Full text of Review (Federal Reserve Bank of St. Louis) : November 1985

View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Vol. 67, No. 9

November 1985

5 Would Lower Federal Deficits Increase
U.S. Farm Exports?
20 Monthly Economic Indicators: A Closer
Look at the Coincident Index

THE
FEDERAL
A RESERVE
J kI1ANK of
Z
ST. LOUIS

T h e Review is p u b lish ed 10 tim es p e r y ea r by th e R esea rch a n d Public Info rm ation D e p a rtm en t o f
th e F ed era l R eserv e Bank o f St. Louis. Single-copy su bscrip tio ns a re available to th e p ublic f r e e o f
ch a rg e. Mail re q u ests f o r su bscrip tio ns, back issues, o r a d d res s ch a n g e s to: R esea rch a n d Public
Inform ation D ep a rtm en t, F ed era l R eserv e Bank o f St. Louis, P.O. Bopc 442, St. Louis, M issouri 63166.
T h e views e x p re s s e d a re th o se o f th e individual a uth ors a n d d o not n ecessa rily reflec t official
positions o f th e F ed era l R eserv e Bank o f St. Louis o r th e F ed era l R eserve System . A rticles h e re in m ay
b e re p rin t e d p ro v id ed th e s o u rc e is cred ite d . Please p rov id e th e Bank's R esea rch a n d Public
Inform ation D ep a rtm en t with a copy o f re p rin te d material.

F ed eral R eserve Bank of St. Louis
Review
November 1985

In This Issue . . .

In the first article of this Review, M ichael T. Belongia and C ourtenay C. Stone
investigate the validity of a conventional policy recom m en d ation am ong agricul
tural econ om ists: that red u cin g the federal budget deficit is a n ecessary co m p o
nent of any effort to increase agricultural exports. W hile widely a ccep ted , Belon
gia and Stone argue that at least three propositions m ust be tru e for this
recom m en d ation to have som e legitim acy. The first two propositions require
positive relationships betw een deficits and the real rate of interest an d betw een
the real rate of interest and the real value of the dollar. The last proposition
requires a negative relationship betw een the real value of the dollar and real farm
exports.
Belongia and Stone investigate these propositions by explaining the relevant
eco n om ic theory behind each and offering both sim ple an d detailed statistical
evidence regarding the predictions of these theories. Their analysis indicates
there is considerable doubt that larger budget deficits have raised real interest
rates and, consequently, the real value of the dollar. They do find su p p ort for the
third proposition relating the exch an ge rate to farm exports. In view of the o th er
conflicting evidence, however, the authors co n clu d e that it is safe to say the
relationship betw een budget deficits and farm exp orts is unresolved.
In the secon d article in this issue, "M onthly E co n om ic Indicators: A Closer
Look at the Coincident Index,” Keith M. Carlson d iscu sses the C om m erce D epart
m en t’s com p osite index of coincident indicators. This index co n d en ses the
inform ation from the m ost im portant m onthly e co n om ic ind icators into one
sum m ary m easure. The co m p o n en t indicators from w hich the co m p o site index is
derived include em ployees on nonagricultural payrolls, personal incom e less
transfer paym ents in con stan t (1972) dollars, industrial produ ction , an d m an u fac
turing and trade sales in co n stan t (1972) dollars.
After describing the coinciden t index and its co m p o n en t indicators, Carlson
d iscu sses the C om m erce D ep artm en t’s scoring system . This system is b ased on
seven criteria. The focus of the article, however, is on the criterion of cyclical
p erform ance. An exam ination of the co m p o n en t ind icators reveals th at each
ind icator generally conform s well with the National Bureau of E co n om ic Re
search 's business cycle reference points, particularly w hen viewed with som e
perspective.

Would Lower Federal Deficits
Increase U.S. Farm Exports?
Michael T. Belongia and Courtenay C. Stone

NY farm policy experts believe that U.S. agri
cultural exports w ould be increased significantly if
federal budget deficits w ere red uced . One su ch ex
pert, for exam ple, has co m m en ted that “a m ore nearly
balanced federal budget probably w ould do as m uch
as anything to im prove o u r agricultural export perfor
m an ce.”1 O ther analysts also have predicted that the
farm outlook will rem ain bleak until this nation d e
velops “a credible plan to red u ce the en orm ous Fed
eral budget deficits.”- This view is so pervasive that it
might now be con sidered the conventional wisdom
on the subject.'
If this view is valid, it has im portant im plications for
d om estic farm policy legislation, including the p en d
ing farm bill.1 If federal budget deficits have seriously
red u ced farm exports and, consequently, real farm
incom e, then legislators should focus primarily on
reducing the deficit to revive farm exports and incom e;
in this case, cu rren t com m odity program s may need
little fundam ental change o nce the deficit has been
red u ced . If budget deficits have not contributed m ate
rially to the decline in farm exports, however, then
focusing attention on deficit red u ction m easures may

Michael T. Belongia and Courtenay C. Stone are senior economists at
the Federal Reserve Bank of St. Louis. David J. Flanagan provided
research assistance.
'Schuh (1984), p. 246.
2Duncan and Drabenstott (1985).
30 f course, this view is not confined solely to farm policy experts. It is
held by a large number, perhaps even the vast majority, of policy
analysts. For similar statements about the effect of deficits on
exports, see Clark (1985), Downs (1985), Kraft (1985) and
Modigliani (1985).
“The omnibus farm legislation currently in effect is a four-year bill that
was passed in 1981. Congress is now debating the issues encom
passed by a new four-year bill, renewal of existing legislation, or
returning to the “ permanent" legislation of the 1930s and 1940s that
covers most major commodities.

divert attention from m ore fundam ental changes that
might be required in farm com m odity program s. The
purpose of this article is to describe the theoretical
links betw een federal budget deficits and U.S. farm
exports and to exam ine the em pirical evidence on
these links.

THE FARM E X P O R T PR O BLEM
singled out as a strong, p erh ap s the prim aiy, suspect
in attem pts to explain w hy farm exp orts have declined
so sharply in recen t years. A p rim a fa c ie case can be
m ade for the deficit explanation bv simply looking at
the recent relationship betw een exp orts and the defi
cit; this com p arison is show n in ch art 1 for sem i
annual data since 1973.
If we look only at the past four years, we see that
nom inal farm exp orts declined from 1981 through
1983, rose marginally last year, and have fallen again
through early 1985. During this sam e period, federal
deficits, as m easured by the national incom e a c
counts, skyrocketed, rising from $64 billion in 1981 to
$176 billion in 1984. The association betw een rising
deficits and falling exp orts app ears to be obvious.
Yet, w hen the entire period is exam ined, this co n
clusion is not so obvious. Deficits w ere rising and
falling from 1973 (when it w as only $6 billion) to 1981.
This period w as one of generally rising farm exports.
Thus, the view that rising deficits adversely affect
exp orts is one that seem s to be based primarily on
evidence from the m ost recen t period. Such a narrow
focus necessarily raises questions about the generality
of the presum ed relationship and the likely effect of
policies designed to exploit the relationship.
To get a b etter perspective on the relationship b e
tween deficits and farm exports, w e m ust focus on the

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

C hart 1

Deficit and Farm Exports

1973

theoretical relationships that tie them togeth er and
the em pirical support for these underlying theories.

TH E LINKS IN T H E D EFICIT-FARM
E X PO R T CHAIN
The conventional view of the links betw een deficits
and farm exp orts is show n in figure 1. In this fram e
work, the problem s of red u ced exports, expan d ed
im ports and political m easures prom oting protection
for d om estic industries can be traced backw ard, stepbv-step, to their so u rce: large, persistent federal
budget deficits. An exam ination of figure 1 show s that
there are at least three key eco n om ic relationships
that m ust exist for this conventional view to be valid.
First, o th er things unch an ged, deficits m ust be re
lated system atically to real interest rates (interest rates

1985

adjusted for exp ected inflationl; specifically, higher
deficits m ust raise real interest rates and low er deficits
m ust red u ce them . Second, real interest rates m ust be
related system atically to the real foreign exchange
value of the dollar (the dollar’s value after adjusting for
differences in inflation betw een the United States and
foreign countries); higher real rates m ust raise the
dollar’s real value an d low er real rates m ust red u ce it:'
Third, the real foreign exch an ge value of the dollar
m ust be related system atically to real agricultural ex
ports (agricultural exp ort receip ts ad ju sted for m ove
m ents in the general p rice level); higher real exch an ge
rates m u st red u ce real farm exp orts and low er e x
change rates m ust increase them . The con clu sion that

5
More correctly, the second link refers to the impact of higher real
interest rates in the United States vis-a-vis those in other countries.
This is explained later in this article.

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Figure 1
A Schematic View of The Theoretical
Linkages Between Deficits and Farm Exports

Large,
persistent
federal
deficits

Low
domestic
investment

High
real
interest
rates

Capital
inflows

High cost
for debt
service

Strong
dollar

Reduced
exports
+
Expanded
imports

Measures
to protect
*
domestic
industries

Lower
inflation

SOURCE: Dobson (1984), p. 49.

low er federal deficits will increase farm exp orts by
reducing real rates of interest and, thus, the real for
eign exchan ge rate of the dollar, relies on the validity
of these links. We analyze these links in turn in the
following sections.

