MONETARY POLICY IN A NEW ERA

and David Wessel for comments and Michael Ng for excellent research assistance.

1

In 2017, the flagship research conferences of the European Central Bank and the Federal

Reserve—held in Sintra, Portugal, and Jackson Hole, Wyoming, respectively—had something in

common: Both had official themes unrelated to monetary policy, or even central banking. The

ECB conference (theme: Investment and Growth in Advanced Economies) did include an

opening speech by President Mario Draghi on monetary policy and the outlook, before turning to

issues like the prospective effects of technological advances on employment. However, the

economic stability in which the pressing economic issues will relate to growth, globalization, and

distribution—issues that are the responsibility of other policymakers and not primarily the

province of central bankers.

Would that it were so simple. Although central bankers can certainly hope that the next

ten years will be less dramatic and demanding than the past ten, there will certainly be important

new challenges to be met. In this note, I focus selectively on two such challenges: the

implications of the secular decline in nominal interest rates for the tools and framework of

monetary policy; and the status of central banks within the government, in particular, the

questions of whether central banks should and will retain their current independence in making

monetary policy. As I will explain, the two challenges are related, in that the low-inflation, low-

investment (Bernanke, 2005); and increased demand for “safe” assets (Del Negro et al., 2017;

Caballero et al., 2017). Some of these factors may reverse in the medium term—for example,

2

recent historically low rates of productivity growth could revert to more-normal levels (Byrne

and Sichel, 2017), and there is some evidence that the global savings glut may be moderating

(Chinn, 2017)—which could lead to somewhat higher rates in the future. For now, though, the

combination of low nominal rates and the difficulty of reducing short-term interest rates (much)

below zero implies that monetary policymakers may have limited scope to address deep

economic slowdowns through the traditional means of cutting short-term interest rates. Recent

research by Kiley and Roberts (2017) illustrates the potential severity of the problem. Based on

fiscal policies, there are two broad possibilities (interrelated and not mutually exclusive). First,

rather than relying on the management of short-term interest rates alone, as assumed by Kiley

and Roberts, monetary policymakers could make greater use of new tools developed in recent

years. In the first main section of this paper, I review some of these tools. I argue that both

forward guidance and quantitative easing are potentially effective supplements to conventional

rate cuts, and that concerns about adverse side effects (particularly in the case of quantitative

easing) are overstated. These two tools can thus serve to ease the ZLB constraint in the future, as

argued by Yellen (2016). Two other tools—negative interest rates and yield curve control—are

less likely to play important roles, at least in the United States. European and Japanese

policymakers have successfully employed negative rates, but overall they appear to have

1

rates, however. Since I am focusing on the Fed here, for simplicity I’ll stick with the ZLB acronym.

3

will discuss, the depth and liquidity of the markets for U.S. government securities would make it

difficult for the Fed to peg rates beyond a horizon of two years or so.

Although unconventional tools can increase the potency of monetary policy, the ZLB is

still likely to be a binding constraint on the monetary response to a downturn that is more

serious, or which occurs when rates remain (like today) below neutral levels. A second broad

response to the problem is to modify the overall policy framework, with the goal of enhancing

monetary policymakers’ ability to deal with such situations (Williams, 2017). Focusing on the

performs better on both counts, as 1) it is fully consistent with the goal of price stability, perhaps

even more so than an inflation target; and 2) it implies a “lower-for-longer” response to periods

when rates are at their ZLB, which approximates what theory tells us is the optimal approach.

However, a price-level target can be problematic in the face of supply shocks, and the switch to a

price-level target from the current inflation targeting approach would be a significant

communications challenge. In the latter part of the section, I propose and discuss a third possible

alternative: a “temporary price-level target” that kicks in only during periods in which rates are

constrained by the ZLB. I argue that the adoption of a temporary price-level target would be

likely to improve economic performance, relative to the current framework. Importantly, it

would do that while both maintaining price stability and requiring only a relatively modest shift

environment, linking this issue to the themes of the first part of the paper. In the United States,

the doctrine of CBI emerged, in part, from the inflationary experience of the 1960s and 1970s,

4

which was blamed in part on undue political influence on monetary policymakers. Following

these events, both formal models and informal conventional wisdom held that, to avoid pressures

to overheat the economy and allow higher inflation, the Fed needed greater independence from

politics. However, the inflation-centric rationale for CBI looks a bit outdated in a world in which

inflation and nominal interest rates are too low, rather than too high; and in which politicians

have criticized central banks for being too expansionary rather than not expansionary enough.

Indeed, the same logic that holds that CBI is necessary to avoid excess inflation can be turned on

the limits of independence appropriately depend on the sphere of activity under consideration

and on economic conditions. The general principle of CBI thus does not preclude coordination

of central bank policies with other parts of the government in certain situations.

