The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
)
January/February 1992
Volume 77, Number 1
Federal Reserve
Bank of Atlanta
In This Issue:
Financial Panics, Bank Failures,
and the Role of Regulatory Policy
What Hath the Fed Wrought?
Interest Rate Smoothing
in Theory and Practice
Forecasting Industrial Production:
Purchasing Managers' versus
Production-Worker Hours Data
/Review Essay
ftcpnpiiiic
January/February 1992, Volume 77, Number 1
—
Federal Reserve
Bank of Atlanta
President
Robert P. Forrestal
S e n i o r Vice President a n d
Director of R e s e a r c h
Sheila L. Tschinkel
Vice President a n d
Associate Director of Research
B. Frank King
Research Department
William Curt Hunter, Vice President, Basic Research
Mary Susan Rosenbaum, Vice President, Macropolicy
William Roberds, Research Officer, Macropolicy
Larry D. Wall, Research Officer, Financial
^mmmamm^mmmm
Public A f f a i r s
Bobbie H. McCrackin, Vice President
Joycelyn T. Woolfolk, Editor
Lynn H. Foley, Managing Editor
Carole L. Starkey, Graphics
Ellen Arth, Circulation
The Economic Review of the Federal Reserve Bank of Atlanta presents analysis of economic
and financial topics relevant to Federal Reserve policy. In a format accessible to the nonspecialist, the publication reflects the work of the Research Department. It is edited, designed, produced, and distributed through the Public Affairs Department.
Views expressed in the Economic
eral Reserve System.
Review are not necessarily those of this Bank or of the Fed-
Material may be reprinted or abstracted if the Review and author are credited. Please provide the
Bank's Public Affairs Department with a copy of any publication containing reprinted material.
Free subscriptions and limited additional copies are available from the Public Affairs Department, Federal Reserve Bank of Atlanta, 104 Marietta Street, N.W., Atlanta, Georgia 30303-2713
(404/521-8020). Change-of-address notices and subscription cancellations should be sent directly to the Public Affairs Department. Please include the current mailing label as well as any new
information. ISSN 0732-1813
ontents
January/February 1992, Volume 77, Number 1
Financial Panics,
Bank Failures,
and the Role of
Regulatory Policy
Stephen D. Smith and
Larry D. Wall
— _ — _
12
—
What Hath the Fed Wrought?
Interest Rate Smoothing
in Theory and Practice
William Roberds
25
During the past two decades financial difficulties have plagued
depository institutions and nonbank financial firms alike. Both discount window lending and closure policies have played important
roles in dealing with such problems. Recent legislative changes will
affect these roles.
The authors look at the impact of these changes on the FDlC's
and the Federal Reserve's ability to keep the financial system stable
at minimum cost to other banks and to taxpayers while ensuring that
the system continues to operate efficiently. Because the legislation
encourages the separation of closure decisions from the handling of
systemic risk, it may improve the financial system's efficiency, the
authors conclude.
FYI—Forecasting
Industrial
Production: Purchasing
Managers' versus ProductionWorker Hours Data
In the real-world economy of financial markets, reducing shortterm interest rate fluctuations—a practice referred to as interest rate
"smoothing"—is often viewed as an important role of monetary
policy. Despite the fact that academic economists often criticize or
ignore it, interest rate smoothing is used to some degree by the monetary authorities in every major industrialized country, including
the United States. This article discusses some of the most important
research on several complex questions concerning interest rate
smoothing and provides a brief history of the Federal Reserve System's approaches to open market policy.
This article examines two broad indicators of current manufacturing activity that analysts incorporate into forecasting models for
monthly industrial production. The author reports on how well these
series serve that purpose and presents several explanations for forecast error in models using these series.
R. Mark Rogers
37
Review Essay
B. Frank King
Breaking Financial Boundaries: Global Capital,
National Deregulation, and Financial Services Firms
by David M. Meerschwam
/Financial Panics,
Bank Failures,
and the Role of
Regulatory Policy
Stephen D. Smith and Larry D. Wall
Smith is the H. Talmage
Dobbs, Jr., Chair of Finance
in the College of Business at
Georgia State University and
a visiting scholar at the Federal Reserve Bank of Atlanta.
Wall is a research officer in
charge of the financial section
of the Atlanta Fed's research
department. This article was
completed while Smith was on
the faculty of the Georgia
Institute of Technology.
Federal
Reserve Bank of Atlanta
rowth in both the price volatility and volume of financial claims
in the United States during the late 1970s and 1980s has far exceeded normal patterns found in post-World War II history. Financial p r o b l e m s h a v e s i m u l t a n e o u s l y p l a g u e d a s i g n i f i c a n t
portion of the financial services sector—markets in which only
depositories are operating, in which only nonbank financial firms are operating, and in which both are operating. W h e n active, financial policy on the
part of the central bank and other federal regulators has often involved use
of both discount window borrowings and, in the case of certain large banks,
blanket coverage of all depository liabilities.
As a part of legislation to recapitalize the Federal Deposit Insurance Corporation (FDIC), Congress recently reconsidered the role of the discount
w i n d o w — t h e m e c h a n i s m by which the Federal Reserve lends to institutions unable to meet their reserve r e q u i r e m e n t s — a n d deposit insurance.
T h e p r o d u c t of this r é é v a l u a t i o n is legislation e n c o u r a g i n g significant
changes in ways the Federal Reserve and the F D I C handle problem institutions. In particular, the legislation requires regulators to intervene before an
institution is insolvent, limits the ability of the Fed to lend through the discount window to financially weak banks, and limits the F D I C ' s ability to
protect deposits in excess of $100,000 and deposits made at foreign branches of U.S. banks.
This study focuses on the implications of these changes for the F D I C ' s
and the Federal Reserve's ability to maintain stability in the financial system at m i n i m u m cost to other banks, the taxpayers, and the efficient operation of the financial system. The existing deposit insurance and discount
window structure are taken as given. Questions of whether an alternative
structure might be more efficient are therefore not considered. Moreover,
Economic Review
1
the article is not a comprehensive look at deposit insurance or the discount window. For example, the disc u s s i o n d o e s n o t e n c o m p a s s t h e u s e of d e p o s i t
insurance to protect small depositors or the use of the
discount window as a mechanism for conducting m o n etary policy in periods when there is no crisis.
The historic policy of mixing closings and systemic
risk via the so-called "too-big-to-fail" doctrine is inefficient to the extent that it increases closure costs without reducing systemic costs. This research concludes
that the recent legislative changes m a y improve the financial s y s t e m ' s e f f i c i e n c y because they e n c o u r a g e
the separation of closure decisions from the handling
of systemic risk. An additional finding is that effective
use of the discount window can control systemic problems without resorting to "too-big-to-fail" as long as
disruptions are not solvency based. Because regulators have more information than the market, the discount w i n d o w provides a m e c h a n i s m for conveying
the additional information, and appropriate use of the
discount window can hasten the end of systemic problems.
In order to evaluate the implications of the recent
congressional actions, the first section reviews the reasons that financial intermediaries m a y encounter financial problems and discusses when and why these
liquidity problems are a social concern. The next section provides a brief review of market and banking
panics and methods used (by private market participants, governments, and central banks) to deal with
them. The costs and benefits of the historic regime are
then compared with those of a restructured discountlending policy that e l i m i n a t e s g u a r a n t e e s to n o n i n sured creditors. T h e latter system is contrasted with
recent legislative changes, especially in terms of their
implications for banking system efficiency and systemic risk.
problem, it can almost always go to other financial intermediaries and the financial markets to obtain adequate funding. 1
If, however, market participants question the continued viability of a financial intermediary, liquidity problems can lead to the institution's demise. Concern about
insolvency can lead to withdrawals by current creditors
and can discourage potential investment by new creditors. To address this dilemma, intermediaries may try to
sell their relatively less liquid assets. Unfortunately,
forced sales of illiquid assets may only compound the
problems by resulting in losses that in turn increase
market concern about an institution's solvency.
T h e traditional explanation f o r the m i s m a t c h between the times to maturity of assets and liabilities is
that it allows intermediaries to exploit the law of large
numbers to increase their profitability. That is, most
intermediaries, especially depositories, experience inflows and outflows from their creditors (depositors). If
the number of creditors is sufficiently large, the bank
can predict with reasonable accuracy what the net outflow is going to be and provide only for this net figure
rather than the gross outflows. The difference between
the gross outflow and net outflow can be profitability
invested in less liquid assets (such as loans) that yield
a higher rate of return than highly liquid assets (such
as short-term securities).
^Reasons for Liquidity and Solvency
Problems at Financial Intermediaries
An alternative e x p l a n a t i o n — a n d the t w o are not
mutually exclusive—for the funding of illiquid loans
with shorter-term deposits suggests that the mismatch
is necessary to provide intermediaries with appropriate
incentives in allocating their portfolios. B a n k s and
other intermediaries can substantially change their asset structure over very short periods of time. Mark J.
Flannery (1991) points out that if these changes result
in a relative increase in high-risk assets, only "shortt e r m " liability holders would be in a position to demand a higher rate of return to cover increased risk.
Thus, the use of shorter-term deposits can substantially reduce any gains shareholders might realize from an
i n t e r m e d i a r y ' s o b t a i n i n g f u n d s in the debt m a r k e t ,
then changing its investment strategy.
The fact that financial intermediaries invest in assets that are less liquid—that is, less readily converted
to money—than their liabilities creates a potential for
withdrawals to exceed the sum of cash inflows and securities sellable for a small discount on the market
value. Intermediaries recognize this risk and ordinarily manage their position to maintain adequate liquidity. M o r e o v e r , s h o u l d an i n s t i t u t i o n that is w i d e l y
recognized as financially viable encounter a liquidity
Charles W. Calomiris and Charles M. Kahn (1991)
explain illiquidity by f o c u s i n g on the incentive for
management and shareholders to engage in inefficient
actions (for example, looting) when an institution approaches insolvency. 2 T h e y argue that the ability of
creditors to withdraw funds on short notice can be especially valuable in controlling such tendencies. Thus,
an implication of Flannery (1991) and Calomiris and
K a h n (1991) is that the decision to issue liabilities
with a maturity shorter than that of assets may be an
2
Economic Review
January/February 1992
integral part of achieving efficient financial intermediation for certain types of loans.
Whatever the cause of the liquidity imbalance, central banks are concerned about cases in which inaccurate market concern about an institution's viability can
m a k e the i n t e r m e d i a r y s u b j e c t to m a s s i v e d e p o s i t
withdrawals (or refusals to roll over lines of credit).
Private market creditors almost never have complete
information on their i n t e r m e d i a r y ' s financial condition, and, occasionally, this lack of information can
lead them to the conclusion that the institution is no
longer viable. W h e n a large fraction of the liabilities
are such that they can be withdrawn on demand, the
potential problem is greater. Creditors may very well
view withdrawal as their best option because leaving
funds in the firm puts them at risk to lose part of their
investment should it fail. Alternatively, if they withdraw from an intermediary that proves viable after all,
they can simply reinvest their funds. A recent paper on
cascades in financial markets by Sushil Bikhchandani,
David Hirshleifer, and Ivo Welch (1991) provides a
rationale for withdrawal by investors. They show that
even those investors that have favorable information
may withdraw funds if a significant number of other
investors signal that they have adverse information by
withdrawing their deposits.
These liquidity "runs" on an otherwise viable institution impose private and social costs. Owners stand to
lose their firm's going-concern value if their ability to
transact routine business is interrupted, and management faces potential loss of their jobs. Social costs bec o m e a factor if the failure of one institution raises
doubts about the solvency of others. These induced runs
can cause additional costs if creditors demand currency
for withdrawn deposits because converting deposits to
currency can lead to an implosion of the money supply
in a fractional reserve system. (In such a system only a
portion of bank deposits are held as reserves and the remainder are returned to circulation, thus enabling the
money supply to expand or contract as a multiple of the
deposit base.) The potential reduction in the money supply is mitigated, according to George G . K a u f m a n ' s
(1987) argument, by the fact that it is impractical for
large institutions to w i t h d r a w c u r r e n c y . T h e r e f o r e ,
Kaufman argues, funds withdrawn from one bank will
be redeposited in another, presumably solvent, institution. Moreover, the central bank can generally offset reductions in reserves through open market operations.
However, as E. Gerald Corrigan (1989/90) has noted,
concern about bank failures can lead to disruption in the
payment settlement or clearing systems. An additional
point, noted earlier, is that a viable intermediary's fail-
Federal
Reserve Bank of Atlanta
ure may still cause some businesses to lose access to
new credit (see Ben S. Bernanke 1983) and simultaneously induce intermediaries to engage in suboptimal actions. For example, the intermediary may take less risk
or hold m o r e equity capital than is socially optimal.
Such activities reduce the potential for real (inflationadjusted) economic growth.
/historical Practices
The liquidity pressures discussed earlier may arise
at both viable and nonviable financial institutions. Prior to the development of formal government institutions that could intervene, the threat of problems at
o n e i n t e r m e d i a r y spilling o v e r to a n o t h e r was frequently great enough that private market participants
would act in concert to stabilize prices and lend to
institutions that were sound but f a c e d temporary liquidity problems. 3 However, the limited resources of
private-sector participants, problems of free-riding behavior (whereby some participants benefit f r o m the
provision of the service without paying), and the lack
of an incentive for intermediaries singly or as a group
to consider all of the social costs associated with panics helped lead to the development of central banks
and other monetary authorities. This development has
centralized the activities of both price stabilization and
the so-called lender-of-last-resort functions in modern
economies. The concern of this study is with the latter
aspect of central banking. 4
T h e classical view of the lender-of-last-resort function, as outlined by Walter B a g e h o t (1873), a r g u e s
that the c e n t r a l b a n k s h o u l d l e n d f r e e l y , but at a
penalty rate, to institutions that are solvent but facing
a temporary liquidity shortage. Interestingly, he does
not limit this prescription to d e p o s i t o r y i n t e r m e d i aries. B a g e h o t n o t e s that " T h e B a n k lof E n g l a n d ]
does, in fact, at every period of pressure, advance to
the bill brokers; the case may be considered ' e x c e p tional,' but the advance is always m a d e if the security
offered is really good' (1873, 282; emphasis added).
Although Bagehot makes the case that the bill brokers
should b e discouraged f r o m accepting very short-term
(on-demand) deposits, h e argues that it is "inevitable"
that secured borrowings f r o m the central bank be allowed during times of crises. Bagehot clearly takes a
broad view of both f i n a n c i a l crises (as o p p o s e d to
bank runs per se) and the corresponding role of the
central bank to provide advances against " g o o d " securities.
Economic Review
3
K a u f m a n (1991, 106) suggests two m a j o r justifications f o r e s t a b l i s h i n g a lender of last resort in the
1800s: (1) concern about a decline in the money supply
in a specie-based monetary system and (2) "to offset
temporary liquidity strains f r o m adverse shocks that induced a large number of market participants to reassess
quickly their asset portfolios and sell some assets without a concurrent threat to the money supply." K a u f m a n
notes that the conversion to fiat money (created by the
government or monetary authority) and the existence
of deposit insurance has substantially reduced the importance of the first rationale but that the second reason
still exists. In K a u f m a n ' s view the justification for recent lender-of-last-resort activity is prevention of losses at "fire-sale" prices when intermediaries experience
a sudden, sharp increase in withdrawals.
The operation of the U.S. system of crisis management in the recent past differs from the Bagehot principles in that (a) the lender-of-last-resort function played
by the Federal Reserve is for the most part directly restricted to depository intermediaries, (b) the recomm e n d a t i o n that the l e n d e r - o f - l a s t - r e s o r t s e r v i c e be
extended only to healthy institutions is routinely violated, and (c) advances are m a d e at rates below, rather
than at or above, market rates of interest. Furthermore,
these changes in lender-of-last-resort activity are not
easily rationalized under K a u f m a n ' s (1991) analysis,
especially the provision of loans to failing banks.
A n o t h e r i m p o r t a n t c h a r a c t e r i s t i c of t h e c u r r e n t
U.S. system is that there exists a separate agency, the
FD1C, that insures the deposits of commercial banks
and thrifts. The FD1C, designed to provide a " b u f f e r "
against losses for small depositors, w a s created by
legislation passed in response to bank runs in the early 1930s. 5 Almost immediately, however, the agency
" o b t a i n e d statutory authority to f a c i l i t a t e m e r g e r s
b e t w e e n s o l v e n t b a n k s w h e n o n e of t h o s e b a n k s
was considered very weak and a future candidate for
failure (and therefore, cost to the F D I C ) " (Golembe
R e p o r t s 1991, 4 ) . A l t h o u g h the stated criteria f o r
dealing with problem banks is now the lower cost of
receivership or merger, it is clear that in either case
the insurance agency's role was intended to be at the
micro- economic (or firm-specific) level. Initially, this
agency had no regulatory role per se in dealing with
financial panics or bank runs beyond the promise that
s m a l l , u n s o p h i s t i c a t e d i n v e s t o r s n e e d not b e c o n cerned (except with loss of interest) about the financial stability of their chosen intermediaries. 6
The role of deposit insurance has expanded beyond
protection of small depositors to encompass, in certain
instances, all deposit liabilities—the so-called too-big4
Economic Review
to-fail doctrine mentioned earlier. Although the F D I C
is sometimes credited with preventing panics that might
otherwise have occurred since its creation, Michael D.
Bordo, a m o n g others, has noted that many countries
have managed to avoid runs without the existence of
d e p o s i t i n s u r a n c e . In particular, the fact that t h e r e
were "other countries that were panic-free before the
1960s and 1970s . . . suggests that such deposit insurance is not required to prevent banking panics" (1990,
27). Neither is the fund, in its current form, sufficient
to stop withdrawals by large depositors, as the case of
Continental Illinois s h o w e d (see, for e x a m p l e , John
M. Berry 1991). This application of the too-big-to-fail
doctrine, which is considered the first, in 1984, has
been repeated on a number of occasions, including as
recently as 1990 in a case involving the Bank of New
England. 7
P o s s i b l e Regulatory Responses and
Recent Legislative Changes
As noted, a liquidity crisis may be confined to a limited set of intermediaries whose failure would have only a small impact on real economic activity. However,
such a crisis could involve liquidity problems threatening to significant portions of the real, or nonfmancial,
sector. The discussion below considers both localized
liquidity problems and systemic problems.
P r o b l e m s at I n d i v i d u a l Institutions. W h e n the
market doubts a firm's viability, private-sector funding
is not likely to be forthcoming. Both the central bank
and the deposit insurer can help reduce the costs associated with runs on viable institutions. T h e central
bank can do so by providing temporary liquidity to
bank and nonbank financial firms, and the deposit insurer can reduce costs at banks (nonbanks are not insured) by agreeing to absorb any losses in the event of
f a i l u r e , t h e r e b y r e m o v i n g the p r i n c i p a l r e a s o n f o r
w i t h d r a w a l s . H o w e v e r , b o t h a g e n c i e s c a n also increase the social costs of b a n k failure if they allow
nonviable institutions to continue in operation.
