The Nominal Anchor Problem: Is Inflation Targeting the Answer?
One issue that is not new is the need for a nominal anchor for the
economy in a regime based ultimately on the central bank’s setting a
nominal interest rate. As Bennett McCallum (1981) showed nearly two
decades ago, the limiting case of Wicksellian instability that Sargent and
Wallace had identified as an indeterminacy of the aggregate price level is
a consequence of a “pure interest rate peg” but is not a consequence of a
regime in which the central bank sets (and, presumably, varies) a nominal
interest rate at least in part as a means of influencing some nominal
magnitude. Although McCallum’s original demonstration of this distinction
relied on an example in which the central bank sets the short-term
interest rate in part as a way of targeting the money stock, the more
general point carries over in full to the case in which the interest rate is
the instrument used to target inflation, among other macroeconomic
variables—which is exactly what the Federal Reserve now does. (McCallum
showed that merely having at least one nominal target, among others
that may be real—for example, output or employment—is sufficient to
dispose of the price indeterminacy.)
These theoretical insights notwithstanding, it is fair to say that the
return to an interest-rate-based monetary policy regime, with neither
money nor any other nominal variable as an intermediate target, has left
many observers uneasy. In some countries—the United Kingdom, Sweden,
Canada, Australia, among others—the chosen solution has been to
adopt a formal “inflation target.” Whether doing so is of positive value
remains to be seen. As Ben Bernanke and coauthors (1999) have shown,
in many cases the countries that have adopted inflation targeting do now
enjoy significantly lower inflation rates than they did earlier on. But in
most cases the slowing of inflation in these countries had occurred before
the new regime was adopted. Hence the value of inflation targeting per
se remains unproved.
Exactly what “inflation targeting” consists of also remains unclear. In
the early stages of debate over this idea, it was sometimes taken to mean
that monetary policy would focus exclusively on inflation, with no regard
for real outcomes. (Parts of the bill offered in the U.S. Congress in 1996 by
the chairman of the Joint Economic Committee read in just this way.)
Although some economists and many central bankers probably would
favor such a change, the idea of relieving central banks of any responsibility
for output and employment attracted widespread criticism, and
advocates of inflation targeting quickly backed away from this interpretation.
In Mervyn King’s (1997) much-quoted phrase, such a regime
would amount to a policy of “inflation nutters.”
As Lars Svensson’s work (1997) has made clear, a different interpretation
of inflation targeting—the interpretation that seems to be accepted
by most of the central banks that have adopted this regime—is that it is,
in formal terms, fully consistent with the standard maximization of a
monetary policy objective function including both inflation and output.
One way to understand Svensson’s point is simply to recall that no matter
how many target variables monetary policymakers seek to influence, in
the end they have only one instrument with which to do so. Once having
decided on the setting of their instrument variable (that is, the level of the
THE ROLE OF INTEREST RATES IN FEDERAL RESERVE POLICYMAKING 55
short-term interest rate), it is then straightforward to express the chosen
policy in terms of any of the target variables, including inflation.
The logical question to ask about all this is what, then, is new or
different? The usual answer given by advocates of inflation targeting is
that it makes explicit, and therefore transparent to the public, the central
focus of monetary policymaking. On close inspection, however, this claim
seems dubious.
The one part of the claim for transparency that seems unquestionable
is that inflation targeting, as it is now conventionally understood,
obligates the central bank to identify, quantitatively, its long-run inflation
objective. All inflation-targeting central banks do so. But if that were all
there is to it, inflation targeting would amount to no more than King’s
“inflation nutter” policy. If it is to be more than that, target(s) for one or
more aspects of real economic activity must also be involved. And if
explicitness and transparency are what inflation targeting is supposed to
be all about, those features should presumably apply to the central bank’s
real target(s) as well.
For example, Svensson has usefully shown that there is a direct
relationship between the relative strength of policymakers’ preferences
with respect to inflation and real output (or employment) and the length
of the time interval over which it is optimal for monetary policy to seek
to return inflation to the target rate, once some unforeseen development
has rendered it different. If the policy weight on output is large vis-a`-vis
that on inflation, it is best to return to the targeted inflation rate slowly,
so as to minimize the associated dislocation of real economic activity.
With only a small weight on output, it is optimal to return to the targeted
inflation rate more rapidly. In the limit, with no weight at all on output
or any other real variable—King’s “inflation nutter” case—the central
bank would always seek to return inflation to the targeted rate immediately
after any disturbance.
Genuine transparency in an inflation targeting regime would therefore
include not only an explicit, publicly stated, quantitative inflation
target but also an explicit statement of the speed with which the central
bank would seek to return to that target after a departure from it.
Although a few central banks that have adopted inflation targeting
regimes have taken this step (the Bank of Canada, for example), most
have not. As a result, either the inflation targeting regime is not really
transparent—in which case the “what is new?” question continues to be
apt—or inflation targeting is really an “inflation nutter” policy after all.
On either interpretation the policy seems unsatisfactory.
An analogy to yet another aspect of central banks’ responsibilities
may help to explain this tension. The point of having a lender of last
resort is that, under some circumstances, the central bank will come to the
rescue of a bank that is facing difficulty in meeting its obligations. But the
more banks come to rely on this potential safety cushion, the more risks
56 Benjamin M. Friedman
they will take in conducting their business and the greater is the prospect
that the central bank will actually be called upon to come to their aid.