DEFEATING THE ZLB: UNCONVENTIONAL POLICY TOOLS

Central bankers in 2008 faced extraordinarily difficult challenges, in particular the

combination of a deep recession—which made a sharp easing of monetary conditions

necessary—and the proximity of short-term interest rates to zero, which made easing difficult.

In response, monetary policymakers employed a number of unconventional policy measures.

Which ones will become part of the standard toolbox? In what order or combination might the

discuss, sequentially, forward guidance, quantitative easing, negative rates, and yield curve

control (the management of longer-term yields).

5

The non-standard tool on which central bankers are most likely to rely in the next easing

cycle is forward guidance, or communication about the expected or intended future path of the

policy rate. The Fed used variants of forward guidance in the Greenspan era, for example, in

references to keeping rates low for “a considerable period” (Federal Open Market Committee,

2003). Even earlier, a number of central banks experimented with forward-looking policy

financial conditions today, even if the short-term policy rate is close to its effective lower bound

(Guthrie and Wright, 2000).

Forward guidance comes in a number of forms. A useful distinction is between Delphic

and Odyssean forward guidance (Campbell et al., 2012). Delphic guidance is a simple statement

of how monetary policymakers see the economy and interest rates as likely to evolve. Delphic

guidance is advisory only and makes no promises about future policy. In contrast, Odyssean

guidance—the phrase is motivated by Odysseus’s decision to tie himself to the mast to be able to

resist the calls of the Sirens—is intended to pre-commit the central bank to some (possibly

contingent) set of future policies.

The goals of Delphic and Odyssean guidance are different. Delphic guidance—as for

increased transparency by central banks, rather than the emergence of the ZLB as an important

6

policy constraint. By improving the clarity of the central bank communication, Delphic

guidance is intended to increase the predictability of monetary policy and make it more effective.

Odyssean guidance, in contrast, is most likely to be relevant when the policy rate is at or

close to the ZLB, so that the scope for short-term rate cuts is limited. Typically, monetary

policymakers use Odyssean guidance to communicate a promise to keep rates lower for longer

than implied by their “normal” reaction function. If the promise is credible, then market

participants should bid down longer-term yields and bid up asset prices today, effectively adding

their institutions, as well as the tendency of market participants to look for focal points around

which expectations can coalesce, appear in practice to provide monetary policymakers some

ability to commit to future policy actions.

The Federal Open Market Committee (FOMC), the Fed’s policymaking body, provided

regular forward guidance during the recovery from the crisis. Some controversy has arisen about

the FOMC’s approach. Michael Woodford (2009) and others have argued that the FOMC

inappropriately used Delphic rather than Odyssean formulations in its guidance, limiting its

benefit. For example, at the same meeting at which the policy rate was cut effectively to zero,

the December 2008 FOMC statement indicated, “…the Committee anticipates that weak

economic conditions are likely to warrant exceptionally low levels of the federal funds rate for

as possible. A real-world complication is that policy committees are not typically unitary actors,

but may include participants of diverse views, trying to reach compromise in an uncertain

7

environment. Some hedging or ambiguity in the committee’s official statements may therefore

be difficult to avoid. In practice, however, the FOMC’s guidance after the crisis—as mediated

by the public comments of policymakers—did seem to have Odyssean effects. Notably, the

Fed’s introduction of forward guidance was typically followed by changes in longer-term interest

rates, exchange rates, and equity prices consistent with substantial increases in monetary

accommodation (Femia et al., 2013; Swanson, 2017) and by reduced sensitivity of near-term rate

expectations to economic news (Williams, 2014). The increases in equity prices in particular

on the fly, is promising for the future use of verbal interventions. As both central bankers and

market participants gain experience with forward guidance, the tool should become increasingly

effective.

Another important distinction is between qualitative guidance (“considerable period”)

and quantitative guidance, for example, describing specific economic conditions that would lead

to a change in policy. Over the years, Fed guidance has evolved from qualitative towards

quantitative, reflecting the desire to enhance transparency as well as the imperative of adding

substantial accommodation during the ZLB period. Economic logic suggests that quantitative

guidance will be more effective, because it is both more precise and more verifiable ex post (and

thus easier to support by reputational concerns). However, again, a policy committee may not

The FOMC used both types after the crisis, indicating first that it expected to hold rates low

through a certain date, then tying rate increases to thresholds based on the prevailing

8

unemployment and inflation rates.

economy, and so state-dependent guidance should be the default in the future (Feroli et al.,

2016).