Central Bank. Central banks may be asked to provide liquidity when the market signals its perception of
problems by withdrawing funds f r o m an intermediary.
The lender of last resort can delay the institution's closing by m a k i n g loans through the discount window.
When the market receives new information suggesting
that the firm is viable, the intermediary can return to
the market for funding. Thus, the key to a successful
lender-of-last-resort operation is the subsequent release
January/February 1992
of good n e w s about the firm in question. If no such
good news is forthcoming, lender-of-last-resort action
only postpones the inevitable closing of the institution. 8
Stated d i f f e r e n t l y , s u c c e s s f u l l e n d e r - o f - l a s t - r e s o r t
activities on behalf of individual institutions require an
information asymmetry such that the regulators know,
at least temporarily, more than the market or are more
willing to lend based on identical information.
Strong proponents of "efficient markets" would argue that market prices fully impound all available information and, hence, the central bank cannot improve
on market actions. This view has considerable validity,
particularly in areas such as the analysis of relative
prices, for which the central bank is unlikely to have
e x c l u s i v e i n f o r m a t i o n . H o w e v e r , in s o m e cases releasing information m a y be costly (for example, w h e n
it would assist the intermediary's competitors) or obtaining information m a y be costly for market particip a n t s . In s u c h c a s e s t h e m a r k e t m a y i n d e e d l a c k
information that regulators have been privy to in examinations and confidential discussions, information
about the composition of an intermediary's assets, for
example, especially its exposure to failing borrowers
and the seniority of claims on failing borrowers. It is
this short-term information advantage that provides a
wedge giving a lender of last resort r o o m to act for
short periods of time.
Lending by the central bank when it does not have
such an information advantage imposes costs on other
creditors and on society. D i s c o u n t w i n d o w lending
must be collateralized in the United States, m e a n i n g
that, should a firm fail after receiving discount wind o w loans, its losses will b e spread over a smaller
number of creditors. Such loans may eventually lead
to transfers of wealth f r o m creditors and those with
contingent liabilities (such as the FDIC) that do not (or
c a n n o t ) w i t h d r a w to t h o s e w h o d o w i t h d r a w their
backing from a failing intermediary. Further, lending
without information ensuring an institution's soundness potentially allows nonviable institutions to continue making inefficient investments.
Another limitation on the effectiveness of the lender
of last resort is that its position of advantage declines
rapidly. Once information has been recognized as relevant and favorable for judging an intermediary's solvency, the firm has a strong incentive to release it, and
the market has an equally strong incentive to seek out
the information. The implication of this set of factors
is that true liquidity crises should not last long. If a
bank cannot reacquire the market's confidence within
a short time, the firm probably is not viable, and continued lending would only postpone its inevitable clos-
Federal
Reserve Bank of Atlanta
ing. That is, long-term lending f r o m the discount wind o w is inconsistent with the idea that true liquidity
problems are short-term in nature because private market participants quickly adjust to new information.
T h e 1991 banking bill limits the Federal Reserve's
discount window lending. Federal Reserve Banks are
generally prohibited f r o m lending to an undercapitalized bank (as defined by the regulators) for m o r e than
60 days in a 120-day period. Reserve B a n k lending is
further limited to five days for a critically undercapitalized bank. A Reserve Bank m a y exceed the 60-day
limit if the relevant agency or the C h a i r m a n of the
Board of G o v e r n o r s of the Federal Reserve System
certifies that the institution is viable. If either lending
limit is violated without the viability certification and
the bank fails, the Board may be liable to the F D I C for
at least part of the losses.
The 60-day period will generally provide more than
adequate time for a troubled bank to prove its viability
to the financial markets. Indeed, if the Federal Reserve
uses the full 60 days to lend to a bank that the market
considers nonviable, a large fraction of the uninsured
depositors are likely to h a v e withdrawn their f u n d s ,
t h e r e b y i n c r e a s i n g the F D I C ' s s h a r e of t h e f a i l e d
bank's losses.
The lender of last resort can strengthen the value of
its discount window actions by lending only to solvent
but illiquid institutions. Indiscriminate discount wind o w activities have two adverse implications for an int e r m e d i a r y ' s r e m a i n i n g c r e d i t o r s : (1) o t h e r m a r k e t
institutions may judge the firm too risky, and (2) these
creditors' portion of any losses on failure is increasing
as the central bank acquires title to the good assets. In
such a case discount window lending may actually increase the rate of withdrawal of funds f r o m the intermediary. However, if loans are k n o w n to be restricted
to solvent institutions, central bank lending signals that
private information suggests an institution's solvency.
A problem arises when the central bank does not necessarily have an information advantage over the market,
as is currently the case in the United States with respect
to investment banks. The lender of last resort can lend to
these intermediaries without increasing the deposit insurer's exposure because investment banks are not insured. H o w e v e r , lending operations of this type are
likely to induce ex post wealth transfers because only the
lender of last resort can perfect its security interest in a
failing firm's collateral (that is, ensure its possession of
the collateral). Because such wealth transfers are undesirable, it would be advisable for the central bank to
avoid lending to intermediaries in nonsystemic situations, especially when it lacks an information advantage.
Economic Review
5
Deposit Insurer. Like the central bank, the deposit
insurer can potentially intervene to stem a run on an individual institution. The deposit insurer can do so by
absorbing the risk of loss should the intermediary fail.
However, absent systemic problems there seem to be
sound reasons that deposit insurance should not be
used to protect intermediaries f r o m failing. For one
thing, in the event of an institution's failure the insurer's function would merely transfer wealth from taxpayers in general to an intermediary's creditors, without
conferring any counterbalancing social benefit. Another consideration is that blanket deposit insurance also
reduces the value of market signals to regulators.
There may, of course, also exist circumstances under which firms that are still able to borrow in private
m a r k e t s will, on the basis of i n f o r m a t i o n that is as
g o o d or b e t t e r than p r i v a t e - s e c t o r i n f o r m a t i o n , b e
deemed unhealthy by regulators. Because these firms
will not generally seek help f r o m the discount window, i n t e r v e n t i o n by the p r o p e r r e g u l a t o r y a g e n c y
would result in an "early" closure (or merger) of the
institution. "Early" is used here to denote the fact that,
b e c a u s e they h a v e less precise information, private
market lenders would not deny the firm credit.
Systemic Risk. Systemic liquidity pressure occurs
w h e n institutions simultaneously attempt to shift to
more liquid and secure investments, adversely affecting the liquidity position of a substantial portion of the
financial system. Increased liquidity n e e d s — f o r example, demand for cash after a stock market crash—can
lead to such pressure. Alternatively, the crisis can result from solvency concerns about a substantial part of
the financial system.
If increased liquidity needs are causing the pressure,
the central b a n k can alleviate the p r o b l e m through
open market operations, as it did in O c t o b e r 1987.
Discount window lending m a y be appropriate if the
amount of Treasury securities is insufficient to meet
the central bank's reserve target.
A crisis arising from doubts about solvency is more
d i f f i c u l t to address. T h e c o n c e r n can arise f r o m a
shock involving only a f e w nonviable intermediaries
that the market cannot distinguish f r o m viable firms.
In turn, the shock may actually threaten the viability of
a s i g n i f i c a n t p o r t i o n of the f i n a n c i a l s y s t e m . T h e
shock can come from a "large" source, such as the default of a major international borrower or it can come,
as Corrigan (1989/90) notes, from a seemingly small
s o u r c e s u c h as E S M G o v e r n m e n t S e c u r i t i e s . T h e
shock can even come from reductions in stock market
values to the extent that purchases of equity securities
are partially financed by debt. 9 Whatever the initiating
6
Economic Review
factor, it is the potential inability by otherwise viable
institutions to pay debt claims on short notice that creates the "systemic" (or macroeconomic) risk that concerns regulatory bodies. In theory multiple avenues
are open to authorities for dealing with systemic risk.
Central Bank. The threat posed to a substantial part
of the financial system is important in determining the
appropriate central bank response. Consider first the
case of depositories. If most are solvent, the central
bank could act as it does when a nonsystemic problem
exists, simply lending to viable intermediaries and refusing credit to nonviable firms. If the crisis does jeopa r d i z e the viability of a s u b s t a n t i a l p o r t i o n of the
financial system, the role of the lender of last resort
becomes m o r e difficult. Lending m a y be required in
order to provide the deposit insurer with time to address the problems at the nonviable banks.
The case of nondepositories is more complicated.
A s n o t e d earlier, the lender of last resort generally
d o e s not h a v e an i n f o r m a t i o n a d v a n t a g e for these
firms and, thus, should not lend in noncrisis situations.
However, these f i r m s also m a y experience liquidity
problems that could threaten the stability of the real
economy, and in such a situation the lender of last resort may wish to assist them. O n e way of coping with
the general crisis is to work indirectly through commercial banks. Doing so shifts the problem of evaluating solvency back to the commercial banks, but in fact
these banks may have an information advantage f r o m
their routine dealings with the nonbank firms.
This indirect mechanism of lender-of-last-resort activities has appeal to the extent that (a) the central
bank is more familiar with the financial health of firms
over which it has some regulatory oversight and (b) such
an action limits the ability of the private sector to borrow at a discount rate that is typically below market
rates of interest. Conversely, the extra step in this process creates other problems, namely, that there must be
a cost/benefit system for banks that in times of systemic stress allows for the use of such regulatory "moral
suasion," whereby regulators would seek to persuade
commercial banks to lend to their troubled peers for
the sake of the social good. Offsetting subsidies (either
implicit or explicit) are therefore required to induce
depository firms to extend funds under circumstances
in which, because of risk aversion or other market factors, t h e y m i g h t o t h e r w i s e not m a k e t h e s e l o a n s .
George Kanatas (1986), for example, provides a model in which the central bank performs a lender-of-lastresort function that amounts to lending to institutions
that a risk-neutral investor would deem creditworthy
even though private risk-averse lenders would not.
January/February 1992
A p p l y i n g this i d e a to the i n d i r e c t p r a c t i c e d i s cussed above, regulatory authorities would n e e d to
absorb the risk by subsidizing "intermediate" lenders
by an a m o u n t e q u a l to the p r e m i u m f o r risk a v e r sion. 1 0 More generally, the very fact that moral suasion must c o m e into play implies that a purely private
cost/benefit analysis m a k e s these loans unattractive
for depository intermediaries. B e c a u s e s u c h m o r a l
suasion m a y not always work, it is risky to count on
any indirect f o r m of liquidity provision to prevent a
period of stress f r o m b e c o m i n g a f u l l - b l o w n crisis.
Historically, r e m e d i e s for this risk involved e i t h e r
(a) open m a r k e t operations or (b) the clause in the
Federal Reserve Act (13-3) that allows for lending to
nondepository firms or individuals during times when
their failure would "adversely affect the e c o n o m y . "
However, because the discounting must involve agricultural, commercial, or industrial purposes, security
brokers previously were effectively constrained f r o m
using the discount window; discounting for purposes
of carrying securities was omitted f r o m the "acceptable purposes" list. The current legislation amends the
clause by eliminating specific requirements for use of
the discount window borrowings, thereby giving the
central bank greater flexibility for providing liquidity
in times of stress. Moreover, because acceptable collateral at the discount window is at the discretion of
the Federal Reserve Bank, the inclusion of securities
firms has the effect of expanding the asset base that
can quickly be turned into cash. O p e n market operations typically have involved a m u c h more homogen e o u s c l a s s of i n s t r u m e n t s (that is, U . S . T r e a s u r y
securities).
While the current legislation enhances the central
bank's ability to deal with systemic risk problems, it
is still possible, in theory, for the entire system to be
s u f f i c i e n t l y illiquid that the a g g r e g a t e d e m a n d f o r
liquidity exceeds the supply of collateralizable securities. An a n a l o g o u s d e m a n d / s u p p l y i m b a l a n c e
might hold at the individual firm level as well. C o n sider an institution that h a s collateral w h i c h u n d e r
normal c o n d i t i o n s w o u l d be s u f f i c i e n t to m e e t its
counterparty obligations but that under certain special circumstances is incapable of posting collateral.
According to the a r g u m e n t s presented here, such a
situation is not a systemic liquidity problem that calls
for m a n a g e m e n t by the central bank. Rather, it represents a risk m a n a g e m e n t decision on the part of the
f i r m , a d e c i s i o n that s h o u l d not be s u b s i d i z e d , ex
post, by the public. B e c a u s e the central b a n k could
lend to the collateralized counterparties of this firm if
problems develop, the systemic risk associated with a
Federal Reserve Bank of Atlanta
possible failure of this particular institution is substantially alleviated.
Federal Deposit Insurance Corporation.
T h e policy e m p l o y e d by the F D I C — t h a t of insuring all deposit claims at certain institutions—has been justified
by systemic risk concerns. However, if only a small
segment of the financial system is in fact nonviable
and the discount w i n d o w is operating properly, deposit insurance seems to have limited value. P r o m p t
extension of deposit insurance to all depositors may
s h o r t e n a c r i s i s , but it w o u l d d o so at the c o s t of
eliminating postcrisis market signals about intermediaries' viability. Moreover, to the extent that a finan-
Recent legislative changes may improve the
financial systems efficiency because they
encourage the separation of closure decisions from the handling of systemic risk.
cial crisis can be brought on by actions of depository
or nondepository intermediaries, such blanket government guarantees are redistributive. T h e s e costs can be
justified only by assuming that either (a) some counterparties to transactions with these depository firms
are not sufficiently solvent to withstand default on this
fraction of their contracts or (b) the discount w i n d o w
policy of the central bank is not capable of dealing
with the s e c o n d a r y liquidity e f f e c t s resulting f r o m
these individual defaults.
The types of collateralization and lending policies
r e c o m m e n d e d earlier would provide for coverage of
liquidity concerns without regard to the secondary activities of other depository firms. Moreover, to justify
a too-big-to-fail policy on the basis of counterparties'
marginal solvency implies that policymakers are ess e n t i a l l y e x t e n d i n g " f a i l u r e i n s u r a n c e " for a w i d e
range of institutions. 11 The cost of such actions is, of
course, borne by the public.
If a significant portion of the financial system is in
fact nonviable, blanket guarantees of deposits at all institutions may be required to give the deposit insurer
Economic Review
7
time to resolve failed institutions without substantial additional damage to the real sector. However, if such a
situation has developed, the mistake was m a d e prior
to the crisis. Letting intermediaries maintain levels
of capital and diversification that create an environment in which a single shock can cause a large n u m ber of failures is fundamentally a different problem
than the temporary lack of liquidity healthy firms may
be experiencing.
The current legislation proposes to reduce resolution costs by prohibiting the F D I C f r o m protecting
uninsured depositors or nondeposit liabilities in most
circumstances. If the F D I C arranges for another bank
to purchase (some of) the failed bank's assets and ass u m e ( s o m e o f ) the failed b a n k ' s liabilities, it m a y
transfer uninsured deposit liabilities to the acquirer so
long as the f u n d suffers a loss on the uninsured deposits no greater than it would have incurred on those
deposits had the institution been liquidated.
The important exception to prohibiting coverage of
uninsured deposits involves systemic risk. This exception c a n be i n v o k e d if f a i l u r e to c o v e r u n i n s u r e d
claims would have serious adverse effects on general
economic conditions or financial stability. Under current legislation the F D I C is allowed to exercise the
systemic risk exception if authorized by (1) at least
two-thirds of the Board of Directors of the FDIC, (2) at
least two-thirds of the Board of Governors of the Federal Reserve System, and (3) the Secretary of the Treasury
(in consultation with the President). If the systemic
risk exception is exercised, the F D I C is required to recover the losses expeditiously with a special assessm e n t o n m e m b e r s of t h e i n s u r a n c e f u n d . T h e
assessment is to be levied on each bank's average assets minus its average tangible equity and average total subordinated debt. 12
Analysis of New Legislation
Table 1 represents what seems to be an "ideal" separation of closure and discount window policies. The
shaded areas depict situations in which either private
market or regulatory actions cause a spillover effect to
"secondary" institutions. Three points merit emphasis.
The first involves the fact that, in the system the table
describes, closure decisions for individual institutions
(and, by extension, "nonclosure" decisions) would not
depend at all on whether there is systemic risk in the
system. This point is reinforced by the observation
that in this system discount window lending is the on-
8
Economic Review
ly tool (other than, possibly, open market operations)
that is used to deal with secondary firms during times
of crisis.
T h e second point of interest in the proposed scenario turns on the fact that, in the absence of systemicrisk conditions, discount w i n d o w lending would be
used only to support individual institutions that the
market has incorrectly (and presumably temporarily)
d e e m e d to be n o n v i a b l e . D i s c o u n t w i n d o w lending
w o u l d never be e m p l o y e d to k e e p open institutions
that both the market and regulators believe to be nonviable.
T h e final element to emphasize involves the case
in which the central b a n k ' s information is "neutral,"
as it might be in the case of dealing with institutions
over which it has no regulatory authority. In this case
no action would be taken in an insolvency that has no
secondary influences. However, should such a closure
cause a crisis for other institutions, the central bank
would stand ready to lend to those secondary f i r m s
that have good collateral, including but not limited to
depositories. Likewise, in other situations causing a
liquidity squeeze for secondary institutions the central
bank could extend discount window lending to those
institutions. T h e recently passed legislation f o l l o w s
these lines with two possible exceptions. The first is
the escape clause for invoking the too-big-to-fail doctrine. Under the current f r a m e w o r k regulators may, in
cases of p e r c e i v e d s y s t e m i c risk (indicated on the
table by the shaded areas), choose both coverage of
all liabilities and lending to s e c o n d a r y institutions.
Such a response would seem justified only if a large
fraction of the banking system is not viable so that refusal to apply the too-big-to-fail principle would result in collapse of the financial system. Second, as a
matter of practice the secondary institutions may involve only depository firms, even though the new law
expands the potential for lending to nondepository intermediaries.
It is r e c o g n i z e d that the n e w r e g u l a t o r y s y s t e m
must start from a less than neutral point. Old regulations and bank actions have combined to provide an
environment in which temporarily continuing to observe the too-big-to-fail doctrine may be needed as a
transitional fallback position in crisis periods. Moreover, limited access to the discount w i n d o w remains
a viable option when the central bank has little or no
information concerning the balance sheets of nondepository financial intermediaries. Despite these "transitional" frictions, the recent legislation appears to be
a s u b s t a n t i a l m o v e t o w a r d s e p a r a t i n g c l o s u r e and
systemic-risk management problems.