Hence most central banks would prefer that bank decision-makers
believe that (and therefore act as if) lender-of-last-resort actions were
much less forthcoming than is actually the case. In short, the goal is the
opposite of transparency: to induce the belief that banks are pretty much
on their own, and so had better structure their balance sheets soundly,
even while maintaining the lender-of-last-resort facility at the ready.
Similarly, many monetary policymakers today seek to benefit from
the “credibility” that goes with being perceived as all-out inflation
fighters, while at the same time in fact taking real considerations like
output and employment into account. The object is not to be an “inflation
nutter,” merely to be seen as one. Once again, the tension with the
often-avowed goal of transparency is clear.
In the United States, which has not adopted inflation targeting, the
“inflation nutter” policy would be precluded by law. The prevailing
legislation charges the Federal Reserve to conduct monetary policy “so as
to promote effectively the goals of maximum employment, stable prices,
and moderate long-term interest rates.” Abjuring any responsibility for
real outcomes would not be legal even if it were somehow thought
desirable. By contrast, there is nothing in the law to prevent the Federal
Reserve from defining “price stability” as a specific rate of change
(perhaps zero) of a particular price index, and quantifying the relative
weight on “maximum employment” in terms of a Svensson-type speed of
return in the event of a departure.
Would the additional explicitness and transparency be valuable? In
particular, would they contribute to addressing the “nominal anchor”
problem inherent in interest rate-based monetary policymaking? No one
really knows. But at least on the basis of the experience of the past decade
and a half, it is hard to know what visible shortcoming in U.S. monetary
policy such a change might be expected to correct. To the extent that
setting interest rates leaves monetary policy without a nominal anchor,
that lacuna has had little apparent consequence recently.
Long-Term Interest Rates and Equity Prices as Information Variables?
The widely discussed recent experiences of two countries, the United
States and Japan, have raised once again a question of long standing: Is
there a role in the monetary policy process for interest rates other than the
short-term rate that the central bank uses as its policy instrument? And,
in parallel, is there a role for equity prices?
The aspect of the U.S. experience that has called renewed attention to
this issue is the dramatic rise in equity prices in the latter half of the 1990s.
Innumerable analyses have examined the effect of higher stock market
wealth in spurring consumer spending, the use of stock market assets as
THE ROLE OF INTEREST RATES IN FEDERAL RESERVE POLICYMAKING 57
credit collateral, the impact of a lower cost of equity capital on corporate
capital spending, and so on. More recently, some of the same questions
have also begun to appear in the context of residential real estate prices.
In short, the issue is whether the Federal Reserve should tighten
monetary policy in light of these developments in U.S. asset markets.
The Japanese experience is more dramatic and also two-sided. In the
1980s both equity prices and real estate prices (including prices for
commercial property) surged in Japan, in what has subsequently become
known as the “bubble economy.” In the early 1990s both equity prices
and real estate prices fell sharply, triggering bankruptcies and rendering
most if not all major Japanese banks insolvent. As a result, the Japanese
economy has stagnated for much of the last decade. Many analyses of
Japanese monetary policy have concluded in retrospect that policy
should have been tighter than it actually was during the “bubble” period,
and easier than it was once the bubble collapsed.
Following the discussion in the first section of this paper, it is clear
that asset prices and long-term interest rates can potentially play a useful
role in the monetary policy process if they contain incremental information
about output or inflation, or any other macroeconomic targets that
policymakers seek to influence. The obvious question is, do they?
Tables 1 to 4 report the results of estimating a series of equations
designed to address just this question for the United States. Table 1 shows
the F-statistics summarizing a pair of baseline equations, one for real
output growth and one for inflation, that do not include any asset price or
long-term interest rate variables. Each equation includes four lags on both
the output variable (real GDP growth) and the price variable (growth of
the chain-weighted GDP price index), four lags on the federal funds rate
(taken as the Federal Reserve’s policy instrument), and four lags on the
growth of the M2 money stock. Data are quarterly. The sample is 1970:I
to 2000:II. Data for all variables other than the federal funds rate are
seasonally adjusted.
The results for the baseline equations reported in Table 1 are roughly
consistent with familiar findings throughout the empirical literature of
U.S. macroeconomic relationships. Inflation in particular is very highly
serially correlated on a quarterly basis. The federal funds rate contains
highly significant further information about the subsequent movement of
both output growth and inflation. The M2 money growth rate contains
information that is marginally significant in predicting output, but not
inflation.
- posted by Mark Pane @ 7:46 PM 0 comments

 
three open questions on interest rate based policy
three open questions surrounding today’s
interest-rate-based policy structure: (1) Has the Federal Reserve solved
the “nominal anchor” problem inherent in interest-rate-based monetary
policymaking? If not, would a shift to explicit inflation targeting (along
the lines of what the Bank of England and the Swedish Riksbank, for
example, have done) be helpful? (2) Is there a role in the monetary
policymaking process for interest rates other than the federal funds rate,
or whatever particular rate the Federal Reserve chooses to set? Equivalently,
is there a role for the prices of financial assets, including equities?
(3) To what extent does the electronic revolution now under way in
44 Benjamin M. Friedman
banking, and in business more generally, threaten the efficacy of interestrate-
based monetary policymaking?
- posted by Mark Pane @ 7:41 PM 0 comments