LINK # 1 : WOULD LO W ER D EFIC ITS
REDUCE REAL IN T ER EST RA TES?
Choosing Appropriate Deficit and
Interest Rate Measures
One basic problem with tiying to discern the rela
tionship betw een deficits and interest rates is that the
m easures w e observe m ust be "ad ju sted ,” both to
eliminate potentially confounding influences and to
focus the analysis on those m easures that are em p h a
sized by the underlying eco n om ic theoiy. The deficit
m easure can be adjusted in a variety of wavs. Three
com m only used p ro ced u res are: (1) to adjust for the
im pact of inflation by using a real deficit m easure; (2)
to adjust for the size of the econ om y by dividing the
deficit by som e m easure of spending o r saving; and (3)
to rem ove the business cycle influences on the deficit.
Interest rates also m ust be adjusted appropriately if
we are to cap tu re the deficit's influence on them .
Market interest rates — the ones that we see quoted

every day — ca n be thought of as the sum of two basic
com p on en ts: the exp ected inflation rate and the ex
p ected (or ex a n te) real rate of interest." Changes in the
deficit per se should have no effect on the exp ected
rate of inflation unless the Federal Reserve is exp ected
to m onetize the deficit (that is, increase its p u rch ases
of governm ent bonds).7 Since ch an ges in the exp ected
rate of inflation, how ever, cau se nom inal interest rates
to change an d obscu re the im p act of ch an ges in the
deficit, w e m ust rem ove the influence of ch an ges in
inflation exp ectation s; w e m ust focus on the deficit’s
im pact on the exp ected real rate of interest.

The Conventional View o f the Deficit’s
Impact on the Real Rate o f Interest
The view that larger deficits increase the exp ected
real rate of interest is based on the validity of the
"interest-rate crow ding o u t” p h en om en on. Interestrate crow ding out is d em onstrated graphically in

6For discussions of the differences between nominal and real interest
rates and between ex ante and ex post interest rate measures, see
Santoni and Stone (1981a, 1981b and 1982); for an analysis of the
problems associated with attempts to measure real interest rates,
see Brown and Santoni (1981).
7For evidence that larger deficits per se are not associated with faster
inflation, see Protopapadakis and Siegel (1984) and Weintraub
(1981).

FEDERAL RESERVE BANK OF ST. LOUIS

figure 2, w hich depicts the d em an d for and supply of
real resou rces allocated through credit markets. The
dem and curve, labeled D, is the dem and for resources
by w ould-be borrow ers: con su m ers, private firms and,
of cou rse, the governm ent. The supply cuive, labeled
S, rep resents the am ou nt of cu rren t saving that w ouldbe lenders (savers) are willing to provide. The price
that influences th ese borrow ing and lending decisions
separately, and that is determ in ed jointly by the inter
actions of all borrow ers and lenders, is the real rate of
interest.

NOVEMBER 1985

F ig u re 2

Com parison

o f D e f i c i t ' s I m p a c t o n In t e r e s t R a t e s

Using figure 2, it is easy to show how a larger deficit
could increase the real rate of interest. If all oth er
things rem ain the sam e, a larger deficit increases the
dem an d in the credit m arket to D' and results in a
higher price of credit as the real rate of interest rises to
r,. The additional resou rces that the governm ent ob
tains com e partly from additional saving (for exam ple,
people red u ce their consum ption) and partly from
reductions in non-governm ent borrow ing (for exam
ple, private investm ent declines). The larger deficit
and the resulting higher real interest rate have thus
crow ded out, or red u ced , private secto r consum ption
and investment.
Of course, the extent to w hich the real rate of inter
est actually in creases u n d er the conventional view
d epends on the specific slopes of the dem and and
supply cuives; the flatter are the dem and and supply
curves, the sm aller the rise in the real rate of interest.

An Alternative Theoretical View o f the
Impact o f Deficits
Interest-rate crow ding out, as d ep icted in figure 2, is
predicated on the view that increases in the deficit per
se have little effect on the supply of o r the dem and for
credit bv co n su m ers and private firms. Instead, co n
sum ers and private firms respon d by moving along
their u nchanged dem an d and supply cuives in re
sponse to changes in the real rate of interest p rodu ced
by the increased governm ent deficit.
An alternative view of how people respon d to
changes in governm ent deficits suggests that the real
rate of interest is essentially unaffected by governm ent
deficits." This view states that people see deficits as

T h is view has been popularized by Barro (1974) and Seater (1982),
among others. For a recent discussion of the conventional and
alternative theoretical relationships between deficits and interest
rates, see Rasche (1985) and Tatom (1985); for recent evidence
supporting this alternative view, see Kormendi (1983) and Protopapadakis and Siegel (1984).

implied taxes that eventually m ust be im posed on
their future incom es to repay the larger governm ent
obligations. Thus, larger deficits today are equated
with larger future taxes an d red u ced future after-tax
incom es. As a result, an increase in the deficit red u ces
peop le’s perm anent incom es; this, in turn, red u ces
the private land, thus, the total) dem an d for credit (to
D), while increasing private saving and, thus, the su p
ply of credit (to S). As show n in figure 2, although defi
cits crow d out private investm ent an d consum ption,
they have no appreciable im pact on the real rate of
in te re st, w h ich re m a in s u n ch a n g e d at r„. The
crow ding-out is d irect; it does not take p lace through
increased real interest rates.

Deficits and Expected Real Interest
Rates: Some Casual Evidence
The conventional view suggests that, o th er things
the sam e, larger deficits are associated w ith higher
exp ected real interest rates; the alternative view sug
gests that they are not.
Chart 2 displays the behavior of one adjusted deficit
m easure and one m easure of the e x p ected real inter
est rate that, accord in g to conventional theory, is in
fluenced by federal deficits. The deficit m easure used
is the real cyclically adjusted deficit div ided by p oten

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Chart 2

Deficit and Ex Ante Real Interest Rate

1973

tial real gross national p ro d u ct (GNP).!I The exp ected
real interest rate m easure is obtained by subtracting
six-m onth inflation forecasts from six-m onth interest
rates at the time the inflation forecasts w ere m ade."1
An exam ination of ch art 2 provides som e evidence
that the real interest rate does not respon d to changes
in the federal deficit in the w av that is generally ex
p ected . For exam ple, average e,v ante real interest rates
w ere m u ch higher in 1 9 7 3 -7 4 than they w ere in 1 9 7 5 77, even though the federal deficit m easure was about

9For discussion of the rationale and use of cyclically adjusted deficit
measures, see Tatom (1984); for discussion of the impact of recent
recessions on deficits, see Malabre (1985).
1 The ex ante real interest rate series was constructed by the following
0
method: six-month-ahead inflation forecasts for the consumer price
index (CPI) were derived from the Livingston survey data. These
expected inflation figures were then subtracted from the quarterly
averages for the six-month Treasury bill rate for the quarters in
which the surveys were taken. For more details on this method, see
Holland (1984).

1985

twice as high in the later y ears than it w as in the earlier
years. Similarly, the exp ected real rate rose sp e cta cu
larly from early 1980 to early 1982 w hen the deficit
m easure w as virtually u n ch an ged ; since then, the real
rate has declined considerably, vet the deficit has
clim bed substantially.
Chart 3 sum m arizes the relationship betw een the
deficit and the real interest rate in an alternative fash
ion. It is a scatter-d iagram of the associated ch an ges in
the deficit and
ante real interest rate m easures. If
increases (decreases) in the deficit generally w ere as
sociated with increases (decreases) in real interest
rates, then the vast m ajority of the associated pairs of
deficit-interest rate ch an ges w ould be in the first (1)
and third (III I quadrants of the ch art. As a perusal of
ch art 3 indicates, however, the points are scattered
fairly random ly with half of them in the “w rong" parts
of the chart.
The solid line, labeled R,, is the regression line

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

C hart 3

Changes in Deficit and Ex Ante Real Interest Rate
A Ex ante re al interest rate
Perc ent

A Ex ante re a l interest rate
Perc ent

A (C y c l i c a l l y adj ust ed d e f i c i t / po t e n t i al GN P , as a percent)

showing the estim ated linear relationship betw een
changes in the deficit m easure and changes in the
exp ected real rate of interest. The conventional theoiy
suggests that the line should slope upw ard from left to
right in ch art 3; in fact, it does not. The slope, however,
is not statistically significant. Thus, a simple analysis
suggests that ch an ges in the deficit have no significant
effect on m ovem ents in the real rate of interest.

Deficits and Interest Rates: Some
Econometric Evidence
Charts 2 and 3 are not intended to d em onstrate that
deficits have no effects on real interest rates; they do
show that there is no easily discernible relationship
betw een them . B ecause a host of o th er influences
could have confounding effects on su ch a simple

com parison, m ore detailed eco n om etric analysis is
required to d ecip h er the im p act of deficits on interest
rates.
Unfortunately, su ch analyses generally have not
been able to isolate the effects of deficits on real
interest rates or draw any firm co n clu sion s. Table 1,
w hich con tain s a su m m aiy of su ch studies, shows
evidence that is highly am biguous." While som e stu d
ies found positive im p acts of deficits on interest rates,
o th er studies found m ixed o r even negative effects;
while the effects w ere statistically significant in som e
studies, they w ere insignificant in oth ers. A nother
sum m aiy of su ch studies rep orted sim ilar findings.

"See Congress (1984), p. 100.
,2U.S. Department of the Treasury (1984).