January/February 1992
Table 1
Alternative Regulatory Responses to Failures,
Based on Regulator and Market-Based Information
Market-Based Assessment
Regulatory
Assessment
Solvent
Insolvent
Solvent
No regulatory action
Discount window lending
to primary institutions
No reason for a crisis at
secondary institutions
Discount window lending
to both primary and
secondary institutions
No regulatory action
Regular closure of primary
No reason for a crisis at
secondary institutions
Discount window lending
to secondary institutions
"Early" closure or merger
for primary institution
for primary institution
"Early" closure for
primary, discount window
lending to secondary
institutions
Regular closure for
primary, discount window
lending to secondary
institutions
institution
Neutral
Insolvent
Regular closure or merger
Market or regulatory action induces crisis for secondary firms
Conclusion
Financial panics and bank runs have been a part of
the U.S. financial system for at least t w o centuries.
Central bank lending via the discount window and closure policies on the part of g o v e r n m e n t a l a g e n c i e s
have played important roles in handling these crises.
This article has provided a review of historical and
current practices relating to these areas of action, emphasizing the distinction between regulatory actions
aiming to avert systemic risk (or crises) and those actions primarily addressing problems isolated at an individual institution.
T h e article suggests that in a world where regulators and m a r k e t participants h a v e d i f f e r i n g (and at
Reserve Bank of Atlanta
DigitizedFederal
for FRASER
times conflicting) information, there exist useful discount w i n d o w and closure policies, which h a v e the
following characteristics:
• Discount w i n d o w lending is used solely to deal
with systemic risk issues except in those cases for
which the central bank's private information suggests that an individual bank is viable even though
market-based financing is unavailable.
• In extreme circumstances discount window lending is made available to secondary institutions that
are nondepository intermediaries.
• Closure rules are not related to systemic risk issues. Systemic risk is dealt with through the discount window policies discussed above or through
open market operations.
Economic Review
9
T h e recently a d o p t e d b a n k i n g legislation m o v e s tow a r d the separation of isolated s o l v e n c y a n d s y s t e m i c
liquidity p r o b l e m s . R e g u l a t o r s are m a n d a t e d to r e s o l v e
p r o b l e m b a n k s promptly. B a n k s w i t h i n a d e q u a t e capital will be required to d e v e l o p and i m p l e m e n t a p l a n to
i n c r e a s e their capital w e l l b e f o r e t h e y b e c o m e i n s o l v e n t . P r o b l e m b a n k s u n a b l e to r a i s e their capital are
s u p p o s e d to b e p l a c e d in r e c e i v e r s h i p o r c o n s e r v a t o r ship p r i o r to b e c o m i n g i n s o l v e n t . T h e ability of t h e
F D I C a n d F e d e r a l R e s e r v e to p r o t e c t u n i n s u r e d dep o s i t o r s is also c u r t a i l e d . In a d d i t i o n , the bill h a s e x p a n d e d the F e d e r a l R e s e r v e ' s ability to s u p p l y liquidity
to n o n b a n k f i r m s in a s y s t e m i c liquidity crisis. T h u s ,
e x c e p t i n g the t r a n s i t i o n a l p e r i o d , t h e n e w legislation
s e e m s to h a v e p r o v i d e d a v i a b l e f r a m e w o r k f o r m a n a g i n g b a n k c l o s u r e s by the F D I C w h i l e a l l o w i n g the
c e n t r a l b a n k m a x i m u m f l e x i b i l i t y to deal with a n y related crises.
Notes
1. The illiquid intermediary may have to pay a premium to obtain funds on short notice, but the extra cost has the side
benefit of encouraging intermediaries to manage their liquidity position prudently.
2. This danger exists for all institutions but may be worse for
intermediaries.
3. See, for example, Chernow (1991, particularly 124-25) for a
discussion of the consortium put together by Morgan to deal
with the potential closing of the New York Stock Exchange
during the panic of 1907. Interestingly, this group also made
decisions concerning the viability of individual trust companies. Much like a central bank, some were allowed to fail
(for example, Knickerbocker) while others (for example,
Trust Company of America) were deemed healthy and provided with emergency credit.
on the dollar and that the crisis was abated by restrictions on
gold payments and the devaluation of the dollar.
6. Emmons (1991) provides a model with roles for both the
lender of last resort and deposit insurance. The lender of last
resort limits the damage from early liquidation of a bank resulting from a run by depositors that monitor a bank's portfolio, while deposit insurance removes the incentives for
nonmonitoring depositors to join a bank run after it has begun.
7. Although Continental is generally cited as the first application of too-big-to-fail, Golembe Reports notes that "in a
very basic sense [too-big-to-fail] was born on December 4,
1956, when the Home National Bank of Ellenville, New
York failed" (1991, 5). The bank had a local box factory as
one of its large depositors, and employees of the box factory
were facing a grim Christmas season.
4. As noted earlier, the purpose of this article is not to enter
the debate over (a) the wisdom of centralizing the fractional
reserve system and the issuance of money (see Bagehot
1873) or (b) the causality between government price stabilization policies (or lack thereof) and the resulting health of
financial intermediaries (see Schwartz 1988). Rather, the
primary concern is with the appropriate use of government
intervention once a period of financial stress occurs, whatever its underlying source.
8. The closing is inevitable absent the existence of a deposit
insurer that will cover the losses for at least some subset of
the firm's creditors.
9. See Tallman and Moen (1990).
10. Technically, the model by Kanatas deals only with riskneutral firms and risk-averse private lenders. Therefore,
lending between private firms is not explicitly considered.
However, the model could be extended to incorporate the
risk aversion discussed in the text.
11. Wall and Peterson (1990) provide a review of the failure of
Continental Illinois and its impact on other allegedly financially weak banking organizations.
12. FDIC insurance assessments are ordinarily levied only on
domestic deposits.
5. Wigmore (1987) provides an interesting argument that neither the timing of the bank holiday nor the calm afterwards
was due primarily to domestic failures on the part of the
Federal Reserve and the subsequent establishment of the
FDIC. He argues that the holiday was necessitated by a run
References
Bagehot, Walter. Lombard Street: A Description of the Money
Market. Philadelphia, Penn.: Orion Editions, 1873.
Bernanke, Ben S. "Nonmonetary Effects of the Financial Crisis
in the Propagation of the Great Depression." American Economic Review 73 (June 1983): 257-76.
Berry, John M. "Can We Do without the 'Too-Big-To-Fail'
Doctrine?" Financier, April 1991, 1-8.
10
Economic
Review
Bikhchandani, Sushil, David Hirshleifer, and Ivo Welch. "A
Theory of Fads, Fashion, Custom, and Cultural Change as
Information Cascades." University of California—Los Angeles AGSM Working Paper 20-90, February 13, 1991.
Bordo, Michael D. "The Lender of Last Resort: Alternative Views
and Historical Experience." Federal Reserve Bank of Richmond Economic Review 76 (January/February 1990): 18-29.
January/February 1992
Calomiris, Charles W., and Charles M. Kahn. "The Role of Demandable Debt in Structuring Optimal Banking Arrangements." American Economic Review 81 (June 1991):
497-513.
Chcrnow, Ron. The House of Morgan: An American Banking
Dynasty and the Rise of Modern Finance. New York, N.Y.:
Simon and Schuster, 1991.
Corrigan, E. Gerald. "A Perspective on Recent Financial Disruptions." Federal Reserve Bank of New York Quarterly
Review 14 (Winter 1989/90): 8-15.
Emmons, William. "Deposit Insurance and Last-Resort Lending as Delegated Monitoring: A Theory of Banking 'Safety
Nets.' " Unpublished paper, Northwestern University,
November 1991.
Flannery, Mark J. "Debt Maturity and the Deadweight Cost of
Leverage: Optimally Financing Banking Finns." University
of Florida Working Paper, July 1991.
Golembe Reports. Too-Big-To-Fail and All That. Washington,
D.C.: CHG Consulting, Inc., May 21, 1991.
Kanatas, George. "Deposit Insurance and the Discount Window: Pricing under Asymmetric Information." Journal of
Finance 41 (June 1986): 437-50.
Reserve Bank of Atlanta
DigitizedFederal
for FRASER
Kaufman, George G. "The Truth about Bank Runs." In The Financial Services Revolution: Policy Directions for the Future, edited by Catherine England and Thomas Huertas,
9-40. Norwell, Mass.: Kluwer Academic Publishers, 1987.
. "Lender of Last Resort: A Contemporary Perspective."
Journal of Financial Services Research 5 (October 1991):
95-110.
Schwartz, Anna J. "The Lender of Last Resort and the Federal
Safety Net." Journal of Financial Services Research 1 (January 1988): 1-18.
Tallman, Ellis W„ and Jon R. Moen. "Lessons from the Panic
of 1907." Federal Reserve Bank of Atlanta Economic Review 75 (May/June 1990): 2-13.
Wall, Larry D., and David R. Peterson. "The Effect of Continental Illinois' Failure on the Financial Performance of Other Banks." Journal of Monetary Economics 26 (August
1990): 77-99.
Wigmore, Barrie A. "Was the Bank Holiday of 1933 Caused by
a Run on the Dollar?" Journal of Economic History 48
(September 1987): 739-55.
Economic
Review
11
W h a t Hath the Fed
Wrought? Interest Rate
Smoothing in Theory
and Practice
William Roberds
h e study of monetary policy has traditionally been hobbled by a
yawning gap between theory and practice. Research economists
who study monetary policy are usually concerned with the behavior of one or m o r e aggregate measures of m o n e y — M 1 , M2, and so
forth. Participants in financial markets, on the other hand, view
monetary policy as primarily affecting interest rates, especially at the short
end of the term structure. Financial market press and wire accounts of monetary policy emphasize its role in reducing short-term interest rate fluctuations. T h i s aspect of m o n e t a r y policy, o f t e n r e f e r r e d to as interest rate
"smoothing," has traditionally been either ignored or criticized by academic
e c o n o m i s t s . D e s p i t e being u n p o p u l a r in a c a d e m i c circles, interest rate
smoothing is nonetheless practiced to a greater or lesser extent by the monetary authorities of every m a j o r industrialized nation and has been practiced
by the Federal Reserve System since its inception. 1
The continued and widespread use of interest rate smoothing as a proximate goal for monetary policy poses a number of questions for economic research. Three of the most important questions are as follows. The first and
most obvious question is, Why smooth interest rates in the first place, in
spite of all academic advice to the contrary? Second, What are the long-term
consequences of interest rate smoothing f o r the e c o n o m y ? And the third
question is, Provided that interest rate smoothing is a desirable objective, to
what extent should interest rate smoothing take precedence over other goals
of monetary policy?
The author is a research
officer and senior economist
in the macropolicy section of
the Atlanta Fed's research
department.
Economic
12
Review
These questions are difficult ones, but they are obviously crucial to the
m a n a g e m e n t of the nation's monetary policy. T h e y remain by and large
unanswered. Still, researchers have made a considerable amount of progress
on these questions in recent years. The following discussion reviews some
January/February 1992
of the most important contributions by economists to
the understanding of these c o m p l e x issues. Because
so much of the discussion of interest rate smoothing
involves historical c o m p a r i s o n s , it is appropriate to
begin with a brief history of the Federal Reserve System's approaches to open market policy, with particular emphasis on the impact of Fed policy on short-term
fluctuations in interest rates.
i n t e r e s t Rate Smoothing and Federal
Reserve Policy: Historical Summary 2
For the purposes of this article, it is assumed that
the principal i n s t r u m e n t of Federal R e s e r v e policy
has been and will continue to be a short-term interest
rate such as the federal f u n d s rate. 3 T h e term policy
" i n s t r u m e n t " refers to the Federal Reserve S y s t e m ' s
practice of a t t e m p t i n g to i n f l u e n c e s o m e e c o n o m i c
variable, such as the fed f u n d s rate, in an attempt to
achieve its policy goals. The Federal Reserve's use of
the fed f u n d s rate as a policy instrument is well k n o w n
and without controversy. What has been less well understood by academic economists, until recently, is the
extent to which Fed policies have traditionally worked
to s m o o t h f l u c t u a t i o n s in s h o r t - t e r m r a t e s . In t h e
1980s, however, a n u m b e r of studies pointed out that
1914, the year the Federal Reserve System was founded, represents s o m e t h i n g of a w a t e r s h e d in the f i n a n c i a l history of the United States, if not the world. 4
A number of changes in the patterns of financial market data are evident after that point. Chief among these
are the disappearance of seasonal fluctuations in most
short-term interest rates, an increase in seasonal fluctuations in the money supply, and the general increase
in the tendency of rates to persist at or near their current levels over time.
How many of the changes in financial markets after 1914 can be attributed to the policies of the Fed
and h o w many are owing to other circumstances is a
matter of continuing debate. It is undeniable, however, that Fed policies did substantially c o n t r i b u t e to
changes in U.S. and world financial markets. A plausible explanation for the impact of the Fed's monetary
p o l i c i e s is c l e a r l y laid out by M a r v i n G o o d f r i e n d
(1988), e x p a n d i n g on the earlier analysis of Milton
Friedman and Anna J. Schwartz (1963). In the early
years of its existence, the Fed found itself in a historically unique position because of its statutory role as a
p u b l i c a g e n c y c h a r g e d with m a n a g i n g the n a t i o n ' s
money supply and also because a large domestic stock
Digitized Federal
for FRASER
Reserve Bank of Atlanta
of gold reserves had built up following the outbreak of
World War I in Europe. T h e fact that the United States
was on the gold standard at the time meant that monetary policy actions had to b e backed by a sufficient
gold reserve to maintain the gold standard. Such reserves were readily available to the United States as
the w a r r i n g E u r o p e a n n a t i o n s a b a n d o n e d the g o l d
standard and started using their o w n gold reserves to
pay for their military operations. Also contributing to
the efficacy of the Fed's actions was its statutory role
as a stabilizing influence on the banking system rather
than as a profit-maximizing entity. Fed officials had a
greater degree of latitude in their attempts to stabilize
financial markets than "central banks" had enjoyed before 1914, when policy-making was often constrained
by the need to safeguard the b a n k s ' (private) shareholders' profits or by an insufficient gold reserve?
In the 1920s the primary focus of the F e d ' s monetary p o l i c y c h a n g e d f r o m the d i s c o u n t w i n d o w to
open market operations, where it has remained ever
since. 6 T h e historical record suggests that this structural change in monetary policy did not result in any
lessening of the Fed's taste for interest rate smoothing.
There was little need for the Fed to smooth interest rates
during the years immediately following the banking crisis of 1933, as rates on short-term government securities
remained at extremely low levels. From April 1942 until
March 1951, Federal Reserve open market policy was
very tightly constrained by the need to finance wartime
expenditures. Over this period, interest rates on shortterm g o v e r n m e n t securities were " p e g g e d " at levels
agreed to by the Treasury Department.
With the signing of the Treasury-Federal Reserve
Accord in March 1951 the Fed regained the ability to
initiate monetary policy. T h e data records of the 1950s
and 1960s again suggest that interest rate smoothing
continued to be an important goal of monetary policy.
In the m e a n t i m e , the d e v e l o p m e n t of the F e d e r a l
Funds Market led to an ever greater amount of emphasis being placed on the fed f u n d s rate as a measure of
credit m a r k e t c o n d i t i o n s . B y the late 1960s m o v e ments in the fed funds rate were frequently mentioned
in the minutes of the Federal Open Market Committee
( F O M C ) . A s open market operations were increasingly concentrated in the fed funds market, stabilization
of day-to-day movements in the funds rate b e c a m e an
important short-term goal of open market policy. In
the mid-1970s relatively narrow ranges for the f u n d s
rate were given in the F O M C directive. Over the latter
part of the 1970s the width of the "intermeeting" (between F O M C meetings) range for the fed f u n d s rate
shrank from around 1 percentage point to as little as
Economic Review
13
1/4 percentage point. As emphasized by A n n - M a r i e
Meulendyke (1990), this extreme concern with shortterm fluctuations in the f u n d s rate developed gradually
over time, not as a result of an active decision to target
the funds rate tightly.
T h e period f r o m O c t o b e r 1979 through O c t o b e r
1982 is officially known as the "nonborrowed reservestargeting" (NBR) period and unofficially as "the Fed's
m o n e t a r i s t e x p e r i m e n t . " D u r i n g this time, n o n b o r rowed reserves officially replaced the fed f u n d s rate
as the short-run target for open m a r k e t o p e r a t i o n s .
W h e t h e r this m o v e was monetarist or not, its effect
was to allow m u c h m o r e variation in the fed f u n d s
rate than had occurred under the funds rate-targeting
regime. The variability of reserves during the N B R targeting period did not change much from the funds
rate-targeting period, c o n f o u n d i n g some predictions
that loosening the peg on fed funds would reduce the
average magnitude of fluctuations in bank reserves.
(See Charts 1A and I B and 2A and 2B.) The continued volatility in reserves during the N B R period reflects at least in part the turbulent economic conditions
in that era. H o w e v e r , at least s o m e p o r t i o n of the
fluctuations in reserves during the NBR period can be
attributed to the continued use of the funds rate as an operating target. A recent study by Timothy Cook (1989)
found that despite the nominal adoption of N B R targeting, two-thirds of the variation in the f u n d s rate
during the October 1979 through October 1982 period
can be attributed to direct policy actions on the part of
the Fed; that is, these m o v e m e n t s in the f u n d s rate
were not necessary to meet the reserves target. If this
estimate is even approximately correct, then the distinction between the open market policies in the NBR
and fed f u n d s rate-targeting periods is perhaps best
characterized as quantitative rather than qualitative.
The borrowed-reserve operating procedure that has
been in place since October 1982 has been seen by
many academic observers and financial market participants as a retreat toward the funds rate-targeting proced u r e of the 1970s. A l t h o u g h the c u r r e n t o p e r a t i n g
procedure incorporates some important reforms, it is
difficult to find fault with this generalization. Comparing the time series data on the fed funds rate and bank
reserves since 1982 (Chart 3) with the data record of the
1976-82 period (Charts 1 and 2) shows that the current
operating procedure falls somewhere between the two
earlier operating procedures in terms of the variability
of interest rates. In particular, Chart 3A indicates that
except for an occasional well-publicized miss, the funds
rate has tended to move within fairly close bounds since
October 1982, particularly since 1986.
14
Economic Review
Why Smooth Interest Rates?
In discussing the desirability of interest rate smoothing, it is necessary to specify carefully the horizon over
w h i c h the relevant interest rate is to be s m o o t h e d .
There is widespread (though not universal) agreement
a m o n g economists that a monetary authority cannot
s u c c e s s f u l l y i n f l u e n c e the real or i n f l a t i o n - a d j u s t e d
rate of interest over the long run. 7 In other words, according to the mainstream view, the Fed does not have
the option of simply pegging the interest rate at some
level d e e m e d to be desirable. At the other extreme,
most economists would agree that it is feasible for the
Fed to exercise considerable influence over at least
n o m i n a l interest rates on a d a y - t o - d a y or w e e k - t o week basis. During the late 1970s, for example, shortterm movements in the fed f u n d s rate were virtually
eliminated by Fed interventions. Consequently, the debate over the "short-run" smoothing of interest rates is
centered to some extent on what should be designated
as the short run.
