Brookings Papers on Economic Activity 2: 2004

Brookings Institution Press

Chapter One

The brookings panel on Economic Activity held its seventy-eighth conference in Washington, D.C., on September 9 and 10, 2004. This issue of Brookings Papers on Economic Activity includes the papers and discussions presented at the conference. The first paper evaluates unconventional measures available to monetary policymakers for stimulating the economy when interest rates are already near zero, a situation that may arise with price stability or negative inflation. The second paper presents empirical evidence on the effects of taxes, federal spending, and deficits on national saving, interest rates, and growth. The third paper explores the impacts on U.S. employment in recent years from conventional foreign trade in goods and from the rise in offshoring of service jobs. The fourth paper examines the effect of tax changes, such as those passed since 2000, on business capital formation.

CENTRAL BANKS USUALLY implement monetary policy by setting the short-term nominal interest rate that the bank controls, such as the federal funds rate in the United States. However, the success of many industrial countries over the years in reducing inflation and, consequently, average nominal interest rates has increased the likelihood that, during a recession, the policy rate will approach its lower bound of zero. When rates are at or near zero, a central bank can no longer stimulate aggregate demand by further rate reductions and must rely instead on "nonstandard" policy alternatives. An extensive literature examines these alternatives, but for the most part from a theoretical or historical perspective. Few studies have presented empirical evidence on their potential effectiveness in modern economies. Such evidence not only would help central banks plan for the contingency of the policy rate approaching zero, but also would bear directly on the choice of the appropriate inflation objective in normal times: the greater the confidence of central bankers that tools exist to help the economy escape the liquidity trap that occurs at the zero bound, the less need there is to maintain an inflation "buffer." Hence evidence of effective alternative policies would bolster the argument for a lower inflation objective. In the first article of this issue, Ben Bernanke, Vincent Reinhart, and Brian Sack apply the tools of modern empirical finance to the recent experiences of the United States and Japan to look for such evidence.

Following earlier work by Bernanke and Reinhart, the authors group nonstandard policy alternatives into three classes: official communications designed to shape public expectations about the future course of interest rates; quantitative easing, which increases both assets (holdings of government securities) and liabilities (unborrowed reserves) on the central bank's balance sheet; and changes in the composition of that balance sheet through, for example, targeted purchases of long-term bonds aimed at reducing long-term interest rates.

The authors' investigation employs two approaches. First, they perform event-study analysis, measuring and analyzing the behavior of selected asset prices and yields over short periods surrounding central bank statements or other financial or economic news. Second, they estimate "no-arbitrage" models of the term structure of interest rates for both the United States and Japan. For any given set of macroeconomic conditions and stance of monetary policy, these models allow the authors to predict interest rates at all maturities. Using the predicted term structure as a benchmark, they are then able to assess whether factors not included in the model-such as the Bank of Japan's quantitative easing policy that began in 2001-have economically significant effects on interest rates.

Bernanke, Reinhart, and Sack begin with a discussion of nonstandard policies that might be effective in stimulating the economy when short-term rates are at the zero bound; the discussion draws on the historical experience with such policies in the United States and Japan as well as on existing theories of potential policy channels and previous empirical analysis. The first type of policy they consider is the use of central bank communications to influence the market's expectations about future policy and hence future short-term rates. According to some theories, shaping expectations about future short-term rates is essentially the only tool central bankers have. But the authors take a broader view, arguing that private sector borrowing and investment decisions are more sensitive to longer-term yields than to short-term rates, and considering the possibility that long-term rates can be moved independently of expectations of short-term rates. This leads to a discussion of the potential importance of policy statements, credibility, and policy rules. Although the authors see "rule-like" central bank behavior, particularly state-contingent behavior, as an important means of shaping the public's policy expectations, they believe that a central bank would find it particularly difficult to establish in advance how it would react to highly unusual circumstances, such as when the short-term rate is near the zero bound. Hence in such cases statements about policy intentions and commitments are likely to be particularly important.

The second type of nonstandard policy, quantitative easing, involves purchasing government securities beyond what is required to drive the short-term rate to zero. The authors discuss three channels through which such a policy might operate to escape the liquidity trap. First, such purchases may lead to private sector rebalancing of portfolios, which in turn would raise the prices of other assets. However, the authors observe that there will be little incentive to rebalance if money is a good substitute for the short-term bills it replaces when the latter are paying close to zero interest. Second, a larger outstanding stock of money raises the prospect of higher seigniorage in the event of future inflation, substituting for direct taxes. The effectiveness of this "fiscal" channel requires that the public, in the midst of a deflation, expect future inflation and expect that the central bank will not withdraw the injected money when that inflation arrives. The authors believe that the fiscal channel could work if pursued aggressively enough and with a clear commitment not to reverse course. Third, the visible signal that quantitative easing provides makes it more believable that the central bank will hesitate to reverse such large purchases soon, perhaps because of the possible shock to money markets.

The third type of nonstandard policy involves altering the composition of the central bank's balance sheet by participating in all segments of the market in government debt, including inflation-indexed debt, so as to influence term, risk, and liquidity premiums. Using emergency provisions dormant since the 1930s, the Federal Reserve could even accept private financial and real assets as collateral for discount window loans. Although many economists are skeptical about the potential effectiveness of this channel, the authors note a number of historical examples of central banks effectively pegging long-term rates.