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Table 1
Federal Deficits and Interest Rates: Some Empirical Findings
Dependent
Variable

Data
Frequency

Statistical
Technique

Deficit
Variable

Short
Long
Federal
Sign of
term
term
Debt
Interest Interest Federal Privately
Deficit Statistically
Other Variable Significant
Rate
Rate
Deficit
Held

Author
(Year of Study)

Time
Period

Bradley (1983)

1966-1982

Monthly

IV»

Canto and Rapp
(1982)

1929-1980

Annual

Granger-Sims
Tests

Mixedc

Carlson (1983)

1952-1983

Quarterly

OLSc

DeLeeuw and
Holloway (1983)

1956-1983

Annual

OLS

Dewald (1983)

1953-1981

Quarterly

OLS

Dwyer (1982)

1952-1978

Quarterly

VAR"

Evans (1983)

1979-1983“

Monthly

OLS

Fackler and McMillin
(1983)

1963-1979

Quarterly

VAR

Feldstein and
Eckstein (1970)

1954-1969

Quarterly

OLS

Feldstein and
Chamberlain
(1973)

1954-1971

Quarterly

OLS

Frankel (1983)

1954-1980

Annual

TheilGoldberger
mixed
estimation

Yes
Yes«

Positive

Mixed” 1

Mixed

Negative

Mixed'

Zero

Positive

Yesk

Negative

Positive
Positive'

Mixed

Hoelscher (1983a)

1952-1976

Quarterly

Hoelscher (1983b)

1953-1982

Annual

IV&OLS

Kudlow (1981)

1958-1980

Annual

OLS

Makin (1983)

1959-1981

Quarterly

OLS

Makin and Tanzi
(1983)

1960-1981

Quarterly

Mascaro and Meltzer
(1983)

1969-1981

Quarterly

OLS

Miller (1982)

1948-1982

Annual

VAR

Motley (1983)

1958-1982

Monthly

OLS

X
X

Positive

Mixed

Mixed'

Positive

Yes

1954-1978

Quarterly

VAR

Sinai and Rathjens
(1983)

1960-1982

Quarterly

OLS

No"

Positive

Mixed

X
X

Mixed"1
Yes

Mixed

Positive
Mixed

Plosser (1982)

•Also 3 major war periods.
instrumental variables.
'Ordinary least squares.
"Vector autoregressions.
'Some positive, some negative.
'Negative in first-difference form.
sNo in first-difference form.
"Some significant, some insignificant.
No short rate, Yes long rate.

Zero

Mixed0

Mixed
X

Nop

Mixed

'No short rate, Yes long rate.
kMixed signs and significance in flow models.
'No and negative sign for short rate, Yes and positive sign for
long rate.
m
Not significant in one case, marginally in other.
"Yes and negative for short rate for 1969-1979.
°No for most recent period.
p instances of significance at 10% level.
3

SOURCE: The Economic Outlook (February 1984), Congressional Budget Office. Simulation references were deleted.

FEDERAL RESERVE BANK OF ST. LOUIS

Thus, it app ears that eco n o m etric studies provide
only weak evidence to support the view that federal
deficits have a significant influence on interest rates."

LINK # 2 : WOULD LO W ER U.S. REAL
IN T ER EST RATES RED UCE TH E
FOREIGN EXCHANGE VALUE O F TH E
DOLLAR?
Most farm com m odities trad ed in international
m arkets are priced in U.S. dollars regardless of w here
they are p ro d u ced . Consequently, a set of events that
raised the value of the dollar in term s of Brazilian
cruzeiros, for exam ple, w ould make Brazilian soy
beans less expensive than U.S. soybeans. Nations that
im port soybeans cou ld use their dollars to p u rch ase
cruzeiros and, hence, p u rch ase Brazilian soybeans
m ore cheaply than before. B ecause of this relation
ship, changes in farm exports are linked to changes in
the value of the dollar.
While we typically think of the value of the dollar visa-vis one or an o th er specific cou n try's cu rren cy — for
exam ple, the Jap an ese yen, the Fren ch franc o r the
West German mark — su ch bilateral exchange rates bv
them selves, do not provide a clear picture of w hat is
happening to the overall value of the dollar in foreign
exchange m arkets. Instead, an /nde,v of the dollar’s
value often is used to incorporate information about
the m ovem ent of the dollar relative to o th er m ajor
cu rrencies. One index, called the trade-w eighted ex
change rate of the U.S. dollar, enables us to determ ine
what is happening to the dollar's value relative to the
cu rren cies of o u r m ajor trading p artners."
The foreign exch ange value of the dollar is the
relative price of the U.S. dollar in term s of o th er n a
tion s’ cu rren cies. The actu al value of the dollar at any
time is d eterm ined by the factors that underlie the
dem and for and supply of dollars in foreign exchange
m arkets.
There currently is som e controversy over w hich
factors determ ine exch an ge rates and the relative in
fluences they have on exchange rate m ovem ents at

1 For a detailed discussion of the major problems associated with
3
empirical estimation of the deficit's impact on interest rates, see
Congress (1985a), pp. 77-84.
,4The trade-weighted exchange rate used in this study is the Federal
Reserve Board index (March 1973 = 100) of the weighted-average
exchange value of the U.S. dollar against the currencies of other G10 countries plus Switzerland. Weights are the 1972-76 average
total trade shares of each of the 10 countries.

NOVEMBER 1985

any p articu lar m om ent.' There is, however, a fairly
general analytical framework that suggests four fac
tors as the main influences on the behavior of ex
change rates: (1) differences in inflation rates betw een
countries; (21 differences in real interest rates betw een
countries; (3) differences in real eco n o m ic conditions
that influence trade p attern s; and (41 differences in
political and o th e r risks associated w ith investm ent in
specific countries."' We focus on the effects of changes
in the first two factors on exch an ge rate m ovem ents.
Unfortunately, ch an ges in the rem aining two factors
can make it difficult to d ecip h er the actu al im p acts of
changes in inflation an d real interest rate differentials
on the exch an ge rate at any given m om ent.

Adjusting the Exchange Rate fo r
Differences in Inflation Across Nations
Theoretical co n sid eration s suggest that ch an ges in
bilateral and trade-w eighted foreign exch an ge values
of the dollar are inversely related to differences be
tw een U.S. and foreign inflation rates. If this inflation
differential (U.S. m inus foreign) is positive, the value of
the dollar will decline over tim e; if the inflation differ
ential is negative, the dollar’s foreign exch an ge value
will rise.
This relationship, called pu rch asin g p ow er parity, is
based on the notion that sim ilar goods traded in world
markets m ust co m m an d sim ilar prices, regardless of
w here they are bought and sold. F or exam ple, if a
bushel of corn co sts $1.50 in the LInited States and C3
in the United Kingdom, an exch an ge rate of £ 2 per
dollar w ould "equalize” the price of U.S. an d U.K. corn
to all p u rch asers. If inflation in the United States drove
the price of corn to S3 p er bushel, then, o th er things
the sam e, the exchange rate w ould have to fall to £ 1
per dollar to bring the price of U.S. co rn back in line
with U.K. corn in w orld m arkets.
Of course, if changes in the value of the dollar were
simply the result of ch an ges in these inflation differen-

,5For example, one analyst has noted that “there is no consensus on
how exchange rates are determined. The interpretations vary widely
among the various theories, ranging from the traditional approach of
trade-oriented demand and supply factors, to the modern approach
of asset-market mechanism and expectations. The analysis of cur
rency determination is complicated by the interdependence of the
exchange rates, monetary and other economic policies, and factors
affecting economic and financial performance.” Poniachek (1983),
p. 2.3.3.
"This discussion is based on the framework developed in Isard
(1980).

FEDERAL RESERVE BANK OF ST. LOUIS

tials, exchange rate m ovem ents w ould be neutral with
respect to trade p atterns. Indeed, o th er things u n
changed, exchange rate m ovem ents con sisten t with
purchasing pow er parity will preserve cu rren t trade
patterns.
Exchange rates are affected bv o th er factors, how
ever, so that their m ovem ents are not con sisten t solely
with purchasing pow er parity conditions. If exchange
rates rise m ore (or fall less) than inflation differentials
w arrant, prices of U.S. goods will rise relative to similar
goods sold by o th er cou n tries; if exchange rates rise
less (or fall more) than inflation differentials w arrant,
prices of U.S. goods will fall relative to foreignp rodu ced goods.
This discussion suggests that, if we want to assess
the effect of exchange rate m ovem ents on exp orts in
general, and farm p ro d u cts in particular, w e should
look at the m ovem ent in exch an ge rates after adjusting
for the effects of inflation differentials. One su ch ex
change rate m easure is called the real trade-w eighted
exchange rate for the U.S. dollar.'7

The Impact o f Real Interest Rate
Differences on the Real Exchange Rate
Theoretical con sid eration s suggest that changes in
the real trade-w eighted exchange rate should be posi
tively related to ch anges in the real interest rate differ
ential (U.S. m inus foreign). If U.S. real interest rates rise
relative to foreign real rates, o th er things the sam e, the
real trade-w eighted value of the dollar should rise; if
U.S. real interest rates fall relative to foreign real rates,
the real trade-w eighted value of the dollar should
decline. The presum ption is that a positive real rate
differential will attract foreign capital, while a negative
differential will make investm ent abroad m ore a ttra c
tive. Thus, ch an ges in the real rate differential should
cau se similar ch an ges in the real trade-w eighted ex
change rate.

Changes in Real Interest Rate
Differentials and the Real Exchange
Rate: Casual Evidence

NOVEMBER 1985

pected real interest rate differential ILLS, m inus foreign
exp ected real interest rates) from 1973 to the p resen t.1
8
These data suggest that the link betw een the real
interest rate and the real exch an ge rate is not esp e
cially reliable. For exam ple, average real interest rate
differentials w ere approxim ately the sam e in the 1 9 7 5 78 and 1 9 8 2 -8 5 periods, vet the real exch an ge rate was
falling in the form er period and rising in the latter one.
Chart 5 show s a som ew hat different w av of looking
at the relationship betw een m ovem ents in the real
interest differential and m ovem ents in the real ex
change rate. It is a scatter-d iagram of changes in the
real interest rate differential and the associated per
cen t changes in the real exch an ge rate. O ther things
unchanged, eco n om ic theoiy predicts that the points
should lie predom inantly in the first (I) and third (III)
quadrants; positive (negative) ch an ges in the real in
terest rate differential should be associated with posi
tive (negative) p ercen t ch an ges in the real exchange
rate. This is not the case: the data points lie mainly in
quadrants II and IV.
The line labeled R, is the regression line relating the
p ercen t ch an ges in tire real trade-w eighted exchange
rate associated with the ch an ges in the exp ected real
interest rate differential. It should slope upw ard from
left to right; instead, it slopes dow nw ard, suggesting
that an increase (decrease) in the real interest rate
differential is associated with a d ecrease (increase) in
the real exch an ge rate. This estim ated inverse rela
tionship, however, is not a statistically significant one;
that is, the claim that there is no sim ple linear relation
ship betw een these variables can n ot be rejected at
standard statistical significance levels. This puzzling
result again suggests that d eciphering the effect of
changes in real interest rate differ entials on exchange
rate m ovem ents requires detailed and careful e co n o
m etric analysis.