The Fed's very strict smoothing of day-to-day movements in the funds rate in the late 1970s generated an
unprecedented amount of attention to Fed operating
procedures by the economics profession, much of it
critical in tone. T h e logical basis for a good deal of
this criticism was laid in an article by William Poole
(1970). Analyzing open market operations from an informational perspective, Poole showed that strict targeting of interest rates, even in the very short run,
would result in what economists call an "identification" problem. That is, strict targeting of interest rates
necessarily entails some loss of information. This loss
of information occurs because successful open market
operations require some knowledge of the banking ind u s t r y ' s d e m a n d schedule for reserves: specifically,
what quantity of reserves would be demanded at various rates of interest. By keeping interest rates c o n stant, even over some short interval, some knowledge
of b a n k s ' d e m a n d schedule will necessarily be forgone. Thus, to smooth interest rates effectively, some
variation in interest rates is desirable, even in the short
run. In everyday terms, Poole's argument would say
that it is easier to steer a car by m a k i n g a series of
small probes and corrections than to keep the same vehicle headed arrow-straight into the great unknown.
Poole's theoretical observations were borne out in
practice. In the late 1970s very narrow targeting of the
funds rate in the short run led to a general hesitancy to
m o v e the f u n d s target. Moves in the funds target were
generally anticipated by financial markets and were
January/February 1992
Chart 1A
V a r i a b i l i t y of Fed F u n d s R a t e
(federal funds rate targeting)
January 1976 to September 1979
(biweekly observations)
C h a r t 1B
V a r i a b i l i t y in R e s e r v e s G r o w t h
Percent
(federal funds rate targeting)
8
6
4
2
0
- 2
-4
- 6
January 1976 to September 1979
(weekly observations)
Source: All charts calculated by the author using data from the Board of Governors of the Federal Reserve System.
Digitized Federal
for FRASER
Reserve B a n k of Atlanta
Economic
Review
15
Chart 2 A
V a r i a b i l i t y o f Fed F u n d s R a t e
Percent
(nonborrowed
reserves
targeting)
3
2
1
0
- 1
- 2
-3
October 1979 to September 1982
(biweekly observations)
Chart 2B
V a r i a b i l i t y in R e s e r v e s G r o w t h
Percent
(nonborrowed
reserves
targeting)
- 6 -
_
8
1
+
1979
Economic
16
1980
J.
1981
1982
October 1979 to September 1982
(weekly observations)
Review
J a n u a r y / F e b r u a r y 1992
Chart 3A
Variability of Fed Funds Rate
October 1982 to July 1991
(biweekly observations)
Chart 3B
Variability in Reserves Growth
Percent
(borrowed
reserves
targeting)
8
6
4
2
0
- 2
-4
- 6
- 8
1982
1983
1984
Reserve Bank of Atlanta
Digitized Federal
for FRASER
1985
1986
1987
1988
October 1982 to July 1991
(weekly observations)
1989
1990
Economic Review
17
often seen both by market participants and academic
observers as too little, too late. T h e consensus view of
the f u n d s rate t a r g e t i n g of the late 1970s is aptly
s u m m a r i z e d by M e u l e n d y k e (1990), w h o observed
that during this period the " [ f e d f u n d s ] rate m o v e s
during the week were so limited that they provided
little or no information about reserve availability or
market forces."
Following the interest rate volatility associated with
the nonborrowed reserves targeting, Fed policy since
1982 has emphasized the stabilization of the fed funds
rate in the short run. This renewed attention to shortterm interest rate stabilization continues to be a subject
of academic controversy. 8 However, potentially important d i f f e r e n c e s exist between the current operating
procedure and that followed in the late 1970s. The first
difference is that the nominal short-run target is borrowed reserves rather than the funds rate per se. The
second is that interventions by the Fed in the f u n d s
market are limited to one per day. The third important
difference is that since early 1984 bank reserves have
to b e maintained with a two-day rather than a twoweek lag. The net effect of these changes has been to
introduce a s o m e w h a t greater amount of day-to-day
variation in the f u n d s rate, particularly on alternate
Wednesdays when reserve balances are settled.
To date, r e s e a r c h e r s h a v e not f u l l y r e s o l v e d the
e x t e n t to w h i c h t h e s e c h a n g e s h a v e h e l p e d a v o i d
some of the difficulties associated with 1970s operating procedures. Bennett T. M c C a l l u m and James G.
Hoehn (1983) provide a theoretical analysis of targeting interest rates versus total reserves under lagged
and c o n t e m p o r a n e o u s r e s e r v e a c c o u n t i n g e n v i r o n ments. They find that under contemporaneous reserve
accounting, targets o n total reserves p e r f o r m better
than interest rate targets in terms of hitting an overall
monetary target. More elaborate versions of M c C a l lum and Hoehn's model, which are capable of distinguishing between borrowed and n o n b o r r o w e d
reserves, are presented by Michael Dotsey (1989) and
David D. Van Hoose (1988). Dotsey also presents simulations suggesting that there is little quantitative distinction between interest rate and borrowed reserves
t a r g e t i n g . T h i s c o n c l u s i o n is c o n s i s t e n t w i t h the
econometric analysis of Daniel P. Thornton (1988).
A number of recent studies have chosen to ignore
the question of what the Fed does or does not do on a
day-to-day basis, focusing instead on the effects of interest rate smoothing over the seasonal horizon. Jeffrey A. Miron (1986), in particular, calls attention to
the changes in seasonal behavior of interest rates and
the monetary aggregates after the founding of the Fed.
18
Economic Review
Miron notes that late nineteenth- and early twentiethcentury money crises almost always occurred during
the fall harvest season. A c c o r d i n g to M i r o n , in the
years immediately following its founding the Fed was
able to prevent the recurrence of such crises by lessening seasonal m o v e m e n t s in interest rates and seasonally expanding the monetary base. Miron's work is
e x t e n d e d in N. G r e g o r y M a n k i w and Miron (1986)
and Mankiw, Miron, and David N. Weil (1987). 9
Historical accounts such as that found in Friedman
and Schwartz (1963) indicate that the lessening of seasonal strains in the money market was one of the most
important, and perhaps the overriding, goal of early
Fed policy. There seems to be widespread professional
agreement, both then and now, that in the early twentieth century some amount of seasonal accommodation
was needed to offset the strains put on the financial
system by the seasonal cycle of agriculture. In recent
years the pace of the U.S. macroeconomy has retained
its seasonal character, but the available evidence suggests that the primary source of this seasonality is now
the Christmas shopping season, not the agricultural
cycle. 1 0 In other words, instead of seasonality induced
by a shock to aggregate supply (that is, the fall harvest), seasonality in the macroeconomy is now driven
by a seasonal shock to aggregate demand (Christmas
shopping). The absence of seasonal movements in interest rates, together with seasonal movements in various monetary aggregates, indicates that smoothing of
seasonal fluctuations continues to be an important focus of Fed policy.
Given the demand-driven nature of seasonal fluctuations in the U.S. economy, a natural question to ask is
whether the smoothing of seasonal shocks to output
should be an important focus of U.S. monetary policy.
A recent study by Mankiw and Miron (1991) argues
that the answer to this question would be yes if the
economy behaves in a sufficiently "Keynesian" fashion—that is, if real quantities such as output and employment are sensitive enough to changes in purely
nominal quantities. Mankiw and Miron estimate that
n o n a c c o m m o d a t i o n of seasonal d e m a n d s for m o n e y
could result in relatively large welfare losses, arising
from seasonal movements in interest rates. Taking the
opposite point of view, McCallum (1991) argues that
stabilization of the economy over the year might not
be inherently desirable, citing the examples of Christmas shopping and the seasonal construction cycle as
two "rational" causes for seasonality. McCallum concludes that the welfare gains to the seasonal smoothing of interest rates are probably less than $1 per U.S.
resident per year.
January/February 1992
In summary, a search of the literature on the possible
rationales for interest rate smoothing raises more questions than it answers. The bulk of the studies appear to
weigh in against very strict short-term smoothing of
interest rates in f a v o r of s o m e sort of reserve-based
operating procedure, but it would be a mistake to characterize this issue as settled. Nor is there anything app r o a c h i n g a p r o f e s s i o n a l c o n s e n s u s c o n c e r n i n g the
issue of the degree to which the Fed should accommodate seasonal pressures in the money market.
What Are the Long-Term Consequences
of Interest Rate Smoothing?
T h e classic criticism of interest rate pegging was
a d v a n c e d by Knut Wicksell in 1898. W i c k s e l l reasoned that if a central bank attempted to peg interest
rates below their natural equilibrium level, over the
long run larger and larger infusions of cash and reserves would be needed to satisfy the resulting excess
demands for money. Any such attempt to peg rates below their natural equilibrium was therefore destabilizing and ultimately doomed to failure.
Wicksell's argument against interest rate pegging
(which is still found in economics textbooks) was concerned with the long-run consequences of a sustained
policy. Partly as a consequence of the Fed's increased
attention to interest rate objectives during the 1970s,
n u m e r o u s articles a p p e a r e d in professional j o u r n a l s
that explored the long-term consequences of smoothing interest rates, even when the smoothing is done
over the very short run.
One of the most influential studies was a theoretical
analysis by Thomas J. Sargent and Neil Wallace (1975).
Sargent and Wallace were able to breathe new life into
Wicksell's argument by introducing the idea of rational expectations into a model similar to that used by
Poole (1970). Specifically, Sargent and Wallace
demonstrated that the potentially destabilizing indeterminacy arises when a monetary authority such as the
Fed attempts to stabilize interest rates, even over the
very short term. For an interest rate target to be credible, it has to be backed by the willingness to defend
this target via open market operations. The degree of
intervention necessary to maintain the interest rate target, however, will depend, a m o n g other things, on the
rate of inflation anticipated by the public over the interval that the interest rate is to be stabilized. T h e indeterminacy arises f r o m the fact that essentially any
expected rate of inflation is consistent with an interest
Reserve Bank of Atlanta
DigitizedFederal
for FRASER
rate target, given a sufficient willingness on the part of
the Fed to defend the targeted rate of interest.
McCallum (1981) reconsiders the Sargent-Wallace
indeterminacy and finds that it can be resolved by the
use of a "nominal a n c h o r " — f o r example, a monetary
target—on the part of the Fed. That is, if policy is adjusted not only to smooth interest rates but also in res p o n s e to c h a n g e s in s o m e o t h e r n o m i n a l t a r g e t
variable, then this type of policy does a better job of
signaling the Fed's intentions. This result occurs because the response to movements in a nominal anchor
gives the public a more definite idea concerning how
staunchly a given interest rate target will be defended.
To return momentarily to the automotive analogy, the
use of a nominal anchor might be compared to a driver's adherence to the rules of the road. By driving on
the right side of the road, stopping at stop signs, and
so forth, individual drivers reduce the potential scope
of their own actions. Yet this reduced freedom of action p r o d u c e s the desirable o u t c o m e of f e w e r collisions because better information about each individual
driver's likely behavior is available to other drivers.
M c C a l l u m ' s results provide some intellectual
b a c k i n g for the c o m b i n a t i o n of s h o r t - t e r m interest
rate s m o o t h i n g and longer-term m o n e t a r y targeting
now in use by the Fed. In practical terms, however,
the u s e f u l n e s s of m o n e t a r y t a r g e t i n g s c h e m e s h a s
been limited by uncertainty about w h i c h m e a s u r e d
m o n e t a r y a g g r e g a t e is the a p p r o p r i a t e m e a s u r e of
" t h e " m o n e y s u p p l y . " Despite these practical d i f f i culties, and despite the usual amount of initial acad e m i c s k e p t i c i s m , t h e p o l i c y c o n c l u s i o n s of t h e
Sargent-Wallace-McCallum a p p r o a c h h a v e c o m e to
be v i e w e d as orthodoxy by a broad cross section of
the e c o n o m i c s p r o f e s s i o n , even as their theoretical
analyses have become seen as somewhat dated.
Numerous refinements have been suggested for the
basic approach used by Sargent and Wallace (1975)
and McCallum (1981). While these refinements have
o f f e r e d some useful insights into the possible longterm consequences of various approaches to monetary
policy, no single point of view on this subject has been
widely accepted within the economics profession. The
waters around this issue have been further muddied by
the publication of another controversial article by Sargent and Wallace in 1982. In the context of a theoretical model Sargent and Wallace were able to show that
u n d e r certain a s s u m p t i o n s , the " b e s t " interest rate
smoothing policy was not adjusted with reference to a
nominal anchor such as a monetary aggregate but according to the needs of people in the model who are
e n g a g e d in trade. 1 2 T h i s c o n c l u s i o n , though widely
Economic Review
19
discussed within the e c o n o m i c s p r o f e s s i o n , has not
been as widely accepted. Dissenting viewpoints can
b e found, for example, in articles by David Laidler
(1984) and McCallum (1986), who view the Sargent
and Wallace (1982) result as lacking in sufficient generality to be applicable to the problems of real-world
monetary policy. 13
D e s p i t e the apparent d i v e r g e n c e of p r o f e s s i o n a l
opinion on the subject, the literature on the long-term
stability of interest rate smoothing policies has made
substantial progress over recent decades. The seminal
article by Sargent and Wallace (1975) provided a useful insight into the possibility that even short-term interest rate smoothing, without reference to some other
policy guidelines, could be destabilizing over the long
term. T h e recent literature in this area has tended to
focus not on whether such guidelines (or "nominal anchors") are needed but on which anchor would provide for the smoothest sailing.
To What Extent Should Interest Rate
Smoothing Take Precedence over
Other Goals of Monetary Policy?
Certainly, m o n e t a r y a u t h o r i t i e s s u c h as the F e d
were not created for the sole purpose of smoothing interest rates. This fact is recognized by researchers in
this area, in that the objective of the Fed is traditionally modeled as incorporating "stable" prices and econ o m i c g r o w t h . S t a b i l i t y m e a n s d i f f e r e n t t h i n g s to
different researchers, but it is traditionally taken to
mean "as close as possible to some desired path."
According to the various models that follow in the
tradition of Sargent and Wallace (1975) (for example,
McCallum 1981; Dotsey 1989; and Goodfriend 1987),
monetary policy actions that smooth interest rates are
essentially in direct conflict with the goals of price and
output stability. In these models, interest rate smoothing leads to increased confusion over real versus nominal quantities and hence to greater uncertainty in both
real output and prices. Thus, the analysis in these articles suggests that the practice of interest rate smoothing will eventually be at odds with other traditional
stabilization objectives.
An alternative perspective on the long-term effects
of interest rate smoothing is presented in an article by
W a l l a c e ( 1 9 8 4 ) . W a l l a c e a r g u e s that d e b a t e s o v e r
monetary policy should not be centered on questions
of stability per se but rather on questions of w h o benefits and who loses, and by how m u c h , as a result of
20
Economic
Review
monetary actions. For example, if open market operations by the Fed are able to lower the real rate of interest permanently, as occurs in Sargent and Wallace's
(1982) model, then holders of nominally denominated
assets (that is, fixed-rate debt) are made worse off, and
people who are borrowing money are better off than
they would be without this open market intervention.
In an example Wallace (1984) uses, it is impossible to
say that the e c o n o m y as a whole is better off, either
with or without the presence of monetary policies designed to influence interest rates. Under each policy
scenario, s o m e p e o p l e are adversely a f f e c t e d while
others are better off.
For many researchers it would be difficult to evaluate the real-world s i g n i f i c a n c e of this a r g u m e n t . If
monetary policy can substantially influence real interest rates over time, as Wallace (1984) argues it can,
then such an argument is almost surely correct. However, as noted above, many economists do not believe
that real interest rates can be influenced by monetary
policy, especially over the long run. For this reason,
those economists would probably dispute the direct
applicability of Wallace's argument. However, it is a
useful exercise to apply a similar line of reasoning to
what is k n o w n about the dynamics of financial markets before and after the founding of the Fed.
From the end of the Civil War to 1914 movements
in interest rates were characterized by sudden, large,
and relatively short-lived swings, as c o m p a r e d with
movements in interest rates since 1914. Movements in
the very short end of the term structure, as typified by
the rate on overnight call money, were particularly subject to large fluctuations. Goodfriend (1991) notes that
on a m o n t h l y a v e r a g e basis o v e r n i g h t m o n e y rates
jumped by more than 5 percentage points on twentysix occasions and changed by more than 10 percentage
points eight times during this forty-nine-year period. 14
There were clearly winners (lenders) and losers (borrowers) who gained and lost money during these brief
episodes of high interest rates, but high levels of interest rates were hardly seen by contemporary observers as
the m o s t n e g a t i v e aspect of these " m o n e y p a n i c s . "
More generally, such episodes were associated (most
often in a causal fashion) with bank runs, contractions
in credit, and a reduction in the overall level of real economic activity. A potentially strong argument in favor
of interest rate smoothing as a goal of monetary policy
would be that smoothing prevents the occurrence of
such crises, given that such crises are, after the fact, in
no one's best interest. Thus, if one accepts the idea that
interest rate smoothing leads to an avoidance of money
panics, with few other serious long-term consequences
January/February 1992
on the economy, then the argument advanced by Wallace (1984) would lose much of its force.
The importance of central bank liquidity provision
in preventing financial panics is an accepted part of
central banking doctrine, dating back to at least Walter
Bagehot (1873). A s was noted above, the elimination
of m o n e y panics associated with the harvest season
has been viewed by many observers as an important
a c c o m p l i s h m e n t of early Fed policy. M o r e recently,
h o w e v e r , e c o n o m i s t s h a v e been revising traditional
views on the causes and remedies of m o n e y panics. In
the recent literature the tendency has been to view preFed money panics as well as the Great Depression of
the 1930s not as exercises in pure mass hysteria but as
being caused by the markets' rational responses to a
poorly designed regulatory structure. 15 In a widely cited paper, Douglas W. Diamond and Philip H. Dybvig
(1983) use such an argument to rationalize deposit insurance as a means of avoiding panics. In a related paper
V.V. Chari (1989) argues that under certain conditions
monetary policy may be superior to deposit insurance
for this purpose because deposit insurance carries with
it the potential for skewing banks' incentives toward
risky investments. On the other side of the question
Wallace (1988, 1990) argues, in effect, that temporary
suspensions of convertibility such as occurred often
during the money panics could be desirable as a means
of eliciting otherwise unavailable information about
the e c o n o m y ' s demand for liquid assets.
