substitutes. Yes, it wouldn’t solve that problem with respect to cash held by the public, but so far most of the hoarding of base money has been done by banks. (This is probably because, unlike during the early 1930s, deposits are now FDIC insured.) I don’t know if the interest penalty idea would work, but the Fed should certainly consider it.
3. Quantitative Easing: This is actually not my ideal solution, as I’ve published many papers advocating Nominal GDP (or CPI) futures targeting. But I also think it is a mistake to adopt an untried scheme in the midst of a crisis. Quantitative easing (although somewhat risky in budgetary terms) seems less uncertain, as it is merely an extreme version of the open market operations that are normally used to control the base. I understand the expectations trap argument at a theoretical level, but in another post I argued that this problem may not limit the Fed’s options as much as one might imagine. The post is here, but the basic idea is that the two famous liquidity traps (the U.S. in the 1930s and Japan more recently) don’t fit the current situation. The Fed is not constrained by a gold price peg (as in the 1930s) and the Fed does seem to have a sincere desire for roughly 2-3% inflation (unlike the BOJ, which raised rates in both 2000 and 2006, despite continual declines in their GDP deflator.)
I think you have acknowledged that there is some level of quantitative easing that would boost demand. If I am not mistaken you are concerned that if such a policy boosted inflation expectations sharply, the Fed would have to quickly sell off these assets, suffering massive capital losses. I understand that argument, but for two reasons I don’t think quantitative easing would be as difficult as many imagine. First, as James Hamilton pointed out, we could begin with Treasury inflation-indexed bonds which might not depreciate if the Fed succeeded in inflating the U.S. price level. I wouldn’t even rule out having the Fed consider buying some riskier U.S. assets (or foreign government bonds), which might actually appreciate if the Fed action succeeded in boosting aggregate demand.
4. Set an explicit NGDP (or CPI) target and engage in “level” targeting: The other reason why I am not so concerned about the possible losses from quantitative easing is that I think that such a policy (especially if combined with my earlier proposal to reduce excess reserves) would not require as much monetary base expansion as one might envision. It is very misleading to look at the huge increase in excess reserves that has occurred in an environment without a credible anti-deflationary policy (and with interest being paid on bank reserves) and extrapolate to what would be required to actually boost AD. Indeed a credible policy along the lines I propose might actually require the Fed to immediately reduce the now bloated base. One key to making the policy credible (as many have already argued) is to set an explicit nominal target, and commit to make up for any shortfall this year with even faster nominal growth in the future (and vice versa.) I know that your expectations trap argument raises questions about credibility. But explicit targets tend to be more credible because it is embarrassing for policymakers to go back on their word–they don’t like to lose credibility (for good reasons.) And Bernanke, et al, already have reputations very different from the members of the BOJ.
I would also point to hints from the financial markets that a bold move might be highly welcome. The strong stock market response to the only slightly more expansionary than expected rate cut in December 2008 (by which time it was already clear to you that we were in a liquidity trap) suggests to me that markets might respond very positively to an announcement similar to the array of steps proposed here. (Multifaceted policy initiatives are more likely to be welcomed by markets, as we don’t know exactly which specific step works best.) I understand that some might argue that the stock market was grasping for straws last December, but as my post here on the earlier December 2007 contractionary policy surprise shows, stock and bond markets often show a very sophisticated understanding of the impact of monetary policy.
5. What do we have to lose?: We can get rid of interest on bank reserves (and consider a penalty rate), set an explicit nominal target, and engage in quite substantial quantitative easing using indexed bonds (and perhaps a few foreign government bonds) without incurring much risk at all. And even if we have to eventually move more heavily into assets more exposed to U.S. inflation risk (long term T-bonds) I don’t see how those risks are any worse that what we are now doing at the Fed. Isn’t a risky policy that has a good chance to boost AD superior to a risky policy that has little chance of achieving that goal?
6. Confusing causality: Lots of people tell me that we need to fix the banking system first. But isn’t that reversing causality? Yes, the original sub-prime crisis was caused by bad decisions by banks (among other factors), but isn’t the current deterioration in higher quality mortgages, commercial loans, industrial loans, etc, mostly due to the precipitous drop in nominal GDP that began late last summer? Even if we end up with some capital losses from unconventional monetary policy, isn’t it also possible that monetary stimulus could vastly reduce the cost of bailing out the banks? And also consider the impact of faster nominal GDP growth on the budget deficit. So yes, there are some potential capital losses from unconventional monetary policy, but given what we are now going through those losses don’t seem quite so scary anymore. FDR showed that boldness can be surprisingly effective–I read somewhere that his housing bailout programs in 1933 ended up costing much less than expected because of his effective steps to boost nominal GDP growth.
Of course there is some risk of overshooting toward high inflation, but I believe those risks are minimal. The Fed can closely monitor yield spreads for signs of a change in inflation expectations. Admittedly (as Bernanke and Woodford pointed out in a 1997 paper on the circularity problem in targeting market expectations) such monitoring does not provide useful information about the proper stance of monetary policy when it is 100% credible–but we are currently far from that situation.
To conclude, I ask you to reconsider your position on monetary policy. If you did change your view, some people might accuse you of inconsistency. But remember what your hero once said:
“When the facts change, I change my mind — what
do you do, sir?”
The Obama administration is obviously struggling in coming up with an effective solution to the banking crisis. The stimulus package seems inadequate, either because (as you believe) it is too small, or (as I believe) the multiplier may be less than we think. The economic data seems to be consistently worse than expected. The facts have changed.