Some Econometric Eiidence
Em pirical studies of real exch an ge rates and real
interest differentials offer a som ew hat qualified view of
their relationship. F or exam ple, one recent investiga
tion of the issue found a small statistically significant
lagged respon se of the real exch an ge rate to the real

Chart 4 show s w hat has h app ened to the real tradew eighted exchange rate and one m easure of the ex-

"The real trade-weighted exchange rate is the nominal tradeweighted exchange rate described earlier (see footnote 14) divided
by the ratio of the U.S. consumer price index (CPI) to the foreign,
trade-weighted CPI, each indexed to March 1973.

1 The ex ante real interest rate differential is obtained from the U.S.
8
three- and four-month money market interest rate minus the tradeweighted average three- and four-month money market rates for six
industrialized countries adjusted by corresponding Organization for
Economic Cooperation and Development (OECD) inflation
forecasts.

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Chart 4

Real Trade-weighted Exchange Rate and Real Interest
Rate Differential

1973

interest rate differential.'" Specifically, the study foui
that a 10-basis-point change in the U.S. m inus foreij
real interest differential w ould cause, after two qua
ters, a 0.23 p ercen t rise in the real value of the dollar.
This study also found no independent effect of deficits
on the real exchan ge rate.-' In general, it ap p ears that
w e know very little about the extent to w hich real
interest rate differentials actually affect real exchange
rates.

,9Batten and Belongia (forthcoming 1986).
“ The estimated coefficients from this type of statistical study are
strictly valid only for small changes in variables. Therefore, the
example presented should not be expanded to conjecture, for
example, that a 100-basis-point change in the interest differential
would cause a 2.3 percent change in the dollar’s real value.
2 Similar results were found by Bisignano (1985).
1

1985

LINK # 3 : WOULD A LO W ER VALUE O F
TH E DOLLAR INCREASE U.S. FARM
EXPO RTS?
Farm ers and legislators w ould like to increase the
real value of U.S. farm exp orts. W ould low er exchange
rates result in a significant in crease in real farm
exports?
We d iscu ssed earlier how exp orts co u ld be affected
bv ch an ges in the exch an ge rate. Purchasing pow er
parity conditions suggest that m ovem ents in ex
change rates should exactly offset ch an ges in the price
of the sam e com m od ity in different cou n tries follow
ing som e adjustm ent period. F or exam ple, the price of
corn should be the sam e acro ss cou n tries after ad ju st
m ents are m ade for exch an ge rate differences and
costs of transportation.

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

C hart 5

G ro w th of Real T ra d e -w e ig h te d Exchange Rate a n d C h a n g e
in Real Interest Rate D iffe re n tia l
7.ARTWEX

%ARTWEX

A Ex ant e i n t e r e s t d i f f e r e n t i a l

There is substantial evidence, however, that p u r
ch asin g pow er parity does not necessarily hold in the
sh ort-ru n and that a considerable period of time,
p erhaps as long as five to 10 years, m ay be required
before it finally is reach ed . If this is the case, deviations
from p u rch asin g pow er parity, ch a racterized by
changes in the real exch an ge rate, m ay have persistent
and significant effects on real farm exports.

Changes in the Real Exchange Rate and
Real Farm Exports: Casual Evidence
Chart 6 displays the behavior of the real exchange
rate and real farm exp orts since 1973. Depending

upon the specific y ears ch osen , a perusal of the ch art
yields both confirm ing and co n tra d icto iy evidence for
the presu m ed inverse relationship betw een m ove
m ents in the exch an ge rate and farm exports. For
exam ple, exchange rates and farm exp orts moved in
opposite directions from 1976 to the first half of 1979,
in 1982, and from the secon d half of 1984 to the first
half of 1985. However, exch an ge rates and farm exports
moved generally in the sam e direction from 1973 to
the first half of 197t> and from 1979 to 1980; m oreover,
farm exports rem ained virtually u n ch an ged from 1980
to the first half of 1982, and from the secon d half of
1982 through 1984, two periods w hen exch an ge rates
w ere rising dram atically.

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

C hart 6

Real Trade-weighted Exchange Rate and Real Farm Exports

Chart 7 displays a scatter-d iagram of ch an ges in the
real exchange rate and associated ch an ges in real farm
exp orts since 1973. O ther things unchanged, e co
nom ic theoiv predicts that the points should lie p re
dom inantly in the secon d (III and fourth (IV) quad
rants; positive (negative) changes in the real exchange
rate should be associated with negative (positive)
changes in real farm exports. This, however, is not the
case: the data points are random ly scattered through
out the four quadrants and nearly half of them lie in
the w rong ones.
The line labeled R, is the regression line relating the
p ercent changes in real farm exports associated with
the p ercen t changes in the real exch an ge rate. It
should slope dow nw ard from left to right and it does.
The negative slope, however, is not statistically signifi
can t. Thus, the possibility that there is no co n tem p o
ran eous relationship betw een ch an ges in the ex
change rate and farm exp orts can n ot be rejected.

Some Econometric Evidence
Em pirical studies have show n that changes in the
real exch an ge rate do affect im ports and exp orts over a
considerable tim e period. W hen these longer-run ef
fects are taken into acco u n t, m ovem ents in the real
exchange rate have the exp ected effects on im ports
and exports. A su m m ary of selected studies exam ining
the long-run im p act of ch an ges in the real exchange
rate on the d em an d for U.S. m erch an d ise exp orts and
im ports is show n in table 2.~ M erchandise exports
consist of all p rodu cts, including farm products, ex
ported to the rest of the w orld; the "long-run price
elasticity of export dem an d is the total percentage
change in export volum e in respon se to a sustained 1
p ercen t ch an ge in the relative price of U.S. exp orts to

2 See Congress (1985b), p. 49.
2

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Chart 7

Growth of Real Trade-weighted Exchange Rate and
Real Farm Exports

7.ARTWEX

foreigners, after it has had time to adjust fully.”--1The
elasticities are all negative as exp ected . Although the
estim ated elasticities range from —0.3 to —2.3, the
m ore recen t ones run close to — 1, indicating that a 1
percen t drop in the real exch an ge rate will, after suf
ficient time passes, ind u ce a 1 percent rise in total
m erchandise exports.
Two recent studies focused specifically on the effect
of changes in the exch an ge rate on agricultural ex
ports.’ 1 After estim ating a simple quarterly red uced form equation for the real value of farm exports, they
find that a 1 p ercen t fall in the real value of the dollar

will increase the real value of farm exp orts bv 0.7
p ercen t w ithin one and o n e-q u arter years.-' Thus,
unlike the previous two links, the third link in the
chain running from deficits to farm exp orts has both
theoretical and em pirical support.

SUMMARY
There is a widely sh ared view that federal deficits
have contributed significantly to higher nom inal and
real interest rates in the United States. M oreover, it is
com m only believed that these higher rates have co n
tributed significantly to the rising foreign exchange

“ See Congress (1985b), p. 48.
2 Batten and Belongia (1984, forthcoming 1986).
4

“ Batten and Belongia (forthcoming 1986).

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Table 2
Long-Run Price Elasticities of Demand for U.S. Merchandise Exports and Imports
Exports
Study or Model
Adams et al.
Houthakker-Magee
Basevi
Hickman-Lau
Samuelson
Stern et al.
Goldstein-Khan
Gylfason
Amano et al.
DRI Model
Helkie
Wharton Model
Average

Imports

Year of
Estimate

Elasticity

1969
1969
1973
1973
1973
1976
1978
1978
1981
1982
1983
1984

-0 .6 0
-1 .5 1
-1 .4 4
-1 .3 8
—1.13
-1 .4 1
-2 .3 2
-0 .6 2
-0 .3 2
-0 .8 3
-0 .9 0
-0 .9 8

Adams et al.
Houthakker-Magee
Armington
Taplin
Beenstock-Minford
Stern et al.
Gylfason
Geraci-Prewo
Goldstein-Khan
DRI Model
Helkie
Wharton Model

-1 .1 2

Average

Study or Model

Year of
Estimate

Elasticity

1969
1969
1970
1973
1976
1976
1978
1980
1980
1982
1983
1984

-1 .1 6
-1 .0 3
-1 .7 3
-1 .0 5
-1 .0 4
-1 .6 6
- 1 .1 2
-1 .2 3
-1 .1 2
-0 .5 6
-0 .8 5
-0 .6 4
—1.10

SOURCE: The Economic and Budget Outlook: An Update (February 1985), Congressional Budget Office.

value of the dollar. Thus, if frequently is argued that
ou r n atio n ’s exporting sectors, p rod u cers of farm
com m odities in particular, will continue to suffer until
federal deficits are red u ced an d U.S. interest rates are
brought down.
In this article, w e exam ined three vital links in the
conventional argum ent that ties the deficit to farm
exports. With resp ect to the first link, w e noted that
there is considerable theoretical controversy over
w heth er larger deficits actually cau se real interest
rates to increase. We found little em pirical evidence to
support this view.
Second, w e noted that, even if low er deficits did
result in low er U.S. real interest rates, they w ould not
necessarily have a salutary im p act on the real ex
change rate. Apparently, o th er influences on the real
exchange rate have offset the effect, if anv, of changes
in real interest rate differentials in recent years.
Among these o th er factors m ay be "the strong perfor
m an ce of the U.S. econom y, confidence in the strength
and stability of the political system in the United
States, capital flight from debtor countries, [and] a
substantial shift in the external position of Am erican
banks. ” 6 The im portant point is that there is little
2

2 Pohl (1985). Similar comments have been made by a wide variety of
6
commentators: e.g., “At various times, other factors, which are
difficult to measure, have also influenced the d o lla r... The reversal

em pirical evidence to show that ch an ges in the real
interest rate differential have had a significant im pact
on m ovem ents in the real exch an ge rate during the
past 13 years.
Finally, w e show ed that, although U.S. farm exp orts
are inversely related to the real exch an ge value of the
dollar, the dem an d relationship is inelastic and ex
change rate m ovem ents have th eir full effect only over
a considerable tim e period. However, even though
low er exch an ge rates would, over time, increase U.S.
farm exports, the failure of the first two links to be
supported suggests that w e can n o t necessarily expect
that low er deficits will result in a low er value of the
dollar in foreign exch an ge markets.-7
None of the d iscussion above should be taken as
evidence that deficits p er se are eith er good or h arm
less. Nor does it prove that larger deficits have had no
adverse effect on real interest rates, on the foreign
exchange value of the dollar o r on farm exports. Unfor
tunately, at the present tim e, there con tin u es to be

in the dollar's fortunes since late 1980 may [be] related to (i) the
election of a new administration committed to a more conservative
approach to financial policies; and (ii) the increased risks associated
with other currencies." Atkinson et al. (1985), pp. 37 and 39.
2 See Poole (1985) for a discussion of why lower budget deficits might
7
be expected to raise the value of the dollar.