As pointed out by Williams (2016) however, date-based guidance may at times be more

effective, perhaps because it is more definitive and more credible to market participants. A

particular situation in which date-based guidance might be desirable arises when policymakers

and the market have different economic outlooks. Suppose the ideal guidance would hold rates

at zero at least until unemployment fell to 6 percent; but suppose also that market participants

I’ve been discussing forward guidance about rates, but guidance can be provided about

aspects of policy other than rates, notably, about plans for asset purchases. Such guidance is a

natural extension of rate guidance and can be Delphic or Odyssean in intent. The main point

here is that guidance about the components of policy needs to be carefully coordinated, so that

the planned sequencing of policy changes is clear. For example, the famous 2013 “taper

tantrum” followed Fed guidance that it anticipated reducing the pace of its asset purchases,

conditional on economic developments. However, the tantrum reflected not so much the

expectation of reduced asset purchases per se, but rather the inference in some quarters of the

market (as could be seen in futures quotes) that increases in short-term rates would quickly

follow the slowing of asset purchases. (See below for more on the “signaling” aspects of

2

and November 2009. In August 2011, January 2012 and September 2012, the FOMC used different versions of

calendar-based forward guidance in which they set a date in which they would keep rates ‘exceptionally low’. In

December 2012 the FOMC switched to a state-dependent form of forward guidance in which they committed to

keeping interest rates ‘exceptionally low’ at least as long as the unemployment rate was above 6.5%, inflation was

below 2.5% based on one to two year ahead forecasts, and inflation expectations remained anchored.

3

discuss this point further below.

15

holdings would ultimately be reflected in reduced seigniorage payments to the treasury. Central

banks naturally see this outcome as a political risk to their independence and institutional

reputations which, all else equal, may make them more hesitant to use QE. However, political

risk to the central bank is not equivalent to a loss in social welfare. From the perspective of

society as a whole, the fiscal risks of QE have to be balanced against the substantial benefits of a

tool that gives monetary policymakers additional scope to respond to a serious economic

downturn or to unwanted disinflation.

losses do occur as the result of a QE program, it will likely be because the economy recovered

more quickly and strongly than expected, resulting in higher interest rates; since losses are most

likely to occur at times when the economy is unexpectedly strong, they are hedged, from a social

perspective. Finally, and importantly, the beneficial fiscal effects of an effective QE program go

well beyond seigniorage, as the government’s budget also benefits from low borrowing rates,

avoidance of deflation or very low inflation, and the higher revenues that result from increased

economic activity.

All that said, the fiscal argument seems to me to be more balanced than some of the other

criticisms of QE. There are difficult governance issues and competing values in play here, and

people could reasonably come to different conclusions. One approach, similar to that taken by

I don’t expect to be used by the Fed in the foreseeable future: negative (nominal) interest rates

and pegging longer-term interest rates (so-called yield curve control).

16

Negative interest rates have been recently employed in Japan and a number of European

countries (Bernanke, 2016a). To enforce negative rates, central banks generally charge a fee on

the reserve holdings of commercial banks. Arbitrage ensures that the negative return to reserves

translates into negative returns to other short-term liquid assets. Negative short rates need not

imply negative rates on longer-term assets, particularly those that are less liquid or involve credit

risk. Rather, negative short-term rates give the central bank a new tool for bringing down the

longer-term rates, like mortgage rates, that matter most for economic activity. The evidence

resistance and on the margin has probably made central bankers more hesitant to use this tool.

Putting aside the politics, and excluding limitations on the use of currency (Rogoff, 2016) as

beyond the scope of this note, negative interest rates appear to provide relatively modest benefits

and have modest costs. So while the tool may well be appropriate and useful in some contexts, it

does not merit the overheated public attention it has received.

Under current institutional arrangements, the potential benefit of negative rates are

relatively modest because attempts to push rates too far below zero will induce substitution into

cash. The most negative rate yet imposed is minus 75 basis points, by Denmark (Danmarks,

2015). To date, negative rates have so far not triggered much movement into cash, as best as we

can tell, but it is likely that more such adjustment would occur if rates were to go much further

not employ negative rates during the post-crisis period (Burke et al., 2010). Reforms undertaken

since the crisis have reduced this risk, by forcing many money market funds that invest in private

17

assets to shift to a system of floating net asset values (which allows for a negative nominal

return) and by inducing a shift toward lower-risk government funds (SEC, 2014; Chen et al.,

2017).

A more frequently heard concern is that negative rates could decapitalize the banking

system, because banks are supposedly unable to pass negative rates on to depositors and thus

would have to absorb the loss. There is little evidence from the European or Japanese

experiences that modestly negative rates have actually hurt bank profits or bank lending. Retail

undertaking such changes is worthwhile, given that political resistance to negative rates appears

disproportionate to their generally modest benefits, is an open question. At the Fed, there was

little support for negative rates during the post-crisis period, a situation that does not appear to

have changed.