An unfortunate aspect of the new literature on panics is that it has generally yielded only limited insights
into the appropriate role of monetary policy in the prevention of panics. Overall there has been a tendency to
focus on the consequences of various regulatory constraints (for example, restrictions on banks' assets, liabilities, reserves, and so forth), with relatively little
attention being paid to the interaction between direct
regulation, emergency credit provision, and open market policies. The implicit message (occasionally made
m o r e explicit in p a p e r s such as Ben S. B e r n a n k e ' s
1983 analysis of the Great Depression in the United
States) has been that the role of monetary policy in
preventing the recurrence of such crises must be subordinate to that of regulation. Whether or not this last
view is correct remains a matter of dispute (see, for
e x a m p l e , J a m e s D. H a m i l t o n 1987 for a s o m e w h a t
contrary view of the Great Depression). Still, the burgeoning literature that seeks to link the structures of fin a n c i a l i n t e r m e d i a t i o n to a g g r e g a t e c o n s e q u e n c e s
suggests that a good understanding of the potential effects of monetary policy depends on a deeper understanding of the workings of financial intermediaries. A
Reserve Bank of Atlanta
DigitizedFederal
for FRASER
number of recent papers—for example, John H. Boyd
and E d w a r d C. Prescott (1986), Bernanke and Mark
Gertler (1987), and Sudipto Bhattacharya and Douglas
Gale ( 1 9 8 7 ) — h a v e suggested that banks may represent a particularly efficient solution to the problem of
how to evaluate the creditworthiness of potential borrowers. If this conclusion is accepted, it is fairly easy
to rationalize the expenditure of public f u n d s to protect the b a n k i n g s y s t e m in m o m e n t s of illiquidity.
W h a t is lacking is a precisely d e f i n e d role for open
m a r k e t policy in this protective m i s s i o n , especially
o p e n m a r k e t p o l i c i e s that i n c o r p o r a t e interest rate
smoothing.
Conclusion
Real-world objectives of monetary policy have historically i n c o r p o r a t e d s o m e d e g r e e of interest rate
smoothing. Traditional economic analyses suggest
that, at best, very strict s m o o t h i n g of interest rates
through open market operations leads to some loss of
information about the d e m a n d for f u n d s provided by
the central bank through such operations. Since 1975
various branches of academic economics have brought
a number of different approaches to this problem.
O n e branch of the m a c r o e c o n o m i c s literature has
had very little good to say about interest rate smoothing. This research has offered no rational basis for the
practice of interest rate smoothing and has often concluded that strict smoothing of interest rates would be
destabilizing in the absence of other observable policy
guidelines ("nominal anchors"). Another common
conclusion of these studies is that interest rate smoothing is inherently in conflict with other stabilization objectives.
Another branch of the recent macroeconomics literature has a m u c h m o r e benign view of interest rate
smoothing. According to these studies, in some cases
it m a y be possible for even a sustained interest rate
peg to result in better monetary policy than if policy
were adjusted according to movements in a monetary
target. Interest rate s m o o t h i n g is seen as inherently
neither good nor bad, but as beneficial to some people
and injurious to others. Although this branch of the literature has been influential, it is fair to say that its
c o n c l u s i o n s h a v e not b e e n w i d e l y a c c e p t e d by the
mainstream of the profession.
Yet another stream of research has focused on the
c h a n g e s in s e a s o n a l p a t t e r n s that a c c o m p a n i e d the
f o u n d i n g of the F e d e r a l R e s e r v e S y s t e m in 1914.
Economic Review
21
T h e s e s t u d i e s h a v e p r o v i d e d interesting c o m p a r a t i v e
a n a l y s e s of the " s m o o t h " financial m a r k e t s c h a r a c t e r istic of the p o s t - 1 9 1 4 w o r l d v e r s u s the " n o i s y " p r e 1 9 1 4 m a r k e t s . T h e t h e m e u n d e r p i n n i n g m u c h of this
research s e e m s to be that p o s t - 1 9 1 4 seasonal s m o o t h ing of interest rates by the F e d c o u l d b e j u s t i f i e d bec a u s e of t h e a s s o c i a t i o n of p r e - 1 9 1 4 m o n e y p a n i c s
w i t h the a g r i c u l t u r a l p r o d u c t i o n c y c l e . H o w e v e r , to
date these studies h a v e not p r o v i d e d a n entirely satisf a c t o r y e x p l a n a t i o n of t h e p r e - 1 9 1 4 p a n i c s . N o r h a s
this b r a n c h of the literature built u p m u c h of a c a s e as
t o w h y s e a s o n a l i n t e r e s t r a t e s m o o t h i n g is n e e d e d
w h e n seasonal f l u c t u a t i o n s are d r i v e n b y s h o c k s to dem a n d rather than supply.
A n o t e w o r t h y c o u n t e r p o i n t t o the l i t e r a t u r e m e n tioned a b o v e h a s b e e n p r o v i d e d by n e w research in the
g e n e r a l area of b a n k i n g , a n d particularly in the area of
m o d e l i n g " r u n s " o r " p a n i c s . " W h i l e not directly f o c u s i n g on the i s s u e o f i n t e r e s t r a t e s m o o t h i n g , this r e -
s e a r c h h a s the p o t e n t i a l to p r o v i d e a m o r e c o m p l e t e
u n d e r s t a n d i n g of t h e m o n e t a r y p o l i c y ' s i n t e r a c t i o n
with the regulatory a n d institutional structure of the fin a n c i a l s y s t e m . B y p r o v i d i n g a stable o p e r a t i n g e n v i ronment for the financial sector, m o n e t a r y policy
actions s u c h as interest rate s m o o t h i n g m a y serve as a
u s e f u l c o m p l e m e n t to m o r e direct r e g u l a t i o n . N o
d o u b t this idea will be e x p l o r e d in g r e a t e r detail as the
literature on financial i n t e r m e d i a t i o n d e v e l o p s .
A s with m a n y s u c h d e b a t e s in e c o n o m i c s , the c o n t r o v e r s y o v e r i n t e r e s t r a t e s m o o t h i n g is a l o n g w a y
f r o m b e i n g r e s o l v e d . It w o u l d be a m i s t a k e , h o w e v e r ,
to say that r e s e a r c h in this area h a s m a d e n o c o n t r i b u tion to k n o w l e d g e a b o u t the issue. To the contrary, the
research has b r o u g h t forth a n u m b e r of i n s i g h t f u l argum e n t s b o t h f o r a n d a g a i n s t t h i s a s p e c t of m o n e t a r y
policy. N o d o u b t there is s o m e d e g r e e of truth in all of
these a r g u m e n t s , a n d e a c h d e s e r v e s c a r e f u l c o n s i d e r a tion by p o l i c y m a k e r s .
Notes
1. On the prevalence of interest rate smoothing in the major
industrialized countries, see Batten et al. (1990), Kneeshaw
and Van den Bergh (1989), and Beaulieu and Miron
(1990). For information about the Federal Reserve System's practice of interest rate smoothing, see the historical
summary below, Goodfriend (1991), or Meulendyke
(1989).
2. Much of the summary below is drawn from Goodfriend
(1991), Meulendyke (1990), and Friedman and Schwartz
(1963).
3. The fed funds rate is the rate paid by banks on funds needed to meet the banks' reserve requirements. Transactions in
the fed funds market are generally limited to very short maturities, and many agreements only last a single night. On
the workings of the Federal Funds Market see Goodfriend
and Whelpley (1986).
4. These studies include papers by Barro (1989); Canova
(1988, 1991); Clark (1986); Mankiw and Miron (1986);
Miron (1986); and Mankiw, Miron, and Weil (1987).
5. The quasi-private nature of the major European central
banks prior to 1914 is recounted in the various essays in
Toniolo (1988). Earlier attempts at activist monetary policies by the U.S. Treasury were hindered by inadequate reserves, according to Timberlake (1978).
6. A helpful summary of the changeover from discount to
open market operations can be found in Chapter 2 of Meulendyke (1989). For a more extensive discussion of this
topic, see Wicker (1966). Both of these sources emphasize
that this change took place gradually during the 1920s and
not as a result of a single decisive change in policy. Evidently, the idea that bank reserves could be managed
22
Economic
Review
through open market operations was not widely understood
at this time, even within the Federal Reserve System. Consequently, many Federal Reserve policymakers in the
1920s viewed open market policy as being subordinate to
discount window operations.
7. At the heart of this belief is the idea that money is just an
accounting device for recording the values of goods and
service, and it is only the relative value between any two
goods that should matter in people's economic decisions.
For example, if next week the United States were to start
phasing in a new currency, the "shmoo," with two of the
new shmoos equal to a dollar, no one would seriously expect such a change to have any real effects. The belief that
monetary policy does not affect real rates over the long run
amounts to a belief that over the long run an increase in the
supply of money amounts to a sort of backhanded dollarshmoo substitution.
8. See, for example, Benjamin Friedman's (1988) comments
on current Fed operating procedures; also see Spindt and
Tarhan (1987).
9. Other important articles in this area differ somewhat in
their interpretation of the historical record. Clark (1986)
disputes the notion that the Fed was responsible for the
disappearance of interest rate seasonals. Arguing on the
basis of a number of statistical tests, Clark concludes that
the international scope of the 1914 changes in the financial
markets requires a more careful explanation than, say, that
advanced in Miron (1986). Two plausible explanations are
offered in Barsky et al. (1988) and Canova (1991).
10. See Barsky and Miron (1989), Beaulieu and Miron (1990),
and Braun and Evans (1991).
January/February 1992
11. See Roberds (1989) for a survey of the difficulties associated with measuring the money supply.
12. The latter point of view is generally associated with a concept known as the "real bills doctrine." The debate over the
validity of the real bills doctrine goes back well into the
nineteenth century, and even a cursory summary of this
debate would go beyond the scope of the present article.
The interested reader should consult The New Palgrave
(Eatwell, Milgate, and Newman 1987).
13. While a full recounting of the academic disagreements about
Sargent and Wallace (1982) goes beyond the scope of the
present article, the essence of the debate has to do with how
to best characterize money in an abstract setting. It turns out
that money is something that is extremely easy to spend but
difficult to characterize in an abstract sense. As a result, various branches of monetary theory have tended to emphasize
differing aspects of "moneyness." The approach used by Sargent and Wallace (1982) emphasizes money's role as a store
of value, so that buying and selling need not happen simultaneously. Other researchers, particularly McCallum (1986),
have found this characterization of money inappropriate.
These researchers have tended to emphasize the role of money as a medium of exchange—that is, its acceptability in
transactions among parties not well known to each other.
14. By way of contrast, under current operating procedures, a
month-to-month movement of 1 percentage point in the fed
funds rate would be considered a large move. As pointed
out in Miron (1986) and other papers summarized above,
interest rates prior to 1914 showed as distinctly seasonal
in the fall of the year, suggesting that even such large interest rate movements as described above were predictable to
some extent.
15. See Tallman (1988) for a nontechnical introduction to the
recent "bank panic" literature. Williamson (1987) and
Gertler (1988) also present useful surveys of this literature.
References
Bagehot, Walter. Lombard Street. London: H.S. King and
Company, 1873.
Barro, Robert. "Interest Rate Targeting." Journal of Monetary
Economics 23 (January 1989): 3-30.
Barsky, Robert B„ N. Gregory Mankiw, Jeffrey A. Miron, and
David N. Weill. "The Worldwide Change in the Behavior of
Interest Rates in 1914." European Economic Review (June
1988): 1123-54.
Barsky. Robert B„ and Jeffrey A. Miron. "The Seasonal Cycle
and the Business Cycle." Journal of Political Economy 97
(June 1989): 503-34.
Batten, Dallas S., Michael P. Blackwell, In-Su Kim, Simon E.
Nocera, and Yuzuru Ozeki. "The Conduct of Monetary Policy in the Major Industrial Countries: Instruments and Operating Procedures." International Monetary Fund Occasional
Paper 170, July 1990.
Beaulieu, J. Joseph, and Jeffrey A. Miron. "A Cross-Country
Comparison of Seasonal Cycles and Business Cycles." National Bureau of Economic Research Working Paper 3459,
October 1990.
Bernanke, Ben S. "Nonmonetary Effects of the Financial Crisis
in the Propagation of the Great Depression." American Economic Review 73 (June 1983): 257-76.
, and Mark Gertler. "Banking and Macroeconomic Equilibrium." In New Approaches to Monetary Economics, edited by William A. Barnett and Kenneth J. Singleton, 89-111.
Cambridge, U.K.: Cambridge University Press, 1987.
Bhattacharya, Sudipto, and Douglas Gale. "Preference Shocks,
Liquidity, and Central Bank Policy." Chap. 4 in New Approaches to Monetary Economics, edited by William A.
Barnett and Kenneth J. Singleton. Cambridge, U.K.: Cambridge University Press, 1987.
Boyd, John H., and Edward C. Prescott. "Financial Intermediary Coalitions." Journal of Economic Theory 38 (April
1986): 211-32.
DigitizedFederal
for FRASER
Reserve Bank of Atlanta
Braun, R. Anton, and Charles L. Evans. "Seasonal Solow
Residuals and Christmas: A Case for Labor Hoarding and
Increasing Returns." Federal Reserve Bank of Chicago
Working Paper 91-20, October 1991.
Canova, Fabio. "An Econometric Analysis of the Disappearance of Interest Rates Seasonals: Real Factors or the Fed?"
Brown University Working Paper 88-9, January 1988.
. "The Sources of Financial Crises: Pre- and Post-Fed Evidence." International Economic Review 32 (August 1991):
689-714.
Chari, V.V. "Banking without Deposit Insurance or Bank Panics: Lessons from a Model of the U.S. National Bank System." Federal Reserve Bank of Minneapolis Quarterly
Review 13 (Summer 1989): 3-19.
Clark, Truman A. "Interest Rate Seasonals and the Federal Reserve."
Journal of Political Economy 94 (February 1986): 76-125.
Cook, Timothy. "Determinants of the Federal Funds Rate:
1979-82." Federal Reserve Bank of Richmond Economic
Review 15 (January/February 1989): 3-19.
Diamond, Douglas W„ and Philip H. Dybvig. "Bank Runs, Deposit Insurance, and Liquidity." Journal of Political Economy 91 (June 1983): 401-19.
Dotsey, Michael. "Monetary Control under Alternative Operating Procedures." Journal of Money, Credit, and Banking 21
(August 1989): 273-90.
Eatwell, John, Murray Milgate, and Peter Newman, eds. The
New Palgrave: A Dictionary of Economics. 4 vols. New
York: The Stockton Press, 1987.
Friedman, Benjamin. "Lessons on Monetary Policy from the
1980s." Journal of Economic Perspectives 2 (Summer
1988): 51-72.
Friedman, Milton, and Anna J. Schwartz. A Monetary History
of the United States: 1867-1960. Princeton, N.J.: Princeton
University Press, 1963.
Economic
Review
23
Gertler, Mark. "Financial Structure and Aggregate Economic
Activity: An Overview." Journal of Money, Credit, and
Banking 20 (August 1988): 559-88.
Goodfriend, Marvin. "Interest Rate Smoothing and Price Level
Trend-Stationarity." Journal of Monetary Economics 19
(May 1987): 335-48.
. "Central Banking under the Gold Standard." CarnegieRochester Conference Series on Public Policy 29 (1988):
85-128.
. "Interest Rates and the Conduct of Monetary Policy."
Carnegie-Rochester
Series on Public Policy 34 (Spring
1991): 7-30.
, and William Whelpley. "Federal Funds." Chap. 2 in Instruments of the Money Market, 6th ed., edited by Timothy
Q. Cook and Timothy D. Rowe. Richmond, Va.: Federal
Reserve Bank of Richmond, 1986.
Hamilton, James D. "Monetary Factors in the Great Depression." Journal of Monetary Economics 19 (March 1987):
145-69.
Heller, H. Robert. "Implementing Monetary Policy." Federal
Reserve Bulletin 74 (July 1988): 419-29.
Kneeshaw, J.T., and P. Van den Bergh. "Changes in Central
Bank Money Market Operating Procedures in the 1980s."
Bank for International Settlements Economic Paper 23, January 1989.
Laidler, David. "Misconceptions about the Real Bills Doctrine:
A Comment." Journal of Political Economy 92 (February
1984): 149-55.
Mankiw, N. Gregory, and Jeffrey A. Miron. "The Changing
Behavior of the Term Structure of Interest Rates." Quarterly
Journal of Economics 101 (May 1986): 211-28.
. "Should the Fed Smooth Interest Rates? The Case of
Seasonal Monetary Policy." Carnegie-Rochester Series on
Public Policy 34 (Spring 1991): 41-70.
, and David N. Weil. "The Adjustment of Expectations to
a Change in Regime: A Study of the Founding of the Federal Reserve." American Economic Review 77 (June 1987):
358-74.
McCallum, Bennett T. "Price Level Determinacy with an Interest Rate Policy Rule and Rational Expectations." Journal of
Monetary Economics 8 (1981): 319-29.
. "Some Issues Concerning Interest Rate Pegging, Price
Level Determinacy, and the Real Bills Doctrine." Journal of
Monetary Economics 17 (1986): 135-60.
. "Seasonality and Monetary Policy: A Comment."
Carnegie-Rochester
Series on Public Policy 34 (Spring
1991): 71-76.
, and James G. Hoehn. "Instrument Choice for Money
Stock Control with Contemporaneous and Lagged Reserve
Requirements: A Note." Journal of Money, Credit, and
Banking 15 (February 1983): 96-101.
Meulendyke, Ann-Marie. U.S. Monetary Policy and Financial
Markets. New York: Federal Reserve Bank of New York,
1989.
. "A Review of Federal Reserve Policy Targets and Operating Guides in Recent Decades." In Intermediate Targets
24
Economic
Review
and Indicators for Monetary Policy: A Critical Survey.
New York: Federal Reserve Bank of New York, 1990.
Miron, Jeffrey A. "Financial Panics, the Seasonality of the
Nominal Interest Rate, and the Founding of the Fed. American Economic Review 76 (March 1986): 125-40.
Poole, William. "Optimal Choice of Monetary Policy Instruments in a Simple Stochastic Macro Model. Quarterly Journal of Economics 84 (May 1970): 197-216.
Roberds, William. "Money and the Economy: Puzzles from the
1980s' Experience." Federal Reserve Bank of Atlanta Economic Review 1A (September/October 1989): 20-35.
Sargent, Thomas J., and Neil Wallace. "Rational Expectations,
the Optimal Instrument, and the Optimal Money Supply
Rule." Journal of Political Economy 83 (April 1975): 241-54.
. "The Real-Bills Doctrine versus the Quantity Theory: A
Reconsideration." Journal of Political Economy 90 (December 1982): 1212-36.
Spindt, Paul A., and Vefa Tarhan. "The Federal Reserve's New
Operating Procedures: A Postmortem." Journal of Monetary
Economics 19 (January 1987): 107-23.
