The Fed: From forward guidance to data dependence
The goal of every central banker is to stabilize the economic and financial system—keeping inflation low, employment high, and the financial system operating smoothly. Success means reacting to unexpected events—changes in financial conditions, business and consumer sentiment, and the like—to limit systematic risk in the economy as a whole. But as they do this, policymakers try their best to respond predictably to news about the economy. That is, there is a central bankers’ version of the Hippocratic Oath: be sure you do not become a source of instability.
This is the context in which monetary policymakers have created elaborate communications frameworks, speaking frequently and publishing profusely. They emphasize their policy guidelines so that households and businesses can count on their responses to new developments. Years ago, in an effort to maximize their discretionary authority, central bankers were zealously secretive. Today, they know that policy is more effective when households and businesses can anticipate it. That means constraining their discretion.
So, why is the Fed highlighting that policy is “data dependent”—responding to new information—while still providing forward guidance, signaling that the tightening is likely to be gradual? Are these two concepts—data dependence and forward guidance—mutually consistent? Or, are Fed officials merely adding to confusion and uncertainty in violation of their fundamental oath?
To answer these questions, we need to step back and analyze the role that forward guidance has come to play over the past decade. This tool takes different forms in different places. For example, a number of central banks publish a forecast of their own policy rate path. (The Swedish Riksbank, the Norwegian Norges Bank, and the Reserve Bank of New Zealand are the best known examples.)
In the United States, forward guidance is integrated into communication policy in the form of phrases in FOMC statements and comments made by the Fed Chair. Yet, as we described in last week’s post, since the onset of the crisis, the Federal Reserve’s policy toolkit has changed substantially; and, communications policy has evolved along with it. It has only been four years since the Federal Open Market Committee (FOMC) began releasing quarterly forecasts for the path of output, inflation and the interest rate; and the Chair started holding a quarterly press conference. Similarly, it was only at the start of 2012—99 years after the Fed’s founding—that U.S. central bankers adopted a quantitative inflation objective.
The mid-2000s provide an excellent pre-crisis example of U.S. forward guidance. A few short excerpts from the FOMC’s statements show what we mean. First, in mid-2003, Fed policymakers emphasized that they would not change policy any time soon:
“In these circumstances, the Committee believes that policy accommodation can be maintained for a considerable period.” FOMC Statement August 12, 2003.
Less than a year later, with conditions changed, policymakers were ready to start raising rates gradually:
“At this juncture, with inflation low and resource use slack, the Committee believes that policy accommodation can be removed at a pace that is likely to be measured.” FOMC Statement May 4, 2004.
The emphasis on a “measured” pace in this second FOMC statement (and in the statements of the next 11 FOMC meetings) was likely a consequence of the unsettling experience with the previous round of Fed tightening a decade earlier. In 1994, the rise in the target federal funds rate triggered a surge of long-term bond yields. By contrast, in the 2004-2006 episode, as the Fed boosted the funds rate from 1% to 5¼% in 17 consecutive 25-basis-points steps, the “measured pace” guidance helped keep long-term yields from soaring. In fact, the policy may have worked too well, leading then-Fed-Chairman Greenspan in 2005 to discuss the “conundrum” that yields had not risen much despite Fed tightening. (Over the period during which the funds rate rose 425 basis points, the-10 year Treasury yield went up by less than 100 basis points.)
Federal Funds Target Rate, weekly, 1990-September 2008
Source: FRED (DFEDTAR).
Note: When the FOMC shifted from a target rate level to a range in 2008, this series was discontinued.
During and after the financial crisis of 2007-09, the FOMC used forward guidance—together with balance sheet policies—not merely to contain, but to depress long-term interest rates in an effort to provide an expansionary impulse when it might otherwise not have been possible. Here is a typical example:
“The Committee also decided to keep the target range for the federal funds rate at 0 to 1/4 percent and currently anticipates that economic conditions … are likely to warrant exceptionally low levels for the federal funds rate at least through mid-2013.” FOMC Statement November 2, 2011.
(A recent paper by Fed economists lists the numerous instances of forward guidance from 2008 to 2013.)