FEDERAL RESERVE BANK OF ST. LOUIS

considerable u ncertainty about the effects that larger
deficits actually have had on these key econom ic
variables.-"

NOVEMBER 1985
Isard, Peter. “ Factors Determining Exchange Rates: the Roles of
Relative Price Levels, Balances of Payments, Interest Rates and
Risk," Federal Reserve Board, International Finance Discussion
Papers, Number 171 (December 1980).
Kormendi, Roger C. “ Government Debt, Government Spending
and Private Sector Behavior," American Economic Review (De
cember 1983), pp. 994-1010.

R EFER E N C ES

Kraft, Joseph. “ Weird Economics, Bizarre Politics,” Washington
Post, October 27, 1985.

Atkinson, P., A. Blundell-Wignall, J., C. Chouragui, and G. Hacche.
Exchange Rate Management and the Conduct of Monetary Policy,
OECD Monetary Studies Series, OECD (1985).

Malabre, Alfred L., Jr. "Whither the Deficit When Recession Ar
rives,” Wall Street Journal, September 20,1985.

Barro, Robert J. “Are Government Bonds Net Wealth?'’ Journal of
Political Economy (November/December 1974), pp. 1095-117.

Modigliani, Franco. “And Why the Deficit Must Be Slashed,” New
York Times, November 3, 1985.

Batten, Dallas S., and Michael T. Belongia. “ The Recent Decline in
Agricultural Exports: Is the Exchange Rate the Culprit?” this
Review (October 1984), pp. 5-14.

Pohl, Karl Otto. Address to the International Industrial Conference
(September 19,1985), reproduced in Bank for International Settle
ments, Press Review (September 20, 1985).

________ . “ Monetary Policy, Real Exchange Rates and U.S. Agri
cultural Exports," American Journal of Agricultural Economics,
forthcoming (May 1986).

Poniachek, Harvey, A. “ The Determination of Exchange Rates,” in
Abraham M. George and Ian H. Giddy, eds., International Finance
Handbook, Volume 1 (Wiley, 1983), pp. 2.3.3-2.3.44.

Bisignano, Joseph. “ Fiscal Deficits and Exchange Rates: A Look at
Recent Policy Assertions and Their Theoretical and Empirical
Support," Working Paper 85-04, Federal Reserve Bank of San
Francisco (June 1985).

Poole, William. “ Summary Comments,” speech delivered at the
Federal Reserve Bank of Kansas City Conference, Jackson Hole,
Wyoming, August 23, 1985.

Brown, W. W., and G. J. Santoni. "Unreal Estimates of the Real
Rate of Interest," this Review (January 1981), pp. 18-26.
Clark, Lindley H., Jr. “ On the Beach In Bermuda With a Book," Wall
Street Journal, October 29, 1985.
Congressional Budget Office. The Economic Outlook, Congress of
the United States (Government Printing Office, February 1984).

Protopapadakis, Aris A., and Jeremy J. Siegel. “ Government Debt,
the Money Supply, and Inflation: Theory and Evidence for Seven
Industrialized Economies," Working Paper No. 84-4, Federal Re
serve Bank of Philadelphia (August 1984).
Rasche, Robert H. “ Views on Deficits and Interest Rates,” Federal
Resen/e Bank of San Francisco Weekly Letter (April 19, 1985).

________ _ The Economic and Budget Outlook: Fiscal Years 19861990 (GPO, February 1985a).

Santoni, G. J., and Courtenay C. Stone. "Navigating Through the
Interest Rate Morass: Some Basic Principles,” this Review (March
1981a), pp. 11-18.

_________ The Economic and Budget Outlook: An Update (GPO,
August 1985b).

_________ “What Really Happened to Interest Rates? A Longer
Run Analysis,” this Review (November 1981b), pp. 3-14.

Dobson, William D. “ Effects of the Macroeconomic Environment on
1985 Farm Legislation," in United States Department of Agricul
ture, Outlook 1985; proceedings of the Agricultural Outlook Confer
ence, December 3-5, 1984, pp. 48-58.

________ _ “The Fed and the Real Rate of Interest,” this Review
(December 1982), pp. 8-18.

Downs, Anthony. “ This Building Boom Shows Something's
Busted,’’ Wall Street Journal, October 29,1985.
Duncan, Marvin, and Mark Drabenstott.
York Times, August 16, 1985.

“ Economic Scene," New

Holland, Steven, A. “ Real Interest Rates: What Accounts for Their
Recent Rise?" this Review (December 1984), pp. 18-29.

Schuh, G. Edward. “ Future Directions for Food and Agricultural
Trade Policy,” American Journal of Agricultural Economics (May
1984), pp. 242-47.
Seater, John J. "Are Future Taxes Discounted?” Journal of Money,
Credit and Banking (August 1982), pp. 376-89.
Tatom, John A. “ A Perspective on the Federal Deficit Problem," this
Review (June/July 1984), pp. 5-17.
________ _ “ Two Views of the Effects of Government Budget Defi
cits in the 1980s,” this Review (October 1985).
U.S. Department of the Treasury. “The Effects of Deficits on Prices
of Financial Assets: Theory and Evidence," (GPO, 1984).

28For additional discussion and empirical evidence supporting this
conclusion, see Bisignano (1985) and Wallis (1985). Even Dobson,
whose diagram of the deficit's effects on the economy appears in
figure 1, notes that “ how much the federal deficits influence the
variables in the diagram is not known with much certainty.” Dobson,
(1984), p. 49.

Wallis, Allen. “ On Deficits and Interest Rates," Washington Post,
November 4, 1985.
Weintraub, Robert E. “ Deficits: Their Impact on Inflation and
Growth,” Staff Study for the Subcommittee on Monetary and Fiscal
Policy of the Joint Economic Committee (GPO, July 30,1981).

Monthly Economic Indicators: A
Closer Look at the Coincident
Index
Keith M. Carlson

ECISIONS relating to m onetaiy policy are based
on a considerable volum e of eco n om ic information.
One of these p ieces of inform ation is the cu rrent
status of eco n om ic activity. Recent eco n om ic perfor
m an ce is a vital foundation required in the p ro cess of
deciding w hat co u rse m on etary policy should take.
There are m any eco n o m ic indicators released each
m onth, and one problem for the m on etaiy policy
maker, as well as for businesses, con su m ers and gov
ernm ents, is to distill from this spate of inform ation
som e assessm en t of just how well the econom y is
performing.
The U.S. D epartm ent of C om m erce's B usiness Con
ditions Digest lists 84 different eco n om ic time series as
m onthly cyclical indicators. An analysis of all, or even
a substantial subset, of these indicators could provide
a confusing picture to even the m ost astu te analyst.
Fortunately, w ading through su ch a m orass is not
n ecessary to determ ine how the econ om y is perform
ing. The C om m erce D epartm ent's Bureau of E co
nom ic Analysis (BEA) has simplified the p ro cess bv
publishing a "com p osite index of four roughly co in ci
dent in d icato rs,” w hich co n d en ses the information
from the m ost im portant m onthly indicators into one
sum m aiy index.
B ecause it is overshadow ed by the sim ultaneous
release of the m ore popular “index of twelve leading
indicators,” the coin cid en t index does not receive
extensive m edia coverage. Yet the coinciden t index
provides valuable an d reliable inform ation. The p u r
pose of this article is to describe the coinciden t index

Keith M. Carlson is an assistant vice president at the Federal Reserve
Bank of St. Louis. Sandra Graham provided research assistance.

Digitized for20
FRASER

and its com p onen ts, and to sum m arize their useful
ness and reliability accord in g to well-known criteria.
Particular attention will be focu sed on the cyclical
perform ance of th ese indicators.

THE COINCIDENT INDEX AND ITS
COMPONENTS: A B R IE F
D ESCRIPTION
The com p osite index of coinciden t ind icators is
published m onthly an d is co n stru cted from four
m onthly indicators p rep ared an d released by four
different governm ent agencies. A tabular su m m aiy of
these indicators is given in table 1. A brief description
of the com p on en t series and the coinciden t index
follows. (A m ore detailed discussion of these series
appears in the appendix.)

Employees on Nonagricultural Payrolls
This series com m on ly is called payroll em ploym ent;
it is prep ared by the Bureau of Labor Statistics of the
D epartm ent of Labor. Usually released on the first
Friday of the m onth, it generally covers the payroll
period including the 12th of the preced in g m onth. It is
based on a survey of business establishm ents, in co n
trast to the estim ate of total em ploym ent, w hich is
based on a survey of households.