Tallman, Ellis. "Some Unanswered Questions about Bank Panics." Federal Reserve Bank of Atlanta Economic Review 13
(November/December 1988): 2-21.
Thornton, Daniel P. "The Borrowed-Reserves Operating Procedure: Theory and Evidence." Federal Reserve Bank of St.
Louis Review 70 (January/February 1988): 30-54.
Timberlake, Richard H. The Origins of Central Banking in the
United States. Cambridge, Mass.: Harvard University Press,
1978.
Toniolo, Gianni, ed. Central Banks' Independence in Historical Perspective. Berlin: Walter de Gruyter, 1988.
Van Hoose, David D. "Discount Rate Policy and Alternative
Federal Reserve Operating Procedures in a Rational Expectations Setting." Journal of Economics and Business 40
(November 1988): 285-94.
Wallace, Neil. "Some of the Choices for Monetary Policy."
Federal Reserve Bank of Minneapolis Quarterly Review 8
(Winter 1984): 15-24.
. "Another Attempt to Explain an Illiquid Banking System: The Diamond and Dybvig Model with Sequential
Service Taken Seriously." Federal Reserve Bank of Minneapolis Quarterly Review 12 (Fall 1988): 3-16.
. "A Banking Model in Which Partial Suspension is
Best." Federal Reserve Bank of Minneapolis Quarterly Review 14 (Fall 1990): 11-23.
Wicker, Elmus R. Federal Reserve Monetary Policy, 19171933. New York: Random House, 1966.
Wicksell, Knut. Geldzins und Guterpreise bestimmenden Ursachen. Jena: G. Fischer, 1898. Translated by R.F. Kahn as
Interest and Prices: A Study of the Causes Regulating the
Value of Money. London: Macmillan, 1936.
Williamson, Stephen D. "Recent Developments in Modelling
Financial Intermediation." Federal Reserve Bank of Minneapolis Quarterly Review 11 (Summer 1987): 19-29.
January/February 1992
FWE
Forecasting Industrial
Production: Purchasing
Managers' versus
Production-Worker
Hours Data
R. Mark Rogers
W
f
m
t
*
The author is forecast coordinator in the macropolicy section of the Atlanta Fed's
research department.
Federal
Reserve Bank of Atlanta
y ach month, various economy watchers—the financial markets in
i
particular—try to gauge the strength of the economy in order to
y
forecast orders, plan production, time investment, or price securij
ties. O n e key broad indicator of the e c o n o m y ' s strength is the index of industrial production (IP index), released by the Board of
Governors of the Federal Reserve around the fifteenth of each month. To
forecast industrial production, analysts u s e data available earlier in the
month. T w o indicators used for forecasting are the purchasing m a n a g e r s '
index and aggregate production-worker hours for manufacturing. Both are
broad measures of manufacturing activity released at the beginning of each
month.
The purchasing m a n a g e r s ' index is released by the National Association
of Purchasing M a n a g e m e n t ( N A P M ) on the first business day of the month
following the reference month. Data for aggregate production-worker hours
for manufacturing are released (generally) on the first Friday of each month
following the reference month as part of the U.S. Labor Department's B u reau of Labor Statistics employment report. These release dates m e a n that
the purchasing managers' index is available about two weeks before industrial production data are announced, and the production-worker hours series
leads the IP index by a period of ten days to two weeks.
Analysts have increasingly used these two series to anticipate the industrial production figure. This article focuses on why they are incorporated into c o n c u r r e n t f o r e c a s t i n g m o d e l s and h o w w e l l t h e y f o r e c a s t . It a l s o
e x a m i n e s s o m e of the r e a s o n s f o r f o r e c a s t e r r o r in t h e s e m o d e l s and
whether such a short forecast horizon is appropriate for judging the current
strength of manufacturing.
Economic Review
25
The basic finding is that while both series provide
some information on upcoming industrial production,
the L a b o r D e p a r t m e n t series of p r o d u c t i o n - w o r k e r
hours is able to project with significantly greater precision on a concurrent basis. Moreover, using initial estimates gives noticeably different results f r o m those
based on revised data; forecasters should view the purchasing m a n a g e m e n t data projections for industrial
production as tentative until confirmed by labor data.
71ie Theory behind the Forecasts
Before focusing on how well these competing m o d els forecast, it may be helpful to define these data series and look at the logic underlying the use of these
indicators to project industrial production.
The N a t i o n a l Association of P u r c h a s i n g M a n agement's Index. The purchasing m a n a g e r s ' index is
a g a u g e of the m a n u f a c t u r i n g s e c t o r ' s s t r e n g t h as
measured by a select group within the National Association of Purchasing Management. The index components and their respective weights (out of 100) are as
follows: new orders (30), production (25), supplier deliveries (15), inventories (10), and employment (20).'
Data c o m e f r o m q u e s t i o n s in the m o n t h l y N A P M
survey. 2
The survey is mailed to about 300 National Association of Purchasing Management members. The response rate varies each month, but it is not published.
The questionnaires are distributed by industry according to shares in total production, although response
rates affect how well the index sample is stratified. An
industry's low response rate (relative to those for other
industries) in a given month would result in a lower
weight for that month and a measurement error.
The survey tallies the responses as to whether conditions are (1) better, (2) worse, or (3) unchanged for
each of the c o m p o n e n t s . F r o m this information, the
N A P M c o m p u t e s a d i f f u s i o n i n d e x , i n d i c a t i n g the
" b r o a d n e s s " of w o r s e n i n g or i m p r o v i n g c o n d i t i o n s
a m o n g surveyed m e m b e r s . This approach is in contrast to other surveys (such as those conducted by the
U.S. C o m m e r c e Department's Bureau of the Census),
which measure actual levels by components for indicators such as new orders, inventories, and shipments.
By the National Association of Purchasing M a n agement's own definition, an overall index above fifty
indicates an e x p a n d i n g m a n u f a c t u r i n g sector, and a
number below fifty suggests a generalized contraction.
These conclusions are based on the component index-
26
Economic
Review
e s ' being equal to the percent responding "better" plus
one-half of the percent indicating "no c h a n g e " (Robert
S. R e i c h a r d 1988, 61). If all r e s p o n d e n t s a n s w e r
" S A M E , " for component / the value is fifty.
INDEX; = %BETTER.
+ (1/2 •
%SAME.).
The N A P M index is thus ordinal, not cardinal: the
index does not reflect precise levels of activity but instead indicates w h e t h e r a g i v e n m o n t h is better or
worse than the preceding one. Diffusion indexes such
as the N A P M series h a v e the merit of being highly
correlated with growth rates, but they are not as precise as surveys that measure actual production levels
from period to period. This is not to say that the N A P M
survey is wrong in any sense by design but simply that
a diffusion index does not measure growth rates based
on precise levels of activity. Because the N A P M index
is designed to gauge whether manufacturing is expanding or contracting in industries represented in the survey,
forecasters believe that it reasonably reflects "current"
c o n d i t i o n s in m a n u f a c t u r i n g . T h a t is, r e s p o n d e n t s '
views are seen as similar to actual production as measured in the Federal Reserve Board's index of industrial p r o d u c t i o n — i n statistical p a r l a n c e , the t w o are
correlated.
P r o d u c t i o n - W o r k e r Hours. Forecasters also use
production-worker hours to project industrial production. Use of this data is based on the idea that output
v a r i e s a c c o r d i n g to its i n p u t s — i n this c a s e , l a b o r
hours. By definition, output equals production hours
times labor productivity:
OUTPUT
= PRODUCTION
HOURS
OUTPUT/HOUR.
•
Any (theoretical) error in this type of concurrent forecasting model would reflect changes in productivity.
In addition, it m a k e s sense f r o m a m e t h o d o l o g i c a l
standpoint to use p r o d u c t i o n - w o r k e r hours f o r predicting the monthly change in the IP index because
the F e d e r a l R e s e r v e B o a r d uses the e m p l o y m e n t based data to d e r i v e initial e s t i m a t e s for m a n y I P
components.
The IP data undergo several revisions during the
early stages of estimating. Because of differences in
timing of data availability for actual goods produced
and inputs such as labor hours and electricity usage,
different methodologies are used for the various revisions. For the first IP release actual output levels are
not yet available for many industries, and estimates for
these are based on inputs such as production-worker
January/February 1992
hours. By value-added weight, about one-fourth of the
basic industrial production series is based on this employment data (Board of Governors 1986, 42). This
portion consists of components initially and ultimately
estimated with the worker hours data. Another 60 percent of the initial e s t i m a t e is d e r i v e d by i n f o r m e d
judgment, with production-worker hours being the primary factor considered. The remaining approximately
15 percent is based on physical product data. 3
T h e Federal Reserve Board uses the historical relat i o n s h i p s b e t w e e n o u t p u t and p r o d u c t i o n - w o r k e r
h o u r s — a l o n g with assumptions for productivity—to
derive initial estimates for various IP components. In
effect, the forecasts using the hours data are mimicking this process. H o w e v e r , such forecasts use broad
levels of aggregation. Usually, these models use broad
labor-data series that have some s u b c o m p o n e n t s for
which the Board of Governors actually has physical
product data available.
Methodology
The Data. The data for this study come from three
original sources: the Federal Reserve Board, the Bureau of Labor Statistics, and the National Association
of Purchasing M a n a g e m e n t . Secondary sources discussed below were used to get original (that is, unrevised) estimates for the purchasing m a n a g e r s ' indexes.
Current data (with full historical revisions) for production workers in manufacturing and for the average
manufacturing work week were obtained from the Bureau of Labor Statistics via a commercial data vendor.
Because initial percent changes are not published for
production-worker hours, initial and first-revised estimates for the levels, found in monthly issues of the
L a b o r D e p a r t m e n t ' s Monthly Labor Review,
were
needed to make this calculation. Use of first revisions
for "the previous m o n t h " is necessary so that initial
percentage changes can be calculated f r o m published
levels. The initial level for period t is the numerator,
and the first-revised level in period t-1 is the base for
this percent change. The relevant variable from the labor data is the percent change in aggregate productionworker hours and the average work week.
For industrial p r o d u c t i o n for m a n u f a c t u r i n g , the
initial percentage estimates were obtained f r o m the
Federal Reserve Board's monthly publication Federal
Reserve Statistical Release: Industrial Production and
Capacity Utilization or from the Federal Reserve Bulletin. Current data—that is, with historical revisions—for
DigitizedFederal
for FRASER
Reserve Bank of Atlanta
levels were obtained from a commercial data vendor,
and m o n t h l y percent c h a n g e s w e r e calculated f r o m
levels. For statistical comparability with percentage
changes for initial estimates, the calculated percentage
changes for revised IP data are rounded to one decimal
place. Because of differences in rounding between officially published index levels and the levels used by
the Board to calculate published percent changes, official percent changes may differ slightly from those derived f r o m levels.
Current values for the indexes produced by the National Association of Purchasing M a n a g e m e n t were
obtained directly f r o m this association through subscription. Historical data for first estimates came from
secondary sources, primarily the Wall Street
Journal,
the New York Times, and the Atlanta Journal. Data for initial estimates for N A P M indexes go back to 1982, when
the data were first released to the public in index form.
E m p i r i c a l C o m p a r i s o n s of N A P M v e r s u s
P r o d u c t i o n - W o r k e r H o u r s M o d e l s . The model
form used in this study to compare the predictive power of the N A P M index and production-worker hours is
simple. The simple monthly percent change in the manufacturing component of industrial production is regressed against the level of the N A P M index or against
the s i m p l e m o n t h l y p e r c e n t c h a n g e in p r o d u c t i o n worker hours. T h e level is used for the p u r c h a s i n g
managers' index because it is a diffusion index reflecting monthly changes in the manufacturing sector. Certainly, more complex models for predicting IP can be
used, but these simple m o d e l s provide a straightforward basis for comparing two competing independent
variables, and they reflect the methods and specifications used by market forecasters to m a k e quick forecasts as data are released.
A regression model compares the strength of the relationship between one variable (the dependent varia b l e ) and o n e or m o r e v a r i a b l e s (the i n d e p e n d e n t
variables). In contrast to correlations, regressions att e m p t t o s p e c i f y t h e m a t h e m a t i c a l f o r m of t h e
relationship between the variables. As is the case with
correlations, a strong statistical relationship does not
guarantee that one or more variables cause changes in
the other.
The models are seen as
%A IPmfg
=f(NAPM,C)
% A IPmfg =f(%&PROD
WORKER
HRS,C),
where C is a constant. The regression technique used
is ordinary least squares.
Economic Review
27
M a r k e t s f o c u s on the initial I P r e l e a s e a n d , as
d i s c u s s e d a b o v e , the initial e s t i m a t e is r e v i s e d . E s sentially, the revised IP n u m b e r s can be considered
a separate data series f r o m the initial e s t i m a t e bec a u s e d i f f e r e n t e s t i m a t i o n p r o c e d u r e s are u s e d . In
consideration of these differences and the m a r k e t s '
greater attention to initial estimates, comparisons are
m a d e using initial estimates for the monthly percent
change in industrial production as the dependent variable. M o d e l s using revised data are c o m p a r e d later.
Importantly, this study uses only the m a n u f a c t u r i n g
component. Explanatory variables also are in initial release f o r m for the N A P M index and the percent change
in production-worker hours. Tables 1 and 2 present the
regression output for each of these two models.
The NAPM-Based
Model. T a b l e 1 p r e s e n t s the
output of the N A P M - b a s e d model. T h e initial m o n t h ly percent change for manufacturing IP is regressed
against the level for the purchasing m a n a g e r s ' c o m posite index and a constant. T h e model has only m o d est predictive p o w e r with an adjusted R 2 of 0.36 and
a m e a n absolute error of the regression of 0.40 perc e n t a g e p o i n t s , c o m p a r e d w i t h t h e m e a n of t h e
absolute value of the dependent variable of 0.55 percentage points. 4
However, the model does have the expected signs
for the coefficients. For each index point change in the
index, the initial m a n u f a c t u r i n g IP changes by 0.06
percentage points minus the constant of 2.87. The purchasing m a n a g e r s ' composite index is positively cor-
related with manufacturing IP, as expected, and the tstatistics are greater than the rule of thumb of 2 (absolute value) for statistical significance.
The degree of confidence one can place in the reliability of the estimate for the coefficients is indicated by
/-statistics. The f-statistic is the ratio of the coefficient
value to its standard error. The greater the coefficient to
its standard error (ignoring the sign of the coefficient),
the greater the confidence that the coefficient is significantly different from zero. If the coefficient is not significantly different f r o m zero, then that variable does
not "explain" changes in the dependent variable.
This model might best be interpreted by setting the
model solution equal to zero and solving for the index
f o r the " n o c h a n g e " v a l u e . Values f o r the N A P M
greater than this solution value would tend to be associated with increases in industrial production, while
lower index levels would suggest declines in industrial
production.
% A IPmfg = b • NAPM + constant
0 = 0.06 • NAPM - 2.87
47.83 = NAPM.
This model s h o w s that " n o c h a n g e " in m a n u f a c turing IP is statistically a s s o c i a t e d with a value of
4 7 . 8 3 f o r the p u r c h a s i n g m a n a g e r s ' index o v e r the
1982-90 period. T h i s zero is the point estimate for
IP manufacturing w h e n the NAPM equals 47.83. T h e
c o n s t a n t is n e g a t i v e b e c a u s e index n u m b e r s below
Table 1
NAPM-Based Model
Dependent Variable: Monthly Percent Change in Industrial Production for Manufacturing, Initial Release
Regression Period: January 1982-November 1991
Independent Variables
Standard Error
-2.87
Constant
Number of Observations:
Mean Absolute Percent Change of
Dependent Variable:
Standard Error of Regression:
Mean Absolute Error:
Root Mean Squared Error:
Review
0.39
119
0.55
0.58
0.40
0.57
i-statistic
0.75E-02
0.60E-01
N A P M Index, Initial
Economic
28
Coefficient
8.03
-7.44
R2:
0.36
Adjusted R2:
Durbin-Watson:
Standard Deviation of MAE:
0.35
1.87
0.41
January/February 1992
47.83 are associated with declines in production. This
figure is somewhat below the level of 50 identified by
the N A P M as associated with a generalized decline in
the manufacturing sector. 5 Although the difference is
not statistically significant, regressions using different
periods consistently estimated values less than 50.
Logically, the N A P M production component might
b e m o r e correlated than the overall N A P M index
with manufacturing production. However, historical
data for initial estimates for the production component were not available prior to late 1988 (Reichard
1988, 61). A sufficient number of observations were
not available to run a statistically meaningful regression. 6
The Production-Worker
Hours Model. For this
model the initial monthly percent change for manufacturing IP is regressed against the percent change in
production-worker hours f r o m the initial estimate.
Production-worker hours are defined as the product of
(for manufacturing) the number of production workers and the average work week. Because initial estimates for IP are being modeled, initial estimates for
the percent change in the independent variable are also
needed. Although initial percent changes are not published, initial and revised levels are. The initial percent
change is calculated from the numerator in the percent, which reflects initial estimates for levels in period t, with the denominator being second estimates
(first revised) for period t-1.
This model, as shown in Table 2, is significantly
more accurate than the NAPM-based model. The adjusted R 2 is twice as high at 0.68, and the mean absolute error of the regression is about one-fourth lower at
0.30 percentage points.
This model also is consistent with standard assumptions about labor input and output. First, the coefficient for production-worker hours is positive. That
is, IP goes up or down with production-worker hours.
The sizes of the coefficients are reasonable. For each
p e r c e n t a g e p o i n t i n c r e a s e in p r o d u c t i o n - w o r k e r
hours, there is a 0.50 percentage point rise in manuf a c t u r i n g p r o d u c t i o n plus 0 . 1 9 p e r c e n t a g e points
from the constant.
The model implies positive growth in labor productivity—that is, output over the long run rises faster
than labor input. This result is shown in terms of a
mathematical relationship between output and labor
input. Rearranging terms in the model provides productivity estimates (which are merely output over labor input): productivity is a mathematical function of
the constant and slope in the regression. The 0.38 figure is the constant divided by the net of 1 minus the
hours coefficient. 7 The result is the break-even point
on positive growth in labor productivity. Most percent
changes in this independent variable are 1 - 0.38,
meaning that changes in output are generally greater
than changes in the labor input. Stated differently, productivity is positive over the observation period. This
Table 2
Production-Worker Hours Model
Dependent Variable: Monthly Percent Change in Industrial Production for Manufacturing, Initial Release
Regression Period: January 1982-November 1991
Coefficient
Standard Error
¿-statistic
Percent Change, Initial
Production-Worker Hours
0.50
0.31E-01
16.02
Constant
0.19
0.37E-01
4.91
Independent Variables
Number of Observations:
Mean Absolute Percent Change of
Dependent Variable:
Standard Error of Regression:
Mean Absolute Error:
Root Mean Squared Error:
DigitizedFederal
for FRASER
Reserve B a n k of Atlanta
119
0.55
0.40
0.30
0.40
R2:
0.69
Adjusted R2:
Durbin-Watson:
Standard Deviation of MAE:
0.68
1.52
0.26
Economic Review
29
finding is consistent with the more rapid rise of m a n ufacturing output from January 1982 through N o v e m ber 1 9 9 1 — a cumulative 41 percent i n c r e a s e — c o m pared with very little net change for production-worker
hours. 8
For the mean value of production-worker hours (initial estimate) over the January 1982-November 1991
period, a 0.05 percent monthly increase gives an expected value of 0.22 percent increase in industrial production. This figure implies roughly a 0.17 percent rise
in labor productivity in manufacturing each month on
average, the same as trend productivity growth of 2.1
percent annually. 9 Certainly, productivity is cyclical, but
the coefficient is consistent with longer-term trends.