Bernanke, Reinhart, and Sack's own empirical investigation begins with an event study measuring the influence of Federal Reserve policy announcements by the response of three market-based indicators following selected decisions of the Federal Open Market Committee (the Federal Reserve's principal policymaking body) since 1991. The indicators are the current-month federal funds futures contract, the Eurodollar futures contract expiring in about a year, and the yield on Treasury securities of five years' maturity. The authors measure the responses of each indicator observed during the forty-five minutes following FOMC announcements; the very short time interval is chosen to minimize the extent to which other factors could be affecting rates. The change in the current-month federal funds futures contract simply measures the markets' reaction to news about the Federal Reserve's near-term funds target. The change in the year-ahead Eurodollar futures contract presumably incorporates both the effect of this funds rate surprise and the effect of any accompanying announcement on market expectations about policy actions and the economy over the coming year. The change in the five-year Treasury yield presumably includes both those effects plus those arising from revisions in expectations beyond a year. The authors decompose the second indicator into the part explained by the first factor, the funds "surprise," and the orthogonal residual, which they label a second factor. The change in the five-year rate not explained by the first two is designated a third factor. The authors find that only about 20 percent of the variance in the one-year-ahead rate during the forty-five-minute policy "window" is explained by the current policy surprise; unexplained movements in the year-ahead rate-the second factor-make up the remaining 80 percent. Again, this factor presumably captures any revisions in the private sector's expectations of future short-term rates due to information contained in the policy statement that accompanies the change in the funds rate.

Perhaps the most striking revelation of the authors' analysis is the high correlation between unexplained movements in the future one-year rate and the five-year Treasury yield: the second factor accounts for 68 percent of the variability of the five-year yield during the event window, and the policy surprise itself explains another 12 percent, leaving only 20 percent unexplained. Informal inspection of the historical behavior of the second factor reveals that it becomes increasingly important in the latter part of the sample, when policy statements came into regular use. In contrast, larger realizations of the first and third factors do not seem to line up with dates of policy statements. This leads the authors to a more formal investigation of the link between FOMC statements and the three factors. First, they regress the squared values of each of the factors on dummy variables indicating dates when statements were issued and the characteristics of those statements: The dummy variable STATEMENT takes a value of 1 on any date on which a statement was released. STATEMENT SURPRISE takes a value of 1 when, in the authors' judgment, the statement included information about the economy or the path of policy that most market participants would not have expected. Finally, PATH SURPRISE is set equal to 1 when, in the authors' judgment, the statement revealed new information about the likely future path of monetary policy. The authors attempt to assign "surprises" as objectively as possible, using commentaries written before and after each statement was released (including those of a leading financial firm that specializes in monitoring FOMC actions), internal Federal Reserve staff analyses of market reactions, pre-FOMC meeting surveys about expectations for the balance-of-risks part of the statement, and the results of a survey of expectations about the statement conducted by the New York Federal Reserve Bank. Of the 116 policy decisions in their sample, statements accompanied 56, and the authors identify 31 of them as having a significant element of surprise, and 9 of these, in turn, as path surprises.

In the regression explaining the first factor, the coefficient on STATEMENT is positive and significant, and the authors attribute this result to the fact that, for much of the sample, statements were made only on days when the federal funds target was changed. These are days on which the policy rate surprise tends to be relatively large. The coefficient on STATEMENT SURPRISE in this regression is negative and significant, suggesting that the FOMC viewed policy rate surprises and statements as substitutes, possibly because the FOMC was reluctant to issue surprising statements at the same time that it was also surprising the markets with the policy action. The PATH SURPRISE variable is insignificant.

The regressions explaining the second and third factors are of more interest, since they provide information about how FOMC statements affect market expectations. The mere issuance of a statement has essentially no effect on the variance of the second factor; in contrast, in periods when there is a statement surprise, the variance is nearly 200 basis points, and when there is also a path surprise, the variance is roughly 230 basis points. When the first two factors are controlled for, the five-year yield is not noticeably affected by policy statements, surprising or otherwise; thus statements do not appear to provide information about long-term yields independent of their influence on expectations about rates one year in the future.

FOMC statements often contain language that suggests the direction of future policy actions. Bernanke, Reinhart, and Sack investigate whether the direction of the response of the factors is consistent with whether a "surprise" statement appears hawkish, dovish, or neutral with regard to interest rates. They define two dummy variables, for statement and path surprises, each of which takes a value of +1 for surprises that are judged to be hawkish, -1 for those judged to be dovish, and zero otherwise. They find no significant response of either the first or the third factor to either of these dummies. However, they find that hawkish statement surprises increase, and dovish surprises decrease, the year-ahead rate by 12 basis points. The response to path surprises is an even greater 16 basis points. Both responses are highly significant. The authors note that, in addition to the official FOMC statements they study, the speeches and congressional testimony of FOMC members may be important in shaping expectations.

Beyond their direct effects, statements that are conditional-that is, that make the central bank's commitments contingent on specific economic developments-are likely to affect the market response to news about those events. For example, the statement that "[monetary] policy accommodation can be maintained for a considerable period," first introduced in Federal Reserve Chairman Alan Greenspan's semiannual report to Congress in July 2003, was, in subsequent FOMC statements, typically tied to labor market conditions and "slack" in the economy. A regression relating changes in the ten-year Treasury yield to surprises in the monthly payroll report shows greater sensitivity after August 2003 than in the preceding twelve years; this is consistent with the claim that this FOMC language heightened the market's attention to employment growth.

Continues...

---

Excerpted from Brookings Papers on Economic Activity 2: 2004 Copyright © 2004 by Brookings Institution Press . Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

---