Personal Income Less Transfer
Payments in Constant (1972) Dollars
This series is estim ated by the BEA in the p rep ara
tion of the national incom e a cco u n ts. The basic data
on personal incom e and tran sfer paym ents for the

The BCD Composite Index of Coincident Indicators1
Employees on
nonagricultural payrolls

Personal income less
transfer payments in
1972 dollars

Industrial
production

Manufacturing and trade
sales in 1972 dollars

Composite index of
coincident indicators

U.S. Department of Labor,
Bureau of Labor Statistics

U.S. Department of
Commerce, Bureau of
Economic Analysis

Federal Reserve Board
of Governors

U.S. Department of
Commerce, Bureau of the
Census and U.S. Department
of Labor, Bureau of Labor
Statistics

U.S. Department of
Commerce, Bureau of
Economic Analysis

Units

Thousands of persons

Billions of 1972 dollars

Index, 1977=100

Millions of 1972 dollars

Index, 1967 = 100

Revisions

Two months back,
and annually in June

Three months back, and
annually in July

Two to three months
back, and annually in
September

Three months back, and
twice annually, usually in
March and May

Three months back and
further as necessary

Name of release

The Employment Situation

Personal Income and
Outlays

Industrial Production
(G. 12.3)

Manufacturing and Trade
Inventories and Sales

Composite Indexes of Leading,
Coincident, and Lagging
Indicators

Date of release

1st Friday of the month
for the previous month

18th-20th of the month for
the previous month

15th—
17th of the
month for the
previous month

13th—
15th of the month for
two months earlier

30th-31 st of the month for
the previous month

Focus of release

Discussion equally divided
between results from
household survey data
and establishment survey
data

Personal income with
discussion of wage vs.
nonwage income. Also
disposable income and
saving

Total industrial
production with some
discussion of market
and industry
groupings

Brief discussion of sales,
inventories and inventoriessales ratio

Highlights discussion of
leading indicators with brief
discussion of coincident and
lagging indicators

Use in BCD

Used "as is" in
coincident index

Transfer payments
subtracted from personal
income and deflated by
personal consumption
expenditures deflator

Used "as is” in
coincident index

Sales deflated with BLS
producer and consumer
price indexes

Constructed from four series
with unequal weights and
detrending (see text)

BCD
classification
by economic
process

Employment and
unemployment

Production and income

Consumption, trade, orders
and deliveries

Composite indexes

1
Business Conditions Digest, a monthly publication prepared by the U.S. Department of Commerce, Bureau of Economic Analysis. All series are seasonally adjusted.

1985

NOVEMBER

Source agency
for basic data

FEDERAL RESERVE BANK O ST. LOUIS
F

Table 1

FEDERAL RESERVE BANK OF ST. LOUIS

previous m on th are released after the m iddle of the
cu rren t m onth. The co n stan t dollar estim ate; how
ever, is not available until the end of the cu rren t
m onth, w hen the com p osite indexes are released.

Industrial Production
This series is an index (1977 = 100) of the output of
m anufacturing and mining establishm ents and utili
ties; it is prepared by the Federal Reserve Board. The
estim ate for the previous m onth is available after the
m iddle of the cu rren t m onth. Industries covered by
the index generate about 30 p ercen t of the gross n a
tional p rodu ct (GNP).

Manufacturing and Trade Sales in
Constant (1972) Dollars
This series is a m easure of m onthly business sales; it
is prep ared by the Bureau of the C ensus of th e D epart
m ent of C om m erce. The cu rren t dollar estim ates are
released about the m iddle of the cu rren t m on th for
two m onths earlier; for exam ple, the estim ate for O cto
b er is released in m id-D ecem ber. The co n stan t dollar
estim ate, w hich is p rep ared by the BEA, is not avail
able until the end of the m onth, w hen it is released as a
part of the report on co m p o site indexes.

Composite Index o f Coincident
Indicators
The com p osite index of coin cid en t indicators is a
sum m ary m easure designed to signal changes in the
direction of eco n o m ic activity. The index m easures
the behavior of the four eco n om ic time series d e
scribed above, w hich show sim ilar timing at business
cycle turn s but rep resen t widely differing activities or
sectors of the econ om y. These four co m p o n en ts w ere
selected w ith the help of a detailed scoring system
that p laces p articu lar em phasis on cyclical timing.
The ch o ice of coin cid en t ind icators w as based up on a
com p arison of the timing ch aracteristics of the p artic
ular series w ith reference dates (business cycle tu rn
ing points) designated by the National Bureau of E co
n om ic R esearch (NBER).'
C onstruction of the com p osite index of coincident
indicators co n sists of several step s.’ First, each of the
com p onen t series is standardized bv dividing the

'For a discussion of an example of how the NBER arrives at such a
designation, see Zarnowitz and Moore (1983).
2For further detail, seethe U.S. Department of Commerce (1984), pp.
65-70, and Ratti (1985).

NOVEMBER 1985

m o n th -to-m on th p ercen t ch an ges by the long-run av
erage of those ch an ges. This prevents the m ore volatile
series from dom inating the index. A w eighted average
of these stan d ard ized ch an ges is then co m p u ted with
the “b etter perform ing” series assigned m ore weight.
These w eighted averages are then cu m u lated into an
index. The final step is to “tren d -ad ju st” this index so
that its long-term average grow th rate (since 1948)
equals the average of the tren d in its four co m p o n en ts.
This trend, w hich is sim ilar to real GNP, can be viewed
as a linear approxim ation of the average grow th rate in
eco n om ic activity.

CRITERIA USED TO EVALUATE
ECONOMIC INDICATORS
The BEA u ses a scorin g system to evaluate e co
nom ic tim e series as cyclical in d icato rs/1This system
provides inform ation about w hich series are to be
included in the com p osite index and the relative im
p ortance, o r w eight, a ttach ed to ea ch series. Seven
criteria are used to evaluate tim e series; the num erical
scores for th ese criteria are su m m arized in table 2. For
the m ost part, these scores are derived qualitatively.
W here possible, however, there is an attem p t to evalu
ate quantitatively.
The scoring system , first developed and applied by
Moore and Shiskin in 1966, w as published in 1967. It
was modified by Zarnow itz and B oschan in 1975, but
continues to be sim ilar in the m ost im portant detail.
According to Zarnow itz and Boschan:
The system disciplines and systematizes the judgment
of both reviewer and user of the indicators. It is an
effort to insure that all the important aspects of the
evaluation problem are considered in a consistent
and, to a significant extent, replicable wav.'

Economic Significance
How im portant is the eco n o m ic p ro cess o r variable
for w hich the p articu lar series stands, and w hat is the
breadth of the series coverage in representing the
activity co n ce rn e d ? Scoring for this criterion is su b jec
tive, depending prim arily on a classification of indica
tors by “type of e co n om ic p ro ce ss.” A broad h ierarch y
of three levels of e co n om ic variables w as postulated:
(1) C om prehensive outp u t and input aggregates in
real an d nom inal term s; for exam ple, real GNP is

3 complete discussion of this scoring system is found in Moore and
A
Shiskin (1967).
"U.S. Department of Commerce (1977), p. 171.

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Table 2
Series Scores
Total
All
All
Statistical Economic
Peaks Troughs turns Conformity Smoothness Timeliness adequacy significance Revisions Peaks Troughs turns
Composite
coincident
index

100

Nonagricultural
employment

100

Industrial
production

100

Real personal
income less
transfer
payments

100

Real manufacturing
and trade
sales
19

SOURCE: 1984 Handbook of Cyclical Indicators, p. 169

the only series that is scored 100.
(2) The m ajor co m p o n en ts of the com prehensive
aggregates and o th er variables to w hich causal
roles in business cycles are attributed; for exam
ple. investment, profits.
(3) Variables w hose role is primarily sym ptom atic
rath er than cau sal; for exam ple, m arginal em
ployment adjustm ents.
All of the co m p o n en ts of the coin ciden t index ap p ar
ently are classified as being included in (1) or (2). With
the excep tion of real m anufacturing and trade sales,
w hich is scored 80, all of the o th er co m p o n en ts are
scored at 90.

Statistical Adequacy
How well does the given series m easure the e co
n om ic variable o r p ro cess in question? The BEA co n
siders eight asp ects of this criterion:
(1) Quality of the reporting system — is it set up for
statistical purp oses or is it the by-product of an
adm inistrative program ?
(2) Coverage of eco n o m ic p ro cess — full en u m era
tion, probability sam ple, etc.
(3) Coverage of time period — full m onth or quarter,
one day per m onth, etc.

(41 Availability of estim ates of sam pling and rep o rt
ing errors.
(51 Frequency of revisions.
(6) Length of series.
(7) Comparability over time — breaks in the series.
(8) O ther considerations.
This detailed evaluation is co n stru cted to be primarily
quantitative rath er than qualitative.

Cyclical Timing
How consistently does the series coincide w ith su c
cessive business cycle turns? B ecau se this criterion is
crucial for timely recognition of business cycle turning
points, it is assigned the greatest weight of all the
criteria. A ccording to the BEA, this criterion has four
phases: (1) identification and dating of specific cycles
for the com p on en t series; (2) deciding on reference
dates (depending m ainly on the reference chronology
established by the NBER); (3) m atch in g the specific
cycle turning points with the reference dates; and (4)
scoring the cyclical perform an ce of the p articular
indicator.

Conformity
How regularly do m ovem ents in the specific in d ica
tor reflect the expansions an d co n tractio n s of the

FEDERAL RESERVE BANK OF ST. LOUIS

overall econ om y? For an ind icator to be useful, its
specific cycles m ust parallel business cycles. C on
formity is defined positively if the ind icato r rises d u r
ing eco n om ic expansions and declines during co n
tractions, and negatively if it m oves countercyclically.
This evaluation consid ers the n um ber of business
cycle phases and how they are m atch ed bv the cycle
m ovem ents of the particu lar indicator. Also exam ined
are false signals associated with specific cycles; that is,
m ovem ents in the indicators that do not m atch gen
eral expansions an d co n tractio n s. In addition, there is
a consideration of am plitude. F or exam ple, larger
m ovem ents than the average will tend to be m ore
identifiable, thu s contributing to the usefulness of an
indicator.