The assumption of constant productivity gains probably
leads to a relatively low r-statistic for this variable.
The only other significant shortcoming of this model is the low Durbin-Watson statistic, which indicates
positive serial correlation. 1 0 That is, the error terms are
positively correlated and should be taken into account.
After the Cochrane-Orcutt technique was used to attempt to remove the serial correlation, the /-statistics
were not appreciably different. 1 1 However, the presence of serial correlation in the original ordinary least
squares model does suggest that the model is missing
a variable, probably a productivity variable. 1 2
Sources of Error. B o t h t h e p r o d u c t i o n - w o r k e r
h o u r s and N A P M index m o d e l s p r o v i d e s o m e w h a t
meaningful forecasts for the pending IP release. H o w ever, the production hours model provides significantly greater accuracy, although the error term for both
models is still large relative to the absolute value of
the dependent variable. The natural question to pursue
next concerns the sources of each model's error.
The purchasing managers' index is closely watched
by the financial markets because of its role as one of
the first indicators of manufacturing strength or weakness for a given month. However, this "timeliness" is
also one source of the series' relative inaccuracy. Several methodological features contribute to the index's
shortcomings: (1) the index is ordinal, not cardinal;
(2) the weighting scheme for respondents does not exactly mirror that for industrial production for manufacturing; (3) the sample is relatively small; and (4) the
timing of the survey does not correspond to the exact
period covered by the IP data.
As a diffusion index the N A P M index does not measure the level of aggregate production. T h e index is
simply a compilation of respondents' answers about
whether or not conditions are better, worse, or the same.
It does not and cannot determine whether a response of
"worse" more or less offsets a response of "better" be-
30
Economic Review
cause the magnitude for each is not known. Statistical
correlations are m u c h m o r e likely to be lower with
these types of ordinal measures (which indicate only the
direction—not level—of production) than with series
such as employment or Bureau of the Census data on
orders that have actual levels estimated. Less precision
is simply the nature of diffusion indexes.
N A P M stratifies its sample according to industry
gross-value weights. However, each respondent's ans w e r is given equal weight regardless of the f i r m ' s
size. Although weighting gross value shares of industry contributions is a valid m e t h o d , these shares or
weights are not the same as for manufacturing production. Industrial production component shares are based
on value-added. Also, while the survey size of about
300 is fairly large for this type of survey, it is quite
small relative to Bureau of the Census manufacturing
surveys or Bureau of Labor Statistics employment surveys, and a larger sampling error results. Taken together, these survey characteristics lead to noticeably
greater error terms.
Finally, the greater timeliness of N A P M ' s survey
may introduce previous m o n t h ' s information into the
survey results along with that of the current month, red u c i n g the i n d e x ' s predictive p o w e r for the current
month's manufacturing IP. To ensure that the purchasing m a n a g e r s ' index is available the first working day
of the month following the reference month, surveys
are sent to respondents early each month, and results
are tallied around the twenty-first of each month. Thus
r e s p o n d e n t s m u s t base their answers on only about
two weeks of the current month and on trends f r o m
the previous month, a pattern borne out statistically.
Forecasting models typically assume that the N A P M
index for the current month has the greatest explanatory
power for current-month IP manufacturing. However,
the large a m o u n t s of information f r o m the previous
month suggest that the current-month manufacturing IP
could more accurately be correlated with the following month's N A P M . For example, the August N A P M
index may be more closely associated with July manufacturing IP than is the July N A P M index.
Table 3 shows that a regression using the purchasing m a n a g e r s ' index (initial estimates) on a m o n t h ahead basis (the month f o l l o w i n g the current
m o n t h ) h a s a higher R2 than the concurrent basis reg r e s s i o n . Although the higher R2 is not statistically
different, it does suggest that the N A P M index has at
least as much information from the previous month as
the current m o n t h . T h i s c a r r y o v e r of the p r e v i o u s
m o n t h ' s information into current m o n t h ' s N A P M data would account for s o m e of the i n d e x ' s relatively
January/February 1992
Table 3
NAPM-Based Model with "Leading" Independent Variable
Dependent Variable: Monthly Percent Change in Industrial Production for Manufacturing, Initial Release
Regression Period: January 1982-November 1991
Independent Variables
Coefficient
N A P M Index, Initial
But Leading By O n e Month
0.62E-01
Constant
Number of Observations:
Mean Absolute Percent Change of
Dependent Variable:
Standard Error of Regression:
Mean Absolute Error:
Root Mean Squared Error:
8.28
0.75E-02
-7.69
0.39
-2.98
119
0.56
0.57
0.39
0.57
low explanatory power for the current month's manufacturing IP. 13 Statistically, it w o u l d m a k e as m u c h
sense to use last m o n t h ' s IP to forecast this m o n t h ' s
N A P M as it does to use this month's N A P M to forecast this month's IP.
T h e p r o d u c t i o n - w o r k e r hours m o d e l has a better
track record on average than N A P M , but it also has
significant error. What are the sources of these errors?
F i r s t , not all c o m p o n e n t s in m a n u f a c t u r i n g IP are
based on production-worker hours for initial estimates.
S e c o n d , by d e f i n i t i o n t h i s m o d e l e x p l a i n s o u t p u t
through changes in either hours or productivity. The
model's errors are innately errors in forecasting productivity.
This fact m a y be explained partly by output shifts
between high productivity sectors and low productivity sectors—on a vastly simplified basis, durables versus nondurables (because durables manufacturing on
average is m o r e intensely capitalized and has higher
labor productivity). T h e total production hours data do
not d i f f e r e n t i a t e any c h a n g e s in r e l a t i v e s h a r e s of
durables and nondurables. It is possible to do so, but if
the researcher intends to continue modeling with initial estimates and percent changes, data and modeling
difficulties c o m p o u n d rapidly. However, a simple ex
post m o d e l (using a v a r i a b l e to m e a s u r e d u r a b l e s '
share in IP) can be used to verify the importance of
shifts b e t w e e n durables and n o n d u r a b l e s . A l t h o u g h
there are nondurables industries with high capital-to-
Federal
Reserve Bank of Atlanta
¿-statistic
Standard Error
R2:
0.37
Adjusted R2:
Durbin-Watson:
Standard Deviation of MAE:
0.36
2.01
0.41
labor ratios, the assumption is that labor productivity
is higher for durables industries on average.
W i t h initial e s t i m a t e s of p r o d u c t i o n i n d e x e s for
durables and nondurables output, a ratio can b e constructed to measure relative shares of durables to nondurables. A p e r c e n t a g e c h a n g e in this index w o u l d
reflect how share is changing m o n t h l y — f o r example,
a rise would suggest a higher durables share and an increase in productivity. A s productivity rises, output
would be expected to increase with hours unchanged,
as t h e m o d e l s h o w n in T a b l e 4 c o n f i r m s . P e r c e n t
changes in manufacturing IP are regressed against percent changes in production-worker hours and in the
percent change in the ratio of initial durables to nondurables output.
Table 4 shows the production-worker hours m o d e l
with a variable added for the percent c h a n g e in the
r a t i o of d u r a b l e s IP ( a d v a n c e e s t i m a t e for the ind e x ) to n o n d u r a b l e s IP. A positive n u m b e r indicates
that durables output rose relative to nondurables. As
e x p e c t e d — b e c a u s e greater output in durables would
suggest higher labor productivity—the coefficient is
positive, with a r-statistic greater than 5. This result indicates that productivity changes attributable to shifts
in output do explain some of the error in the simple
production-worker hours model. Of course, the ratio of
durables to nondurables output is known only after the
fact and cannot be added to a predictive m o d e l . T h e
ratio of durables to n o n d u r a b l e s hours can b e used
Economic Review
31
Table 4
Production-Worker Hours Model with Durables Share Variable
Dependent Variable: Monthly Percent Change in Industrial Production for Manufacturing, Initial Release
Regression Period: January 1982-November 1991, excluding June 1985 and May 1990
Independent Variables
Coefficient
Standard Error
f-statistic
Percent Change, Initial
Production-Worker Hours
0.46
0.30E-01
15.59
Percent Change, Ratio of Durables
to Nondurables
0.16
0.30E-01
5.19
Constant
0.18
0.34E-01
5.32
Number of Observations:
Mean Absolute Percent Change of
Dependent Variable:
Standard Error of Regression:
Mean Absolute Error:
Root Mean Squared Error:
117
0.55
0.37
0.27
0.36
instead, but using initial e s t i m a t e s c o m p l i c a t e s the
data-gathering process.
T h e r e are o t h e r r e a s o n s the p r o d u c t i o n - w o r k e r
hours model has error. First, the Labor Department data are reflected only for the pay period up to and including the t w e l f t h of each m o n t h . Actual m o n t h l y
h o u r s m a y d i f f e r a n d , in f a c t , the Federal R e s e r v e
Board does attempt judgmental adjustments to account
for the differences when appropriate. Another problem
is that the hours data simply cannot reflect h o w intensively labor works. Production lines in some industries
m a y be run at varying speeds, within limits, with a
constant level of labor hours, providing a very significant source of error in this type of model.
This version of the production-worker hours model
still has a low Durbin-Watson statistic. Including a
productivity variable would probably yield additional
information.
C o m b i n e d Model. O n e might ask whether either
of the two data series has information content that the
other does not have. Could combined use of both series i m p r o v e forecasting? Table 5 shows the simple
production hours model with the initial N A P M index
variable added. This model gives slightly better forecasts than the simple production-worker hours model
with an adjusted R2 of 0.79 versus 0.69. Even though
m u l t i c o l l i n e a r i t y p r o b a b l y e x i s t s , the c o e f f i c i e n t s
32
Economic
Review
R¿:
0.75
Adjusted R2:
Durbin-Watson:
Standard Deviation of MAE:
0.74
1.45
0.24
are r e a s o n a b l e . B o t h d e p e n d e n t v a r i a b l e s ' c o e f f i cients remain strongly positive although slightly lower in value. However, the constant term's value turns
negative, offsetting the c o m b i n e d positive values of
c o e f f i c i e n t s f o r p r o d u c t i o n - w o r k e r h o u r s a n d the
N A P M index.
7Tie Longer Run
Economists use the types of models discussed here
to predict c u r r e n t - m o n t h changes in m a n u f a c t u r i n g .
H o w e v e r , these predictions are for initial estimates.
H o w well do these models predict more reliable revised output data? T h e production hours model m a y
have a significant advantage in predicting initial estimates simply because of the methodology for initial
IP. Does using revised data affect the relative accuracy of the competing models? Specifically, how well
do initial e s t i m a t e s f o r p u r c h a s i n g m a n a g e r s ' data
and for production-worker hours predict revised estim a t e s for m a n u f a c t u r i n g production? Table 6 gives
comparable statistics for the "initial-initial" models.
Table 7 provides a summary of the error measures for
models using revised data as the dependent variable.
For an interesting baseline of c o m p a r i s o n , forecast
January/February 1992
Table 5
Combined Model
Dependent Variable: Monthly Percent Change in Industrial Production for Manufacturing, Initial Release
Regression Period: January 1982-November 1991
Independent Variables
Coefficient
Standard Error
f-statistic
Percent Change, Initial
Production-Worker Hours
0.43
0.28E-01
14.95
N A P M Index, Initial
0.33E-01
0.47E-02
6.99
Constant
Number of Observations:
Mean Absolute Percent Change of
Dependent Variable:
Standard Error of Regression:
Mean Absolute Error:
Root Mean Squared Error:
119
0.55
0.34
0.25
0.33
numbers are compared with the revised n u m b e r "predicted" by using the initial IP figure.
With the revised IP data, the absolute mean value
of the d e p e n d e n t v a r i a b l e is l a r g e r — 0 . 6 3 v e r s u s
0.55—indicating larger errors in absolute magnitude.
T h u s , the e r r o r in p e r c e n t a g e t e r m s is p e r h a p s a
slightly m o r e valid measure because it takes into account the different magnitudes of the variable being
explained. Table 6 shows the production-worker
hours model and pooled model having a significant
advantage over the N A P M model for initial IP. Table 7
shows the production-worker hours and pooled model
with r e v i s e d IP n u m b e r s m a i n t a i n i n g their a d v a n tage—but just barely. All three models lose explanatory p o w e r . H o w e v e r , the " a d v a n c e IP m o d e l " f o r
predicting revised IP numbers indicates that revisions
have been substantial and that the loss of explanatory
power is to be expected.
Each month, forecasters predict the initial estimate
for IP. However, most use the readily available revised
data for both the dependent and independent variables.
H o w do models with fully revised data p e r f o r m ? Is
there a difference between using initial estimates versus revised estimates?
As seen in Table 8, the model using the composite
N A P M index is the weakest performer. With a sufficient number of observations available, a model using
the N A P M production component shows essentially no
Federal
Reserve Bank of Atlanta
-6.19
0.24
-1.51
R2:
0.79
Adjusted R2:
Durbin-Watson:
Standard Deviation of MAE:
0.79
1.81
0.22
improvement over the overall N A P M index. This finding is surprising because it seems likely that the other
f o u r c o m p o n e n t s in the overall index, theoretically
not as closely tied to production, would generate some
error. The pooled model—with the overall N A P M index included—does not perform quite as well as the
production-worker hours model. Over this period the
production-worker hours model with all revised data is
slightly m o r e accurate than with initial estimates for
the i n d e p e n d e n t v a r i a b l e . O v e r a l l , the p r o d u c t i o n worker hours model maintains its superiority when the
models use revised data—but not by the same degree.
Conclusion
Several findings bear keeping in mind as forecasters
use N A P M and the production-worker hours models to
predict IP: (1) For a given month, N A P M is not particularly accurate for predicting IP although over several
m o n t h s it reflects trends reasonably well. (2) Of the
two, the production-worker hours model is a better indicator for initial manufacturing IP releases, partly because of the IP methodology. (3) Both types of models
c o r r o b o r a t e near-term trends but are only m o d e s t l y
successful in forecasting the current month's IP number. (4) The models "miss" for different reasons. The
Economic Review
33
Table 6
Relative Error Estimating Initial Industrial Production
With Initial Estimates for Independent Variables 3
Adjusted
R2
Mean Absolute
Error
Standard Deviation
of MAE
Mean Absolute
Percent Errorb
N A P M Composite Index
0.35
0.40
0.41
0.32
0.41
0.42
93.59
N A P M Production Index0
Production-Worker Hours
0.68
0.30
0.26
71.85
Combined Model
0.78
0.25
0.22
60.72
90.28
Mean of Absolute Value of Dependent Variable: 0.55
Table 7
Relative Error Estimating Revised Industrial Production
With Initial Estimates for Independent Variables 3
Adjusted
R2
Mean Absolute
Error
Standard Deviation
of MAE
Mean Absolute
Percent Error'1
N A P M Composite Index, Initial
0.23
0.49
0.46
N A P M Production Index, Initial
n.a.
n.a.
n.a.
Production-Worker Hours, Initial
0.42
0.45
0.37
92.47
Combined Model, Initial
0.48
0.42
0.35
93.29
Initial Estimate IP
0.58
0.38
0.31
108.32
n.a.
85.24
Mean of Absolute Value of Dependent Variable: 0.63
Table 8
Relative Error Estimating Revised Industrial Production
With Revised Estimates for Independent Variables 3
Mean Absolute
Error
Adjusted
R2
Standard Deviation
of MAE
Mean Absolute
Percent Error15
N A P M Composite Index, Revised
0.21
0.50
0.46
N A P M Production Index, Revised
0.20
0.50
0.47
Production-Worker Hours, Revised
0.50
0.41
0.35
87.92
Pooled Model, Revised
0.51
0.41
0.34
91.15
108.89
108.73
Mean of Absolute Value of Dependent Variable: 0.63
3
Error
measures
b
Nine
observations
ro results
cover
in a divisor
c
Initial
estimates
d
Nine
observations
the January
in which
of
1982-November
the dependent
1991
variable
equals
period.
zero
are omitted
prior
to
from
this measure
because
a dependent
variable
equal
to ze-
zero.
for NAPM
(different
components
were
from previous
34
Economic Review
not available
table)
in which
October.
the dependent
variable
equals
zero
are omitted
from this
measure.
January/February 1992
that, as with other indicators, these m e a s u r e s of c o n c u r -
p r o d u c t i o n - w o r k e r h o u r s m o d e l falls short b e c a u s e of
the r a p i d c h a n g e s in o v e r a l l m a n u f a c t u r i n g p r o d u c t i v i t y — i n c l u d i n g the e f f e c t s o f s h i f t s in the p r o d u c t i o n
m i x ( m o r e or less p r o d u c t i o n of d u r a b l e s relative to
nondurables). T h e N A P M m o d e l ' s errors arise f r o m a
lack of c o n t r o l l e d w e i g h t i n g of the N A P M s u r v e y
a n d its i n h e r e n t lack of p r e c i s i o n as a d i f f u s i o n i n d e x .
(5) T h e size of revisions to industrial p r o d u c t i o n s h o w
rent p r o d u c t i o n can fail to p r e d i c t the strength of the
m a n u f a c t u r i n g sector. F i n a l l y , (6) s i g n i f i c a n t d i f f e r e n c e s arise in using initial e s t i m a t e s in the m o d e l s as
o p p o s e d t o using r e v i s e d e s t i m a t e s . In particular, the
p r o d u c t i o n - w o r k e r h o u r s m o d e l has a d e c i d e d a d v a n t a g e w i t h initial e s t i m a t e s but m u c h less so w i t h rev i s e d data.
Notes
1. This index is discussed in greater detail in The Report on
Business: Information Kit, compiled by the NAPM Information Center, Tempe, Arizona.
2. The survey also has questions about commodity prices and
exports and imports, but these are not factored into the overall index.
3. There are three types of data used in estimates for industrial
production. From initial estimates through benchmark revisions, these are physical product data, production-worker
hours, and kilowatt hours. However, no kilowatt hours data
are available for the initial estimates.
