Skip to main content
On the Record

Comments on “The Liquidity Trap: What to do about it”

Editor’s note: Donald Kohn delivered the remarks below at the 18th Geneva Conference on the World Economy, titled “The Liquidity Trap: What to do about it,” held in Geneva from May 26–27.

This paper couldn’t be more timely or important. One of the narratives that has taken hold and contributed to roiling global markets in the early part of 2016 is that central banks were running out of room to use even unconventional policies to boost growth, dooming the global economy to a prolonged period of resource underutilization and inflation rates well below central bank targets.

This paper is a very helpful and comprehensive guide to monetary policy in a liquidity trap—at the zero lower bound for policy rates. It aggregates and evaluates the many studies of the effects of unconventional polices undertaken in the past several years, marshals new evidence on the effects of negative policy rates, and uses those results to make concrete suggestions for policy at the ZLB, which it tests in a number of counterfactual exercises

The authors arrive at several conclusions: 1. The zero lower bound and liquidity trap are far more likely to be issues in the future than in the past, given the decline on the equilibrium short-term interest rate; 2. Just about everything central banks tried in terms of low/negative interest rates and asset purchases/QE worked to ease financial conditions and by extension to bolster growth and prevent even worse outcomes, and the negative side effects if any have been small and far outweighed by the benefits of these policies; 3. So to escape the liquidity traps today and in the future, the advice to central banks is to just do more—lower more, buy more, buy over a wider array of assets; 4. And while they are at it, central banks should drive inflation rates to new higher targets of say, 4 percent to elevate nominal rates and make episodes at the ZLB less frequent.

My comment is of the character “yes, but”. I broadly agree with many of the conclusions reached by the authors, but I think there are nuances, costs, and complications they could examine more carefully and that central banks need to consider when making policy in a liquidity trap or raising their inflation targets.

I agree that central banks are likely to find themselves dealing with liquidity traps more in the future than in the past. Sluggish growth in potential GDP reflecting slower technological change, weak capital investment, and the demographics of aging populations has contributed to a prolonged decline in r*. But I would caution that future episodes in the liquidity trap may be less serious than one might infer from recent history and this different character could have implications for policies at the ZLB. That’s because of the very substantial beefing up of the regulation of the financial sector in the wake of the global financial crisis. Standards for bank capital, liquidity, and risk management have been raised quite substantially; plans for bank resolution that minimize risks to financial stability are being made; and some nonbank markets that contributed to financial instability in 2008 have been made more transparent and safer, say through the use of central counterparties for clearing.

This regulation probably has contributed to the decline in r* by making intermediation more expensive. But it also means that the amplification of real economy shocks by the financial sector is considerably less likely and when it occurs should be less serious. Well-capitalized and highly liquid banks will be much less subject to runs with forced fire sales of assets; interconnections in derivative markets will be more transparent and the risk easier to manage; credit should keep flowing to the real economy following an adverse development. Of course business and financial cycles will persist fueled by waves of greed and fear and miscalculations by private and public sectors, but the resulting recessions are more likely to be of the garden variety type, with implications for time spent in a liquidity trap and perhaps for the appropriate type of response.

I agree that unconventional monetary policies have been effective at easing financial conditions and boosting the economy. The paper concentrates a lot on the policy of quantitative easing—large-scale asset purchases in Fed jargon. The authors identify three channels through which such purchases should work. First, a market-calming channel in which central bank purchases help to restore market liquidity when trading conditions are disrupted. Second, purchases can reinforce the signals that the central bank is sending about its intention to keep interest rates at unusually low levels for unusually long periods. Third is the portfolio balance channel in which central bank purchase of longer-term or risker assets reduces term or risk premiums directly and which is transmitted more widely as the previous holders of these assets rebalance their portfolios.

But, I wonder whether the paper’s conclusions about the effectiveness of QE derived from recent experience will be applicable to future episodes of liquidity traps and if they are not, what that says about the efficacy and ordering of unconventional policy tools. Without a doubt, the Fed’s first QE in a dysfunctional MBS market was highly effective. But, as I noted, financial disruption is likely to be much milder in future episodes, reducing any impact from the market-calming channel.

With regard to signaling future policy intentions, central banks have developed various means of forward guidance for policy interest rates over this episode of policy at the ZLB. Guidance has become more sophisticated and more economy based, as, in my view, it should be. The Fed and other central banks have seen forward guidance as a key element and a separate tool from purchases in unconventional policy, and it will come into play in a more developed state in the next episode, reducing the need for and impact of the signaling channel for QE.

Trimming expected future short-term rates is a powerful tool for convincing households and businesses to bring future spending forward to the present, as it emphasizes that yields on holding liquid assets into the future are going to present a poor alternative use of funds to spending today. The paper barely touches on the experience of central banks with forward guidance and its efficacy in promoting spending; greater attention to this tool would enhance the utility of this paper for researchers and central banks.