Yield curve control, recently introduced by the Bank of Japan, is the targeting of yields

on longer-term bonds; e.g., the BOJ is currently targeting the yield of ten-year government bonds

at around zero. Yield curve control is “dual” to conventional QE: Instead of setting targets for

securities purchases and letting the market determine yields, as in ordinary QE, under yield curve

control the central bank targets the yield on one or more securities and adjusts its purchases as

necessary to hit the targets (Bernanke, 2002; Chaurushiya & Kuttner, 2003; Bernanke, 2016b).

10

of the stack in terms of net effectiveness” (Mui, 2016). However, interestingly, Chair Yellen has indicated that she

believes that the Fed has the legal authority to impose negative interest rates should it choose to do so (United

States, 2016).

18

control can also be an efficient strategy when the securities available for purchase are potentially

limited in quantity and supplied with less-than-perfect elasticity, as is the case with the Japanese

government bonds that make up the bulk of the BOJ’s purchase program. In particular, the

adoption of yield curve control by the BOJ has allowed the Japanese authorities to maintain

substantial stimulus, even as the supply of bonds available for purchase by the BOJ has shrunk

(Bernanke 2016d).

On the other hand, in jurisdictions with deep and liquid securities markets, like those for

limiting the utility of the program (Bowman et al., 2010; Bernanke 2016b). An intriguing

possibility, however, is that a relatively short-horizon peg could be used to complement forward

guidance about future short rates.

What policy tools will be used, and in what sequence, when the next recession hits?

Yellen (2016) has described the Fed’s prospective toolbox. In the face of an economic

slowdown, the FOMC would respond first with conventional rate cuts. In an environment of

super-abundant bank reserves, in practice that would involve reductions in the Fed’s key

administered rates, including the interest rate paid to banks on excess reserves, the rate offered to

11

a date two years from the announcement. This strategy would be difficult to implement for state-dependent (as

opposed to time-dependent) guidance, however.

19

Forward guidance, of the Odyssean variety, would come next (substantial Delphic

guidance is already in place). Relative to earlier experience, I would expect a much earlier

adoption of state-contingent, quantitative commitments to hold rates low.

What about QE? Fed policymakers have been clear that QE is now part of the toolkit and

that it would be used if necessary (United States, 2017; FOMC, 2017). I am sure that’s true, but

I would not be surprised if there is a period of hesitation before the FOMC starts up new rounds

of asset purchases. The effects of QE on financial markets and the economy are less well

Speaking positively rather than normatively, I don’t see much likelihood that negative

rates or yield curve control will be employed in the United States in the foreseeable future,

unless circumstances become dire. Given current institutional arrangements, negative rates have

only modest benefits and may create problems for some financial institutions. Targeting longer-

term interest rates, at least at maturities out beyond a couple of years, could be a hazardous

undertaking given the deep and liquid markets for U.S. government obligations. It is not

inconceivable though that the Fed might consider targeting yields at somewhat shorter horizons,

particularly as a way of reinforcing its forward guidance on rates.

DEFEATING THE ZLB: THE POLICY FRAMEWORK

limit the flexibility and timeliness of fiscal tools; that’s why monetary policy has normally been

the first line of defense against short-term economic instability. But recent experience suggests

20

that fiscal policy can provide some backstop in the most severe slowdowns, as in the United

States in 2008 and 2009 (Matthews, 2011; Auerbach and Gorodnichenko, 2017).

All that said, I am less sanguine than Yellen (2016) that the current monetary toolbox

would prove sufficient to address a sharp downturn. In particular, there is no guarantee that the

next recessionary shock will occur only when policy rates are at or above neutral levels. If the

Fed had to react to a new slowdown today, it would have only 100 basis points or so of room to

cut short rates, and other major central banks (such as the ECB and BOJ, whose short rates are

options, I’ll suggest a compromise approach.

As noted, one proposal for modifying the policy framework is to keep the current

inflation targeting framework of the Fed and other major central banks, but to raise the level of

the target—from 2 percent or so to 3 or even 4 percent. Presumably, after a period of transition,

an increase in the inflation target would result in a comparable increase in nominal interest rates.

Higher nominal rates would in turn expand the scope for short-term rate cuts and reduce the

salience of the zero lower bound at all points of the yield curve.

Most recently, a group of

economists signed a letter to the Fed arguing for a higher inflation target, and the aforementioned

Fed Up group held a seminar at this year’s Jackson Hole conference endorsing the idea (Baker et

al., 2017; Leubsdorf, 2017).

12

practical issue with this approach is that measurements of nominal GDP are not as timely as those of inflation and

unemployment, and more subject to revision.

13

increase to 4 percent provides more room for future rate cuts, but it might be difficult to defend 4 percent inflation as

being consistent with central bank mandates for price stability. An increase to 3 percent adds only modest scope for

rate cuts; if that is the increase contemplated, it would be desirable to compare the costs and benefits of that increase

with the adoption of negative rates, which add similar amounts of policy space (Bernanke 2016c).