Smoothness
How prom ptly can a cyclical turn in a series be
distinguished from a directional change associated
with sh o rter m ovem ents? It is desirable that the cycli
cal m ovem ents of the ind icator are not obscu red by
relatively large and frequent irregular variations. Al
though there are o th er ways of improving sm ooth n ess
(for exam ple, by calculating a moving average), gen er
ally su ch p ro ced u res imply a loss of tim eliness.

Timeliness
How prom ptly available is the particu lar series and
how often is it reported ? Two asp ects are considered:
(1) how frequently the figures are com piled, and (2)
how prom ptly the figures are available.

NOVEMBER 1985

Depending on the total perform ance score, weights
are assigned to ea ch in d icator in the co n stru ction of
the com p osite index. F or exam ple, nonagricultural
em ploym ent receives the greatest weight (1.064) be
cau se of its 87 score, followed by industrial p roduction
(1.028), personal incom e less transfer paym ents in
con stan t (1972) dollars (1.003), and m anufacturing and
trade sales in co n stan t (1972) dollars (.905). These
weights do not differ markedly, although the dif
ference betw een the largest and the sm allest is
substantial.
Nonagricultural em ploym ent sco res high on all cri
teria, followed closely by industrial p rodu ction . A c
cording to table 2, the weight on industrial p roduction
is red u ced by its perform an ce at cyclical peaks and by
the extent to w hich it is revised. Personal incom e
receives a relatively low score, mainly b ecau se of the
revisions. Clearly, the w orst of the co m p o n en ts is
m an u factu rin g an d trad e sales, w hich generally
scores low est by ea ch criterion.
The scores for the co n stru cte d com p osite index are
som ew hat surprising. In particular, the total sco re is
low er than that of both nonagricultural em ploym ent
and industrial p rod u ction . Apparently, this reflects
the fact that the com p osite index is subject to su b stan
tial revision, since it has to be revised every tim e any
one of the co m p o n en t series is revised.

CYCLICAL PERFORM ANCE

How large are the revisions? This criteria w as added
separately to the list prep ared by Zarnow itz and
B oschan in 1975. Series that are subject to large revi
sions, especially if they involve directional change, can
be troublesom e, providing m isleading signals about
the p ace of eco n om ic activity.

The m ost im portant criterion used in the evaluation
of eco n om ic indicators is the cyclical perform an ce of
the ind icator in question. Generally, to be classified as
a coinciden t indicator, an in d icato r m ust turn on
average betw een —3 (3-m onth lead) an d + 1 (1-m onth
lag) at peaks and betw een —1 an d + 3 at troughs. The
differences at peaks an d troughs reflect the historical
distribution of timing. Specific peak and trough dates
for each of the coinciden t indicators are sum m arized
in table 3.

TH E BEA SCORING SUMMARY FO R
TH E COINCIDENT INDEX

Specific Cycles vs. Reference Dates

Revisions

The scores show n in table 2 cover the period from
1948 to 1980. The BEA’s objective w as to develop a
com posite index that red u ces the n um ber of false
signals that m ight arise if one w ere to rely on a single
indicator. The advantage of a com p osite index is that it
will sm ooth out the noise in the co m p o nen t series and
also cap tu re the different eco n om ic p ro cesses rep re
sen ted — produ ction , em ploym ent, real incom e and
real sales.
Digitized for24
FRASER

Business cycle turning points are designated by a
special com m ittee appointed by the NBER. This desig
nation usually o ccu rs several m on th s after the fact
and is based on all available inform ation at th at tim e.
No au tom atic rule is u sed in the determ ination of
these reference d ates/’

5See Zarnowitz and Moore.

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Table 3
Specific Peak and Trough Dates for Coincident Indicators
Reference peak date
July 1981

Jan 1980

Nov 1973

Dec 1969
10/69( - 2)

Apr 1960

Aug 1957

Jul 1953

1/60( - 3)

2 /5 7 (-6 )

3/53( - 2)

4/60(0)

3/57( - 5)

6/53( - 1 )

10/69( - 2)

1/60( - 3)

2 /5 7 (-6 )

7/53(0)

11/73(0)

10/69( - 2)

1/60( —3)

2 /5 7 (-6 )

3/53( - 4)

11/73(0)

NSC’

5/60( + 1)

8/57(0)

6/53( - 1 )

Composite coincident index

7/81(0)

1/80(0)

11/73(0)

Employees on nonagricultural
payrolls

7/81(0)

3/80( + 2)

10/74( + 11)

Index of industrial
production

7/81 (0)

3/80( + 2)

11/73(0)

Manufacturing and trade
sales in 1972 dollars

4/81 ( - 3 )

3/79( - 1 0 )

Personal income less
transfer payments in 1972
dollars

8/81 (+ 1 )

1/80(0)

3/70( + 3)

Reference trough date
Nov 1982

Jul 1980

Composite coincident index

12/82( +1)

7/80(0)

3/75(0)

11/70(0)

2/61(0)

4/58(0)

8/54( + 3)

Employees on nonagricultural
payrolls

12/82( +1)

7/80(0)

4/75( + 1)

11/70(0)

2/61(0)

5/58( +1)

8/54( + 3)

Index of industrial
production

12/82( +1)

7/80(0)

3/75(0)

11/70(0)

2/61(0)

4/58(0)

4/54( - 1 )

Manufacturing and trade
sales in 1972 dollars

10/82( - 1)

I
o

3/75(0)

11/70(0)

1/61( - 1 )

4/58(0)

12/53( —5)

Personal income less
transfer payments
in 1972 dollars

10/82( - 1 )

7/80(0)

3/75(0)

NSC'

12/60( - 2)

4/58(0)

4/54( - 1 )

Mar 1975

Nov 1970

Feb 1961

Apr 1958

May 1954

SOURCE: U.S. Department of Commerce, Business Conditions Digest (August 1985), p. 104.
'No specific cycle, that is, no specific turning point corresponding to the indicated reference is discernible.

T here is considerable variation in the lead-lag time
for the co m p o n en t indicators, although the range gen
erally is sm allest at cycle troughs. Personal incom e
less transfer paym ents in 1972 dollars app ears to co in
cide closest with the reference dates. Note, however,
this series had n either a peak n or a trough associated
with the 19 6 9 -7 0 recession .
M anufacturing and trade sales show s the greatest
variation around reference dates. F or only four of the
14 turning points did this series coincide. This is the
only series that always leads o r coincides with the
reference dates, however.
Nonagricultural em ploym ent has an excellent re c
ord excep t for the 1 9 7 3 -7 5 recession. This recession
was initiated by the OPEC oil shock and w as thus
unique am ong postw ar recession s. The un certain re
sponse of firms in deciding w h eth er to retain o r lay off
em ployees in the face of this recession w as no doubt

related to the difficulty of interpreting the shock.
O ther than that period, nonagricultural em ploym ent
turns n ear reference dates, although it tend s to lag
rath er than lead.
The index of industrial p rodu ction has an excellent
record around reference dates, especially the trough
dates. It is not clear, therefore, w hy the BEA scores this
ind icator so low arou n d cycle peaks, but it m ay be due
to subsequent revisions arou n d su ch dates. In recent
years, the specific cycle turning point has been within
± two m onths.
The coinciden t co m p o site index should be ex
p ected to coincide alm ost identically with the refer
en ce dates, and in recen t y ears it is very close. In
earlier cycles, however, the difference w as as m u ch as
six m onths. The explanation probably lies in the revi
sion of data b ecau se reference dates are seldom
ch an ged and are based on d ata a few m onths after the

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Table 4
One-Month Negative Changes during Expansions

Industrial
production

Real personal
income less
transfer
payments

Real mfg. &
trade sales

5
10
9
4
7
5
4
1
3

4
11
9
7
19
5
7
1
4

3
10
6
7
10
5
8
2
2

3
18
13
7
32
9
16
6
6

110

Industrial
production

Real personal
income less
transfer
payments

Real mfg. &
trade sales

No. of
months

Composite
coincident
index

Nonagricultural
employment

2/4&-11/48
11/49-7/53
6/54-8/57
5/58-4/60
3/61-12/69
12/70-11/73
4/75-1/80
8/80-7/81
12/82-12/84

10
45
39
24
106
36
58
12
25

4
13
9
7
10
6
8
2
2

Total

355

Expansion
period

Table 5
One-Month Positive Changes during Contractions
Composite
coincident
index

Contraction
period

No. of
months

12/48-10/49
8/53-5/54
9/57-4/58
5/60-2/61
1/70-11/70
12/73-3/75
2/80-7/80
8/81-11/82

11
10
8
10
11
16
6
16

2
0
0
0
1
2
0
2

2
0
0
0
4
11
2
1

2
2
0
1
1
4
1
2

4
3
1
5
6
4
1
4

4
3
0
3
3
4
1
6

Total

Nonagricultural
employment

fact. The specific cycle turning points are derived from
series as they are currently published.