That portfolio balance channel will continue to have effects on term and risk premiums—depending on the type and duration of the assets purchased. The question is just how powerful this will be in stimulating spending if purchases have no effect on market functioning and little effect on expected short-term rates, given the use also of forward guidance. QE should remain effective to some extent; for example It will increase incentives to barrow and it will raise asset prices, activating a wealth channel. But with expected rates little affected, the impact of the portfolio balance channel by itself on bringing spending forward could be muted. And that might affect the most effective ordering and mix of unconventional policies in a liquidity trap.

Asset purchases generally also involve the central bank taking some fiscal risk. This is true even when the assets purchased are longer-term government bonds without credit risk, as the Fed has been doing. The expected profits from holding long-term government obligations financed with short-term debt (bank reserves) will be positive because of the term premiums that were prevalent when the central bank undertook its purchases. But the returns on carry trades can vary considerably—that’s why there are term premiums under normal circumstances. And of course purchases of corporate bonds or private mortgage securities or equities entail much more risk as well as decisions about government intervention credit allocation that are normally made by the fiscal authorities.

So my second “yes but” for unconventional policies is to consider the governance and accountability issues of independent central banks engaging in QE. Governments and central banks have found ways to deal with this during the crisis for some of the facilities. For example, the Fed’s TALF facility that lent against private securitizations came with a backstop from the Treasury’s TARP program; and the Bank of England’s corporate bond purchases were made pursuant to a public letter form the Chancellor concerning compensation for any loses. Discomfort at the Federal Reserve with some of these issues was reflected in an “accord” with the Treasury department in the spring of 2009, acknowledging, among other things the temporary allocative implications of intervening in government backed MBS markets. Questions have also been raised about the debt management implications of the Federal Reserve’s purchase of long-term Treasury securities—how should this coordinated in the future with the debt managers at the Treasury department.

In sum, QE raises serious issues of governance and accountability for independent central banks that should at least be acknowledged and discussed if public and legislative support for central bank independence in the conduct of monetary policy is to be maintained. I suspect that one reason some central banks were slower to adopt QE types of policies was their concern about provoking questions about authorities and independence; better these issues be confronted openly in dialogue with legislators.

I agree that a 4 percent inflation target that was achieved and credibly committed to would have substantial benefits because it would materially lower the odds on future liquidity traps and encounters with the ZLB, given the higher average level of nominal interest rates implied by a higher target. Of course there are costs to weigh against these benefits. The report argues convincingly that some of the costs often cited can be reduced over time. One such cost is the greater variability of higher levels of inflation. Another is the concern that raising the inflation target will undermine the credibility of any such target—raising it once will arouse suspicions that it can be changed and make it much harder to build credibility for the new target. Both of these can be overcome by central bank actions—though it may take some time and entail transition costs advocates of higher targets rarely consider. For example building credibility for a 4 percent target is likely to require leaning extra hard against any tendencies for inflation to overshoot this objective for a while. This kind of asymmetrical reaction function implies that embedding 4 is likely to require inflation averaging below 4 for some time.

But more broadly, higher inflation could well have costs that go beyond these last two or the shoe leather, menu, and tax distortion costs the report discusses. Alan Greenspan defined price stability as “best thought of as an environment in which inflation is so low and stable over time that it does not materially enter into the decisions of households and firms.” That definition has resonated with many over the years. It reflects, in my view, recognition that inflation has costs beyond the technical issues already discussed. It complicates decision making and distorts market signals. And higher inflation probably create difficulties for households and businesses with limited financial expertise in particular, and is likely to have disproportionate effects on small businesses who can’t hire that expertise and less-educated households.

This sense of broader costs could be one reason ”price stability” is so deeply entrenched in legislation and treaties that establish central bank mandates. Four percent is not price stability in that the price level would double every 18 years. The benefits could well exceed the costs in a world of r* close to zero. But at a minimum, good governance and accountability would suggest consultation with the legislature o r other authorizing authority so the costs and benefits could be fully aired by elected officials.


Finally I agree that unconventional policies at the zero lower bound have been effective, but what lessons should we draw from the experience of Japan? Japan seems to be a distinct natural experiment in the all-in use of unconventional policies to escape a liquidity trap. Under its current governor, the BoJ has used all the tools advocated by the authors of the paper and used them aggressively, but it has not had the kind of success one might have anticipated from reading this paper. It raised its inflation target (to 2 percent from 1); it has purchased large quantities of both government and private longer-term obligations, including equities in the form of ETFs; it has reduced its target rate into negative territory. Many financial market measures responded, at least initially. Core inflation rose for a while, albeit partly reflecting the rise in the level of import prices after the yen depreciated, and inflation expectations also increased. But progress stopped well short of the two percent target and recent signs are that the higher inflation has not become entrenched in wages and that core is slipping back—though it remains higher than before the program began. Perhaps the BoJ should just do more of everything, as the report implies, but it is troubling that more sustainable success at hitting a higher target has not been achieved; the Japanese experience is worthy of extra study for its implications for what works to overcome a liquidity trap.


Get daily updates from Brookings