21

policy frameworks. These are important benefits. At the same time, I see some problems with

this proposal:

First, proponents may be underestimating the costs, uncertainties, and delays associated

with the transition to a higher target. We have seen, most recently in Japan, that managing

inflation expectations through central bank announcements can be tricky. Given that inflation

expectations in advanced economies seem well anchored at two percent or below, trying to raise

expectations and to re-anchor them at a higher level could well be a protracted and uncertain

hysteresis channels, for example—but it might also prove to be destabilizing and difficult to

manage, particularly if inflation expectations became volatile.

Second, to be fully effective in raising longer-term nominal yields, the increase in the

inflation target must be perceived as permanent and irrevocable. However, that perception

would be undermined by the apparent willingness of the Fed to raise its target for what might

appear to be tactical reasons. Looking forward, it is likely that the determinants of the “optimal”

inflation target—such as the prevailing equilibrium real interest rate, the costs of inflation, and

aspects of the monetary transmission mechanism—will change over time. If the Fed raised its

inflation target today based primarily on the low level of real interest rates, would it change the

target again in response to future changes in fundamentals? That would be important to clarify

public perceives costs of inflation—the greater difficulty of planning and calculation when

inflation is high, for example—that economic models don’t well capture. In any case, it’s not a

22

coincidence that the promotion of “price stability” is a key part of the mandate of the Fed and

most other central banks. Certainly, a substantial increase in targeted inflation would invite a

backlash, perhaps even a legal challenge. Proponents have suggested convening a national

commission to approve the increase in the inflation target, to increase its legitimacy and

durability. Those proponents should be careful what they wish for. In the United States, rather

than validating a higher inflation target to afford scope for discretionary monetary policy, I

suspect that the political process would be more likely instead to reaffirm the centrality of “price

preferred response is for the central bank to promise to follow a “make-up” policy (or, in

Woodford’s term, for policy to be “history-dependent”). Specifically, suppose the ZLB binds for

a period, keeping monetary policy tighter than it otherwise would have been. Then (speaking

very loosely) the optimal policy involves the central bank promising to keep rates lower for

longer than it otherwise would have, where the length of the “make-up” period increases with the

severity of the episode and the cumulative shortfall in monetary ease. If the public understands

and believes this promise, then the expectation of easier policy and more rapid growth in the

future should act to mitigate declines in output and inflation during the period in which the ZLB

is binding. Note, by the way, the close analogy to Odyssean forward guidance, discussed earlier.

The difference is that, rather than being implemented by ad hoc guidance, the optimal policy is

target does not optimally calibrate the vigor of the policy response to a given ZLB episode to the

duration or severity of the episode.

23

A somewhat better option than raising the inflation target is to adopt a price level target,

an approach advocated by a number of economists and policymakers (Svensson, 1999; Gaspar et

al., 2007; Williams, 2017). Effectively, a price-level targeting central bank tries to keep the long-

run average inflation rate close to a targeted value, say 2 percent. The principal difference

between price-level targeting and conventional inflation targeting is the treatment of “bygones.”

An inflation-targeting central bank aims to keep inflation stationary around its target, an

approach that allows policymakers to “look through” a temporary change in the inflation rate, so

usually optimally depends) on the state of the real economy.

In this section I consider only

“flexible” variants of policy rules.

Switching to a price level target has at least two principal advantages over raising the

inflation target. The first is that price-level targeting is consistent with low average inflation

(say, 2 percent) over time and thus with the price stability mandate. Indeed, price-level targeting

arguably promotes price stability better than does inflation targeting, because its commitment to

stabilizing long-run average inflation should lead to considerably less uncertainty about the level

of prices far in the future. The second advantage is that price-level targeting has the desirable

“make-up” feature of the theoretically optimal monetary policy. In particular, under price-level

targeting, periods of below-target inflation (as is likely to happen when interest rates are stuck at

14

and employment, does not perform well.

24

switching from the inflation concept to the price level concept might require considerable

education and explanation by policymakers. How quickly, for example, would markets and the

public adjust to the implication of price-level targeting that a burst of inflation today should lead

them rationally to expect lower-than-normal inflation in the future?

Another possible concern about price-level targeting is that the “bygones are not

bygones” aspect of this approach is a two-edged sword. Under price-level targeting, the central

bank cannot “look through” supply shocks that temporarily drive up inflation, but must commit

credible; if not, efforts to offset positive inflation shocks would likely be costly.

Is there a compromise approach? One possibility, that I will describe briefly here, is to

apply a price level target and the associated “make-up” principle only to periods around ZLB

episodes, retaining the inflation-targeting framework and the current 2 percent target at other

times. As I will explain, the central bank can explain this combined policy in familiar inflation-

targeting terminology, which I take to be an advantage.