Over the past year, however, as the economic recovery has continued, the tone of comments shifted from forward guidance to data dependence. Fed Chair Janet Yellen’s March statement is indicative:
“To conclude, let me emphasize that in determining when to initially increase its target range for the federal funds rate and how to adjust it thereafter, the Committee's decisions will be data dependent, reflecting evolving judgments concerning the implications of incoming information for the economic outlook.” Federal Reserve Board Chair Janet Yellen, March 27, 2015.
What should we make of this? How big is this change? There is a temptation to say that the move is trivial. Isn’t all policy data dependent to some degree? After all, as Keynes may (or may not) have said, "When the facts change, I change my mind. What do you do, sir?"
Our view is less cynical. We see a whole spectrum of combinations between pure forms of data dependence and forward guidance. What the Fed has done in recent months is move along this spectrum toward the former and away from the latter. While one part of the story is that policymakers no longer wish to lower long-term bond yields, we suspect that the Fed’s shift also reflects uncertainty about the appropriate future policy path.
In a stable environment, where policymakers can be relatively sure of their forecasts, forward guidance makes sense. But in an environment that is less predictable, statements about the likely future actions have much less value and can even be misleading.
To see what we mean, consider again the 2004 form of forward guidance. One way to read the statement we reproduced above is as a conditional promise of the following form: Unless something drastic happens that leads to a material change in our forecast, we will raise interest rates by 25 basis points per meeting for the foreseeable future. Now, consider what happens as forecast uncertainty increases. The bigger the uncertainty, the bigger the chance that at each decision date there will be a material change in the environment. There is a point where the forward guidance becomes meaningless. Put another way, when uncertainty is extreme, there is no statement about the future path of the interest rate that a policymaker is likely to make good on, so it is better to say nothing—letting policy be purely data dependent.
Is uncertainty unusually high today?
On an intuitive level, the answer would seem to be yes. As we have written in the past, we know relatively little about the current level of either the neutral real rate of interest or the potential rate of real growth (see here and here). More generally, following the worst financial crisis since the Great Depression, the economic recovery has been the slowest since World War II, and is today characterized by puzzling patterns like the mix of high corporate profitability and low business investment, and of relatively low unemployment and stagnant wages.
Interestingly, however, common (if imperfect) proxies for uncertainty do not exhibit an obvious increase in recent years. The first chart below plots the dispersion in four-quarter ahead forecasts of real GDP made by professional forecasters. It shows that the current level is more than one-standard deviation below its mean of the past 25 years. The second chart displays the (options-implied) volatility of the U.S. equity market. This measure, known as the VIX, climbed to 24.4 as of December 11, but remains less than one standard deviation above the median since 1990. (Similarly, recent work on inflation prediction suggests no greater difficulty forecasting now than in earlier periods.)
Dispersion in Professional Forecasts of 4 quarter ahead Real GDP (quarterly)
CBOE Volatility Index (monthly)
Source: FRED2 (VIXCLS).
But, looking closely at Chair Yellen’s comments—she mentions “evolving judgments concerning the implications of incoming information for the economic outlook”—we are led to wonder if policymakers might also be concerned about another source of uncertainty that plagues their decisions today. This is the impact that policy itself can have on the economy. On an operational level, central bankers are in uncharted territory. No one has ever tried to tighten policy using the tools the Fed will employ. And, with the changes in the structure of the financial system brought on by the combination of the crisis and the subsequent financial reforms, no one can say with any precision how rising interest rates will be transmitted to prices, employment and output.
In either case—whether we are unsure about the structure of the economy or about policy effectiveness—we are led to the same conclusion: policymakers will be naturally inclined to proceed with caution. Previously, we used the analogy of driving on a dark and foggy night on a road with a sheer cliff on one side. In such a circumstance, you go slowly (with your foot closer to the brake than the accelerator). And, if you are on your way to visit a friend, you are vague about when you are going to get there. Instead, you say that your arrival will depend on the road conditions along the way. That is, as uncertainty increases, you shift from forward guidance to data dependence. Doing otherwise would imply false precision—that you know more than you really do.