4. The R2, or coefficient of determination, is a measure of how
much of the variation in the dependent variable is explained
statistically by the variations in the independent variables.
As variables are added to a model, the R2 can only go up or
show no change. The adjusted R2 is a variant of the R2 measure that takes into account the number of independent variables, allowing competing models with differing numbers
of variables to be compared more easily.
The mean absolute error is merely the average error
without regard to the sign of the error. This error measure is
more intuitive than the standard error of the regression. One
can use this measure to compare the error with the magnitude of the dependent variable.
The standard error (as shown in Table 1) of the regression provides an approximate confidence interval for model estimates. A point estimate—taken by plugging in values
for the independent variables—plus the standard error (or
multiples of the standard error) on either side of the point
estimate gives an approximate confidence bond in which
one believes the actual values of the dependent variable lie.
The approximate 95 percent confidence bond is the point
estimate plus or minus two times the standard error. The
smaller the standard error, the narrower the confidence
bond and the greater confidence one has in the point estimate.
5. Over the entire January 1948-November 1991 period, using
revised data series, the NAPM composite index level that is
statistically associated with no change is 49.1. This value
differed noticeably depending on the estimation period used
in the regression.
6. Because revisions to NAPM data primarily reflect new seasonal factors, revised data are a close approximation of ini-
Federal
Reserve Bank of Atlanta
tial estimates. A regression using revised production component data as the independent variable to explain initial IP
is shown in Table 6. Over the September 1988-November
1991 period, the two series have a correlation coefficient of
0.99.
7. For productivity growth to be positive, the percent change
in IP must be greater than the percent change in productionworker hours. The model solution results in % A / P = ft, •
% A HRS + CONSTANT, where ft, is the slope. Because the
percent change for hours equals that for IP, %A HRS can
be substituted for %MP, so that %AHRS = ft, • %AHRS
+ CONSTANT. Rearranging the terms yields %A HRS =
CONSTANT/^ 1 - ft,). This is the formula for deriving the
break-even point for hours for maintaining positive labor
productivity.
8. Based on revised levels for both series. The comparison periods make these series difficult to compare because both
November 1991 and January 1982 were during downturns
when both employment and output were deviating from trend.
9. This 2.1 percent productivity figure is low relative to traditional measures, probably owing to definitional differences
inherent in the model.
10. The Durbin-Watson statistic measures the degree that model errors are correlated over time, called serial correlation or
autocorrelation. That is, the error in the current period may
be affected (statistically) by the error in the previous period;
an error may carry over from period to period. Classical regression analysis assumes that error terms are unrelated. A
value of 2 or "close" to 2 suggests no or very little serial
correlation in a regression model.
11. The Cochrane-Orcutt technique is one of several methods
used to remove the effects of serial correlation. This technique introduces an error term into the model to try to quantify the relationship between errors over time.
12. As discussed in a later section, rather than looking at the
problem as a missing variable that would reflect changes in
the quality of labor, the variable should be one that lakes into account shifts in production between sectors with different capital-to-labor ratios.
13. The same comparison was made for the production-worker
hours model. With the dependent variable "led" by one
month, there was an R2 of near zero.
Economic
Review
35
References
Board of Governors of the Federal Reserve System. Industrial
Production, With a Description of the Methodology. Washington, D.C., 1986.
National Association of Purchasing Management. The Report
on Business: Information Kit, 1990.
36
Economic
Review
Reichard, Robert S. "NAPM Upgrades Indicators." Purchasing
World, September 1988, 61-62.
January/February 1992
R e v i e w Essay
Breaking Financial Boundaries:
Global Capital, National Deregulation,
a/zd Financial Services Firms
by David M. Meerschwam.
Boston: Harvard Business School Press, 1991.
306 pages. $35.00.
B. Frank King
The reviewer is vice president
and associate director of the
Atlanta Fed's research
department.
Federal
Reserve Bank of Atlanta
here is a large body of literature analyzing the d e v e l o p m e n t of
America's financial system during the last t w o decades, but it says
little of the international forces that helped to bring about the transition or of similar experiences in other nations. David M. Meerschwam adds a global perspective to the literature with Breaking
Financial Boundaries, which examines post-World War II changes both in
the international financial system and in domestic systems in Japan and the
United Kingdom as well as the United States. After outlining the forces for
change and their results and implications, he discusses the impact of newly
integrated, and newly volatile, financial markets on (large) financial institutions' m a n a g e m e n t strategies.
Meerschwam, a m e m b e r of the Harvard Business School faculty, sees the
vast, jarring changes in financial systems and institutions in the two decades
since the demise of the Bretton Woods international exchange rate system as
a story of the breakdown of barriers between financial markets in the global
economy and within nations. He approaches the complex story primarily by
analyzing the history of economies, institutions, and ideas and concludes
that market integration has brought increased flexibility, freedom, and efficiency, as well as instability. These results, of course, have implications for
financial institutions and their customers and for governments, which seek
both efficient intermediation and economic stability.
Economic Review
37
71ie International Financial System
M e e r s c h w a m begins with one of the subjects designed to set his book apart—international forces that
have affected financial evolution. T h e history he presents interweaves economists' ideas about international
balance-of-payments adjustment and developments in
the international exchange system. Its focal point is the
impact on national financial systems of the early 1970s'
breakup of the Bretton Woods fixed exchange-rate system adopted at the end of World War n . This is the first
of a series of discussions that cover the effects of the
same basic forces in four only somewhat different econ o m i c e n v i r o n m e n t s — t h e United States, the United
Kingdom, Japan, and the international economy.
In contrast to later parts of the book, Meerschwam's
treatment of international forces is steeped in the history of e c o n o m i c thought. His text is a h e l p f u l laym a n ' s tour of two centuries of theory of international
trade and finance; it is footnoted for more seriously interested readers. The author outlines the development
of intellectual backing for fixed exchange rates supported by the gold standard. H e illustrates h o w the
Keynesian revolution in the 1930s ushered in the development of more or less nihilistic theories that cast
doubt on whether economies would m o v e toward an
equilibrium balance of trade under any exchange-rate
system.
What had emerged out of the interplay of ideas by
the mid-1960s, Meerschwam argues, was professional
confusion among economists about the nature of the
international adjustment m e c h a n i s m and strong support of flexible, m a r k e t - d e t e r m i n e d e x c h a n g e rates.
There were three main arguments for flexible rates:
that they yielded e c o n o m i c e f f i c i e n c y at little cost,
that they secured exchange rate stability through stabilizing speculation, and that they protected governm e n t s ' domestic m o n e t a r y and fiscal policies f r o m
external influences.
By the early 1970s substantial trade imbalances in
several large trading countries and, eventually, a run
on the dollar led to crises which demonstrated that the
Bretton Woods system was unworkable. And, Meerschwam contends, with no strong intellectual support
for a remodeled fixed rate system, the nations of the
world chose to adopt a market-oriented flexible rate
system.
The main point the author makes is that rate flexibility did not protect countries' domestic policies from
external forces. Rather, during the oil price shock of
1973-74, as g o v e r n m e n t s around the world found it
38
Economic
Review
quite difficult to accommodate the externally generated supply restrictions, flexible rates merely allowed
national policymakers to delay controlling inflation.
Their procrastination p r o m o t e d m u c h m o r e volatile
e c o n o m i e s , w h i c h in t u r n g e n e r a t e d m u c h m o r e
volatile domestic inflation rates and international capital flows.
Existing markets evolved and new markets and financial instruments were developed to facilitate these
flows and m a n a g e the risks that accompanied them.
As markets changed, so did ways of doing financial
b u s i n e s s . G o n e w e r e the old " r u l e s of the g a m e , "
which had emphasized controlled prices, strong customer relationships, and segmented product markets.
Large financial institutions were virtually required to
compete in the new and evolving markets if they were
to survive.
iVational Financial Systems
T h e double volatility of inflation rates and international capital flows inevitably affected national
economies around the globe. M e e r s c h w a m examines
three i m p o r t a n t d o m e s t i c financial s y s t e m s — i n the
United States, Japan, and the United K i n g d o m — t h a t
c o n v e r g e d t o w a r d s i m i l a r o r g a n i z a t i o n d u r i n g the
1970s and 1980s. T h e s e three systems had evolved
over long periods of time in differing economic and
cultural conditions. By the 1940s these countries had
reached different levels of (a) concentration of private financial institutions and government influence,
(b) formality of relationships between government and
financial institutions, and (c) the extent to which government operated to allocate credit.
M e e r s c h w a m points out, h o w e v e r , that even diverse cultural and economic conditions did not prevent marked c o m m o n organizational characteristics.
In each e c o n o m y financial institutions operated under
strong government influence. In each, financial product markets were segmented by institution and important financial prices were controlled. Dealings between
financial institutions and their customers were influenced much more by relationships than by (often controlled) prices. In the early post-World War II era,
these conditions combined with relatively stable economic climates in each country to m a k e most financial institutions profitable and to limit the incentive to
innovate.
The period during which the three financial systems
fit the a b o v e d e s c r i p t i o n w a s a c t u a l l y q u i t e short.
January/February 1992
Meerschwam suggests the 1950s and 1960s; however,
one might take a year or two off each end. Even during
these two decades forces for change were operating.
By 1963, for instance, the negotiable certificate of deposit allowed larger U.S. banks to tap the money markets at home and abroad for funding, providing a m a j o r
addition to competition in money and loan markets.
It is in the discussion of forces for change, both domestic and international, that M e e r s c h w a m makes one
of his most v a l u a b l e contributions. Building on his
earlier discussion of developing international capital
flows and markets, the influence of flexible exchange
rates on m a c r o e c o n o m i c policy formulation, and the
attraction of foreign operations to large financial institutions, M e e r s c h w a m argues convincingly that these
forces o v e r c a m e national e c o n o m i c and cultural idiosyncracies to bring about further convergence among
national systems. National parallels are clearly displayed: national markets became increasingly international, and in each market prices became increasingly
market-determined and flexible, market segmentation
was substantially reduced, and government relations
were amended.
Less convincingly, the author contends that crumbling market barriers have caused significant financial
instability. Each of the three nations he discusses certainly had its financial crises in the 1970s and 1980s,
e x h i b i t i n g m o r e v o l a t i l e i n t e r e s t r a t e s and c a p i t a l
flows. While inflation promoted market integration, it
probably also played a m a j o r role in financial instability by generating excessive expectations of asset
appreciation and volatile interest rates. In addition,
Meerschwam mentions government subsidies to bank
risk-taking, which appear in each country's financial
s y s t e m , but g i v e s t h e s e no s e r i o u s d i s c u s s i o n as
s o u r c e s of p r o b l e m s f o r f i n a n c i a l i n s t i t u t i o n s and
volatility for financial systems.
Meerschwam claims that the forces he discusses lie
at the base of the evolution in both international and
national systems. However, he gives only superficial
treatment to a crucial set of developments—the rapid
advancement of information-communications and datamanipulation technology and of intellectual techniques
for v a l u i n g c o m p l e x f i n a n c i a l assets and a s s e s s i n g
their risk that h a v e played a m a j o r role in financial
evolution since World War II. This progress has made
the information that those engaged in financial transactions require—and that financial institutions m a k e
their m o n e y by p r o v i d i n g — c h e a p e r to acquire, process, and communicate and has increased the precision
and accuracy of analysis. Meerschwam's failure to address this topic adequately is a significant shortcom-
Federal Reserve Bank of Atlanta
ing, for without these developments it is difficult to
imagine that segmented national and international fin a n c i a l m a r k e t s w o u l d h a v e o p e n e d up, t h a t t h e
n u m b e r a n d c o m p l e x i t y of f i n a n c i a l i n s t r u m e n t s
and t r a n s a c t i o n s could h a v e i n c r e a s e d , and that the
price and relationship flexibility that have been characteristic of the past two decades would have been int r o d u c e d . It is not that the state of the art entirely
d e t e r m i n e s d e v e l o p m e n t s , but it can be argued that
many private and public decisions about h o w to react
to the pressures of the international e c o n o m y would
have been different had not substantial innovations in
communication and analysis occurred.
institutional Strategies
W h a t e v e r their root c a u s e s , c h a n g e s in financial
systems present financial institutions' managers with a
new set of challenges. Meerschwam turns to these in
the final section of his book. Unfortunately, this discussion suffers from the author's certainty that financial systems will continue to u n d e r g o c o m p l e x and
unpredictable change. In the face of such change, confidence and concreteness fade and generic discussion
takes their place.
His first attempts to outline strategy are appropriately cautionary. Through discussions of Continental's
energy lending, Citibank's L D C adventures, Morgan
Stanley's demanding entry into the Tokyo market, and
the Crédit Suisse joint venture with First Boston, he illustrates some of the pitfalls of seemingly appropriate
strategies gone awry.
Building on these introductory discussions, Meerschwam emphasizes the rapidity of change in financial
products and markets, the integrated nature of financial products, and the potential contestability of most
financial markets. This line of thought leads h i m to the
dilemma so often apparent in others' discussion of financial f i r m s ' strategy: Should a financial institution
choose a niche or offer a wide selection of integrated
products? To his credit, M e e r s c h w a m does not back
d o w n f r o m the difficult question by selecting both.
Rather, h e points out that o f f e r i n g a broad array of
products involves substantial difficulties: transfer pricing, a l l o c a t i n g c o m p e n s a t i o n , and c o s t c o n t r o l , to
which one might add risk management. He then suggests that value added depends on taking advantage of
market anomalies, production efficiency, superior
ability to join firm and employee resources, and information advantages. These ideas do not differ greatly
Economic Review
39
from suggestions many others have offered for many
other types of firms.
The author's final foray into strategy leads him to
the raison d ' ê t r e for financial intermediaries—information advantage. He believes that future advances in
communications and analysis technology may m a k e
information processing far less costly to members of
the public, thus eliminating or severely diminishing
the need for financial intermediaries' value added. In
such an event, financial intermediaries would be left
with (possibly transitory) knowledge advantages to be
exploited through advisory services.
.Summary
Breaking Financial Boundaries makes four claims
to value added: It bridges a gap between the extensive
academic literature on recent financial evolution and
the business managers who need to be in the know but
h a v e little time for or familiarity with the argot of
academia. Second, M e e r s c h w a m ' s work e m p h a s i z e s
international forces that are important bases for unders t a n d i n g d e v e l o p m e n t s in f i n a n c i a l s y s t e m s in the
1970s and 1980s and developing business strategies
for the future. Further, the book shows how this evolution has been influenced by interactions among financial firms, national financial systems, and the global
financial system. Finally, Breaking Financial
Boundaries provides guidance for managers of financial in-
40
Economic
Review
stitutions on their best approach to the global financial
system's changing condition.
The author clearly achieves his first goal. The
bridge f r o m academics to the practitioner is solidly anchored with copious footnotes to academic literature at
the s a m e t i m e that it is, f o r the m o s t part, built of
s t r a i g h t f o r w a r d , u n d e r s t a n d a b l e stories of developments in international and national markets. For those
familiar with the m o r e scholarly literature there is little
new here, but the author's aim was integration rather
than creation.
T h e other three goals are not as well achieved as
the first. The book's international perspective does ind e e d b r o a d e n r e a d e r s ' u n d e r s t a n d i n g of i m p o r t a n t
evolutionary forces. However, it gives short shrift to
important elements related to information processing
and risk subsidies.
Analyses of interactions among financial markets,
financial firms, and governments are common in scholarly work as well the popular press. M e e r s c h w a m is
neither the first nor the most successful in pointing out
interactions.
Meerschwam's concluding consideration of corporate strategies is his most difficult challenge, and it is
this section of the book that disappoints most. His discussion reads rather like a long conversation, carried
on well past midnight—interesting and insightful if one
was there but difficult to follow when recounted. Readers not willing to work hard on integrating its elements
would be well advised to skip it, settling for the significant added value in the book's first three sections.
January/February 1992
Federal Reserve Bank of Atlanta Working Papers
T
h e R e s e a r c h D e p a r t m e n t of the Federal R e s e r v e B a n k of Atlanta p u b l i s h e s a w o r k i n g p a p e r series to
c o n v e y the research of staff e c o n o m i s t s and visiting scholars and stimulate p r o f e s s i o n a l discussion and
e x p l o r a t i o n of e c o n o m i c subjects. To subscribe to the w o r k i n g p a p e r series or to receive single copies of recently p u b l i s h e d p a p e r s (sent automatically to current subscribers), please c h e c k the a p p r o p r i a t e b o x ( e s ) below and return the f o r m as indicated.
91-7
Estimating the Minimum Risk Maturity Gap
James E. McNulty, George E. Morgan III, and
Stephen D. Smith
91-13
The Wild Card Option in T-Bond Futures Is
Relatively Worthless
Hugh I. Cohen
91 -8
Money Demand and Relative Prices in the
German
Hyperinflation
Ellis W. Tall man and Ping Wang
91-14
Legal Restrictions and Welfare in a Simple
Model of Money
William Roberds
91 -9
Generalized Put-Call Parity
David F. Babbel and Laurence K. Eisenberg
91-15
Quantity-Adjusting Options and Forward
Contracts
David F. Babbel and Laurence K. Eisenberg
91-10
Misspecification Bias in Tests of the Forward
Foreign Exchange Rate Unhiasedness
Hypothesis
Janice L. Boucher
91-16
Option Pricing with Random Volatilities in
Complete Markets
Laurence K. Eisenberg and Robert A. Jarrow
91-17
In Search of the Liquidity Effect
Eric M. Leeper and David B. Gordon
91-18
Competition for More Than One Class of Borrowers Using Different Credit-Worthiness
Tests
Larry D. Wall
91-11
Consumer Attitudes and Business
Eric ML Leeper
Cycles
91-12
Multiple Reserve Requirements : The Case of
Small Open Economies
Marco Espinosa
•
Please start my subscription to the Working Paper Series.
•
I currently receive other Federal Reserve Bank of Atlanta publications. (To expedite your order, please
provide the subscriber number that appears above your mailing address: R
)
•
Please send me a copy of the following working papers:
• 91-7
• 91-8
• 91-9
• 91-10
•91-11
•
91-12
• 91-13
• 91-14
• 91-15
• 91-16
• 91-17
Q91-18
Name
Address
City
——
State
ZIP
Return this form to the Public Affairs Department, Federal Reserve Bank of Atlanta, 104 Marietta Street,
N.W., Atlanta, Georgia 30303-2713. Some previously published working papers are also available. For more
information write to this address or call 404/521-8020.
1
I
Bulk Rate
U.S. Postage
PAID
Atlanta, GA
Permit 292
B A N K OF
Public Affairs
Department
104 M a r i e t t a S t r e e t ,
N.W.
Atlanta, Georgia 30303-2713
( 4 0 4 )
5 2 1 - 8 0 2 0
printed on recycled paper