So, to be concrete, at some moment when the economy is away from the ZLB, suppose

the Fed were to make an announcement like the following:

(1) The FOMC has determined that it will retain its inflation-targeting framework, with a

symmetric inflation target of 2 percent. The FOMC will continue to pursue its

situations in which the funds rate is at or near zero, a necessary condition for raising

the funds rate will be that average inflation since the date at which the funds rate first

31

explicitly guarded by statute or treaty, as is the case in some other jurisdictions, including Japan,

the United Kingdom, and the euro area.

A standard description of the terms of central bank independence might end here; but I

think we should recognize that, at least in the medium term, democratically elected officials must

be able to respond to institutional deficiencies of or poor execution by the central bank.

reasons).

The most dramatic reform of the Fed, in 1935, generally increased the independence

and effectiveness of its monetary function; the Dodd-Frank reforms of 2010 focused on lending

and regulatory authorities and did not materially change the Fed’s monetary role. Likewise, it is

not inconsistent with CBI that the reappointment of a policymaker depends on his or her success

at meeting mandated goals, so long as the promise of reappointment is not used to influence or

politicize particular policy decisions. This latter distinction is admittedly not always sharp, but it

is supported by the tradition of appointing central bank governors based on professional

qualifications rather than political experience and by cultures of professionalism and

nonpartisanship at central banks, including the Fed of course.

This description of CBI captures, in general terms, the current institutional and political

15

good times but acts to reform the institution after episodes of poor economic or financial performance.

32

On the first question, the most prominent argument of the research literature is that CBI

helps to solve time consistency problems of monetary policy (Kydland and Prescott, 1977; Barro

and Gordon, 1980). The basic story, with some simplification, posits that promises of low

inflation may be difficult to make credible if decisionmakers are too focused on the short term.

In particular, if the public expects low inflation, then policymakers may be tempted to achieve a

short-run increase in output by engineering a near-term inflation “surprise.” But the public will

come to expect such surprises, with the result in equilibrium being that inflation and inflation

the Fed (Rogoff, 1985), and, indeed, it’s very plausible that central bank independence has

contributed to the global reduction in inflation of the past couple of decades. Essentially, in this

view, an independent central bank is a commitment device by which politicians control their own

inflationary biases.

The general thesis that independent central bankers have longer horizons, and are thus

better able than short-term-oriented politicians to take into account the longer-term interests of

the economy, rings true, as I’ll discuss. However, the specific argument that CBI reduces

inflation bias is looking shaky these days. In recent years, the problem faced by central banks

has been too little, rather than too much, inflation. Moreover, some of the strongest political

opposition to the strong monetary measures of the Fed and other major central banks have come

diverse causes, including the political power of creditors (savers and financial institutions), but in

any case it is difficult to reconcile with the traditional time consistency story.

33

One might be tempted to conclude that, in a low-inflation world, CBI is no longer

desirable, and that central banks should coordinate directly with governments when needed to get

higher inflation (Eggertsson, 2013). I’ll return to the issue of central bank-government

coordination below. But the point I want to make here is that the time consistency argument for

CBI was never the complete rationale, and that there are other good reasons to delegate monetary

policy to an independent central bank:

First, monetary policymaking can be highly technical. Congress hires the Fed to manage

with financial markets. Legislatures are not equipped to do this, and even ex post interference

with central bank decisions or communication will create uncertainty in markets and reduce the

effectiveness of monetary policy.

If monetary policy’s technical aspects, need for timeliness, and market sensitivity were

the only reasons for delegating this responsibility, then it might be sufficient to have monetary

policy run out of the finance ministry or Treasury, rather than by an independent central bank.

Indeed, there have been significant examples of monetary policy oversight by developed-

economy finance ministries, as recently as the 1990s in the United Kingdom and Japan for

example. An argument for delegating monetary policy to the Treasury is precisely that it might

the weight of experience. First, monetary policy operates with substantial lags, and campaigns

of policy easing and tightening can play out over a number of years. So, even if inflation bias is

34

itself not a problem, it is important that monetary policymakers keep a longer-term perspective

and ensure policy continuity and coherence over time. Finance ministers (who in most

jurisdictions are elected politicians) may turn over quickly—as Fed chair I worked with four

Treasury secretaries (two Republicans and two Democrats), and Alan Greenspan worked with

seven—and changes in government shorten time horizons as well. For similar reasons,

independent central banks are better placed to develop long-run institutional credibility—to

anchor inflation expectations, establish a reputation for carrying through on forward guidance, to

above as an argument for keeping monetary policy in the Treasury, might cut the other way in a

short-run political context: Would there be concerns, for example, that the Treasury is setting

interest rates to influence the size of the government’s current deficit?

In sum, the avoidance of inflation bias is far from the only rationale for delegating

monetary policy to an independent central bank. Other reasons include the technical, time-

sensitive, and market-sensitive aspects of monetary policy-making; the benefits of the longer-

term perspective on the economy that apolitical central bankers bring; the development over time

of institutional credibility and predictability; and the avoidance of incentives to react to short-

term political developments. The delegation of monetary policy to central banks may also be in

the interest of politicians, as well as in the public interest, which helps explain why it’s

35

These arguments for CBI pertain primarily to monetary policy. What about other central

bank activities? Should they be carried out with the same presumption of independence as

monetary policy?

CBI is an instrumental principle, not an ideological or philosophical proposition: It

makes sense when it leads to better policy outcomes (on average) and not otherwise. Thus, the

applicability of CBI may vary according to policy function or to economic circumstances. In its

supervisory functions, for example, the central bank should not make special claims to

function rapidly and effectively, making use of supervisory and other confidential information.

However, in a financial crisis, coordination of the central bank with the executive (including

other agencies) and the legislature may be essential for achieving stability. In the 2007-09 crisis,

the Fed and other central banks worked closely with finance ministers, without compromising

their independence in monetary policy or as lenders of last resort.

A special case of interest arises when an economy faces significant deflation risk, and

when monetary policy is unable to contain the problem on its own. As Eggertsson (2013) has

pointed out, that situation turns the traditional “inflation bias” argument for independence on its

head: If the purpose of CBI is to avoid inflation bias in normal times, then it would seem to

follow that CBI should deliberately be sacrificed when more rather than less inflation is desired.

necessarily preclude coordination of monetary policymakers with the fiscal authorities, if two

conditions are satisfied:

36

First, the goals of the coordination should be both consistent with the central bank’s

mandate and not achievable in the absence of coordination. So, if fiscal-monetary coordination

is essential for meeting the central bank’s inflation goal or to preserve financial stability, then it’s

better to coordinate than to fail to meet the mandated objectives. In contrast, the central bank

should not coordinate if the result is inconsistent with or outside of its mandates, or if its

mandated objectives can be achieved without coordination. In particular, explicit fiscal-

monetary coordination to raise inflation should not be considered except as a last resort, when

As a positive matter, will CBI survive? There are certainly worrying signs. In the United

States, the Fed remains unpopular in many quarters, and hostility between some members of

Congress and Fed leadership has been frequently evident. The Fed also has fewer institutional

protections than most other advanced-economy central banks; in particular, as noted, its policy

independence is not explicitly protected by law or treaty (except indirectly, through the speed

bumps mentioned earlier).

However, there are also reasons to be guardedly optimistic: First, recent economic

performance—at least the dimensions of it amenable to monetary policy—has been good, with

the Fed near its inflation and unemployment objectives. At the same time, post-crisis reforms

have addressed many aspects of the crisis response, including the most unpopular interventions.

financial crisis and the ensuing recession were fought with financial and monetary policies that,

while ultimately successful, engendered political backlash. After the crisis, necessary reforms

37

helped to strengthen the system and to rationalize crisis-fighting authorities. The risk is that

politicians may be tempted to constrain or eliminate some of the powers, including monetary

powers, that proved essential in the past decade. Such changes might have little immediate

effect, but unless carefully thought through they could expose the American and world

economies to severe financial and economic risks in the future.

frameworks may prove insufficient to offset a severe economic slowdown. That leads to the

question of whether alternative monetary frameworks could increase the potency of monetary

policy. One leading suggestion, to keep the current inflation-targeting regime but to raise the

inflation target to 3 or 4 percent, raises a number of concerns. Targeting the price level is a

better approach, but it also has drawbacks, including the communications challenge of making

such a large change in the policy framework and the requirement that central banks reverse

(rather than “look through”) temporary inflation shocks. In this paper I have proposed for

consideration a “temporary price-level targeting” approach, which applies only at times at which

policy rates are at or very close to zero; at other times, standard inflation targeting would prevail.

Under this approach, monetary policymakers would commit in advance not to raise rates from

inflation. I agree with Eggertsson (2013) that fiscal-monetary coordination, rather than

independent monetary policy, may at times be necessary to fight persistent deflation or low

38

inflation. However, I argue that the case for CBI goes well beyond the avoidance of inflation

bias and is based on factors that include technical competence, institutional credibility, and the

need for a longer-term perspective. I am hopeful that CBI will survive in the United States and

other advanced economies, but memories are short and thus vigilance in defense of CBI will be

important.

APPENDIX: THE RELATIONSHIP BETWEEN THE TEMPORARY PRICE LEVEL

However, unlike KR, who assume that the shadow rate rule applies at all times, my approach

assumes that a conventional policy rule is used away from the ZLB.

To be concrete, suppose that the ordinary policy rule followed by the central bank (away

from the ZLB) is a Taylor rule with a lagged interest rate term:



  . I include the lagged term both for realism and because lagged responses tend to improve policy performance

when the ZLB is relevant. Empirical evaluations of this approach should therefore allow for the lagged term.

39

An alternative means of dealing with the ZLB constraint, explored by KR, begins by

specifying a shadow policy rate, which is non-zero (negative) only when the actual policy rate is

constrained by the ZLB. Following KR, I define the shadow rate by:

17

letting the shadow rate rule apply during ZLB episodes and setting policy by the inertial Taylor rule, equation (1), at

other times.

40

not directly observable by the public, and its value depends on specific choices of the form and

target. After the exit, in this approach, the policy rate is determined by the standard Taylor rule.

47

the-stimulus-work-a-review-of-the-nine-best-studies-on-the-

subject/2011/08/16/gIQAThbibJ_blog.html?utm_term=.5a6527454988.

Mui, Ylan Q., “Why this top official thinks the Federal Reserve should raise interest rates this

federal-reserve-should-raise-interest-rates-this-year/?utm_term=.33fb6184fc76.

England Staff Working Paper No. 571, December 2015.

Reifschneider, David and John C. Williams, “Three Lessons for Monetary Policy in a Low-

Rogoff, Kenneth S., “The Optimal Degree of Commitment to an Intermediate Monetary Target”

Rogoff, Kenneth S., “The Curse of Cash,” Princeton University Press, Princeton, NJ, September

Summers, Lawrence H., “Have We Entered an Age of Secular Stagnation?” IMF Fourteenth

Annual Research Conference in Honor of Stanley Fischer, IMF Economic Review, Vol. 63,

Issue 1, Washington DC, May 2015, 277-280, http://dx.doi.org/10.1057/imfer.2015.6.

Svensson, Lars E. O., “Price-Level Targeting versus Inflation Targeting: A Free Lunch?,”

Swanson, Eric T., “Measuring the Effects of Federal Reserve Forward Guidance and Asset

Taylor, John B., “Discretion versus Policy Rules in Practice,” Carnegie-Rochester Conference

Series on Public Policy, 39, North-Holland, 195-214, 1993.

Taylor, John B. and Paul D. Ryan, “Refocus the Fed on Price Stability Instead of Bailing Out

Fiscal Policy,” Investor’s Business Daily Op-ed, December 1

, 2010,

http://media.hoover.org/sites/default/files/documents/Refocus-Fed-on-Price-Stability-Instead-of-

Tobin, James, “How Dead is Keynes?,” Cowles Foundation for Research in Economics at Yale

United States. Senate Committee on Banking, Housing, and Urban Affairs, “Federal Reserve’s

th

Washington: Federal Reserve Board, 2016 (statement of Janet L. Yellen, Chair, Federal Board).

48

United States. Cong. House Committee on Financial Services and Senate Committee on

Banking, Housing, and Urban Affairs, “Semiannual Monetary Policy Report to Congress” July

12

and 13

th

, 115th Cong. 1st sess. Washington: Federal Reserve Board, 2017 (statement

of Janet L. Yellen, Chair, Federal Board).

Rules,” Washington DC, July 23

rd

15

2010, https://blogs.wsj.com/economics/2010/11/15/open-letter-to-ben-bernanke/.

Wall Street Journal, “Full Text: Republicans’ Letter to Bernanke Questioning More Fed Action,”

th

https://blogs.wsj.com/economics/2011/09/20/full-text-republicans-letter-to-bernanke-

questioning-more-fed-action/.

Werning, Ivan, “Managing a Liquidity Trap: Monetary and Fiscal Policy”, NBER Working

Paper No. 17344, Cambridge, MA, August 2011.

Williams, John C., “Monetary Policy at the Zero Lower Bound: Putting Theory into Practice,”

Hutchins Center on Fiscal and Monetary Policy at the Brookings Institution, Hutchins Center

Working Paper #3, Washington DC, January 16

, 2014.

Zero Lower Bound,” Presentation at the US. Monetary Policy Forum, New York, New York,

th

Williams, John C., “Preparing for the Next Storm: Reassessing Frameworks & Strategies in a

th

Woodford, Michael, “Comment on ‘Heeding Daedalus: Optimal Inflation and the Zero Lower

Bound’,” Brookings Papers on Economic Activity, Washington DC, Spring, 2009, 38–45.

Yellen, Janet L., “Opening Remarks: The Federal Reserve’s Monetary Policy Toolkit: Past,

Present and Future,” Remarks at the Jackson Hole Economic Symposium, Jackson Hole, WY,