False Signals
A nother question of interest is the extent to w hich a
p articular series em its false signals. Does an ind icator
suggest a rise o r fall in eco n om ic activity that is not
confirm ed bv later inform ation? To determ ine the
extent of this problem , m o n th -to -m on th chan ges in
the com p on en t indicators and the com p osite w ere
exam ined. A false signal is defined as a decline in the
series during an expansion and an increase during a
recession. Tables 4 and 5 sum m arize these results.
Digitized for 26
FRASER

Table 4 indicates that one-m on th negative ch an ges
are co m m o n during expansions, even for the co m p o s
ite index. The frequency of th ese perverse m ovem ents
ranges from 48 (or 13.5 percent) for payroll em ploy
m ent to 110 (or 31 percen t) for m anufacturing and
trade sales. The frequency of false signals also app ears
high in table 5, w hich show s positive m ovem ents of
the indicators during recession s. The com p osite in
dex perform s b etter than the individual co m p o n en ts
during recession s.
M onth-to-m onth variation is exp ected , n ecessitat
ing that som e longer-term perspective be m aintained.
As an exam ple of w hat h ap p en s w hen d ata are ana-

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Table 6
Three-Month Negative Changes during Expansions

Nonagricultural
employment

Industrial
production

Real personal
income less
transfer
payments

Expansion
period

No. of
months

Composite
coincident
index

2/48-11/48
11/49-7/53
6/54-8/57
5/58-4/60
3/61-12/69
12/70-11/73
4/75-1/80
8/80-7/81
12/82-12/84

10
45
39
24
106
36
58
12
25

0
0
1
2
0
0
2
0
0

0
0
2
0
0
0
0
0
0

0
4
2
3
0
0
0
0
0

0
0
0
1
0
0
0
0
0

0
3
1
1
0
2
1
0
0

Total

355

Real mfg. &
trade sales

Real ml
trade s

Table 7
Three-Month Positive Changes during Contractions

Nonagricultural
employment

Industrial
production

Real personal
income less
transfer
payments

Contraction
period

No. of
months

Composite
coincident
index

12/48-10/49
8/53-5/54
9/57—
4/58
5/60-2/61
1/70-11/70
12/73-3/75
2/80-7/80
8/81-11/82

11
10
8
10
11
16
6
16

0
0
0
0
0
0
0
0

0
0
0
0
0
9
0
0

0
0
0
0
0
0
0
0

0
0
0
0
2
1
0
0

0
0
0
0
0
0
0
0

Total

lyzed with som e perspective, co n sid er tables 6 and 7.
These tables are co n stru cted like tables 4 and 5, excep t
that a false signal is defined as three successive
m onths of perverse m ovem ent.

During recession s, the record is even better. The
coincident index has a perfect record , as do two of the
com p on en ts. Payroll em ploym ent is perfect excep t
that it kept clim bing during the early stages of the
unusual 1 9 7 3-75 recession.

As tables 6 an d 7 show, the frequency of false signals
drops dram atically. On a p ercen tage basis, the fre
quency is m inuscule. During expansions, real per
sonal incom e less tran sfer paym ents em itted only one
false signal in the p ostw ar period, followed closely bv
payroll em ploym ent with only two. B ecause it is a
w eighted average of all four com p on en ts, the co m p o s
ite index em itted m ore false signals than eith er payroll
em ploym ent o r real personal incom e less transfer
pavm ents.

Correlation Between Indicators

F u rth er analysis of the com p osite index an d its
com p onen ts is su m m arized in the correlation m atrix
in table 8.6 The highest co m p o n en t correlation is be6
Shown is the Pearsonian coefficient of correlation, which is simply a
measure of the closeness of association between two variables. If
there is no association, the coefficient is zero; if the relationship is
perfect, it equals 1. Also shown for comparison purposes is the
correlation of the quarterly average with real GNP.

FEDERAL RESERVE BANK OF ST. LOUIS

NOVEMBER 1985

Table 8
Correlation Matrix: Coincident Indicators (monthly data, compounded annual rates of
change)

Indicator
Employees on nonagricultural
payrolls

Employees on
nonagricultural
payrolls

Index of
industrial
production

Real personal
income less
transfer payments

Real manufacturing
and trade sales

Composite index
of coincident
indicators

1.00

Index of industrial
production

.71

1.00

Real personal income less
transfer payments

.58

1.00

Real manufacturing and
trade sales

.38

.47

.35

1.00

Composite index of
coincident indicators

.84

.88

.77

.62

1.00

Real GNP (quarterly)

.70

.81

.83

.72

.84

tween nonagricultural em ploym ent and industrial
production, an d these two series also correlate m ost
highly with the com p osite index. Clearly, the m an u
facturing and trad e sales series perform s least satisfac
torily. W hen quarterly averages of the com p on en t in
d icators are co m p ared with GNP, real personal
incom e less transfer paym ents correlates m ost highly,
followed closely bv industrial produ ction .

SUMMARY
M any e co n om ic ind icators are potential can d id ates
for inclusion in a list of sensitive m easures of the
econom y's cyclical m ovem ents. To simplify the p ro
cess of selection, th e Bureau of E co n o m ic Analysis of
the U.S. D epartm ent of C om m erce has co n d en sed the
list to four key indicators and, in turn, has co n stru cted
a single com p osite index of coin cid en t indicators.
This article describes the com p osite index and its four
com p onen t indicators, with special focus on their
cyclical perform ance.
The four co m p o nen ts of the coinciden t index are
industrial produ ction, nonagricultural em ploym ent,
real personal incom e less transfer paym ents, and real
m anufacturing an d trad e sales. These four w ere rated
highest by the BEA using a scorin g system based on

seven criteria. On the basis of these criteria, no single
ind icator dom inates the others as a m onthly ind icator
of eco n om ic conditions.
F o r the casual e co n om ic analyst, the least satisfactoiy ind icator w ould be real m anufacturing and trade
sales becau se, of the four series, it is the m ost volatile
on a m onthly basis. Also, it is slow to be released; the
estim ate for a p articu lar m on th is not available until a
m onth and a half later. The release lag for two of the
oth er series is about one w eek for nonagricultural
em ploym ent an d about 15 days for industrial p ro d u c
tion. Although the basic d ata for real personal incom e
less transfer paym ents are released in about 20 days,
the inflation-adjusted series is not released until the
end of the m onth for the preceding m onth. The co m
posite index is released at the end of the m onth, but
the first estim ate for the preced in g m onth is based on
only three of the co m p o n en ts; b ecau se of the rep o rt
ing lag, real m anufacturing an d trad e sales is exclu d ed
from this first estim ate.
An exam ination of th e cyclical perform an ce of the
coinciden t index and its co m p o n en ts revealed that
each series generally con form ed well with the NBER
business cycle reference dates. This result is not su r
prising, since turning points in eco n o m ic activity are
am ong the m ost im p ortan t criteria used in the se le c
tion of coinciden t series.

NOVEMBER 1985

FEDERAL RESERVE BANK OF ST. LOUIS

One key question relating to the cyclical reliability of
an indicator is the extent to w hich it em its false sig
nals. An exam ination of m onthly m ovem ents of the
indicators show s that they do, in fact, give false sig
nals, betw een 15 p ercen t and 30 p ercen t of the time.
All of the indicators perform very well, however, w hen
false signals are defined as three successive m on th s of
perverse movement.
Finally, som e insights into the com parative perfor
m an ce of the coinciden t indicators w ere gleaned from
simple correlation analysis. This analysis indicates
that nonagricultural em ploym ent and industrial p ro
d u ction are m ost closely related on a m onthly basis
with the com posite index. These two co m p o n en t se
ries also receive the greatest weight in the co n stru c
tion of the index. On a quarterly basis, however, real
personal incom e less transfer paym ents correlates

m ost highly with real GNP. Real m anufacturing and
trade sales generally perform s least satisfactorily.

R E FE R E N C E S
Moore, Geoffrey H., and Julius Shiskin. Indicators of Business Ex
pansions and Contractions, Occasional Paper 103 (National Bu
reau of Economic Research, 1967).
Ratti, Ronald A. “A Descriptive Analysis of Economic Indicators,”
this Review (January 1985), pp. 14-24.
U.S. Department of Commerce, Bureau of Economic Analysis. Busi
ness Conditions Digest (U.S. Government Printing Office, August
1985).
________ . Handbook of Cyclical Indicators, A Supplement to the
Business Conditions Digest (GPO, 1977).
________ . Handbook of Cyclical Indicators, A Supplement to the
Business Conditions Digest (GPO, 1984).
Zarnowitz, Victor, and Geoffrey H. Moore. “ The Timing and Severity
of the 1980 Recession,” in Geoffrey H. Moore, Business Cycles,
Inflation, and Forecasting, 2nd ed. (Ballinger Publishing Company,
1983), pp. 11-17.

APPENDIX
Detailed Discussion of the Component Series of the
Coincident Index
Employees on Nonagricultural
Payrolls

Personal Incom e Less Transfer
Payments in Constant (1972) Dollars

This series m easures the n u m ber of persons em
ployed in nonagricultural establishm ents. Data are
obtained from the establishm ent survey co n d u cted
each m onth by the Bureau of Labor Statistics. The
data are primarily from payroll record s voluntarily
rep orted each m on th to state em ploym ent security
agencies by em ployers in the 50 states and the
District of Columbia. Most of these d ata relate to the
payroll period that includes the 12th of the m onth;
data for federal governm ent em ployees represent
positions o ccu p ied on the last day of the m onth.
Included are full-time, part-tim e, tem p oraiy and
perm an en t w orkers. W orkers on paid leave and
those w ho w orked part of the pay period are in
cluded. Persons on the payroll of m ore than one
establishm ent are co u n ted each tim e they are re
ported. E xclu d ed are p ersons on nonpav status for
the entire period due to layoff, strike or leave w ith
out pay; the self-em ployed an d unpaid volunteer
and family w orkers; farm and dom estic w orkers;
and noncivilian governm ent em ployees.

This series m easures personal incom e less tran s
fer paym ents in co n stan t (19721 dollars. Because
transfer paym ents rep resen t the largest part of p er
sonal incom e not a ccru e d in p rodu ction and som e
types of transfer paym ents tend to be co u n te r
cyclical, th e ir rem oval from p erso n al in com e
p rod u ces a series w ith greater cyclical am plitude.
The cu rrent dollar series is deflated with the im
plicit price deflator for pei'sonal consum ption
expenditures.

Personal incom e is the incom e received by p er
sons from all sou rces, that is, from participation in
production, from transfer paym ents from govern
m ent and business and from governm ent interest,
w hich is treated like a transfer paym ent. Persons
consist of individuals, nonprofit institutions, pri
vate noninsured welfare funds an d private trust
funds.
Alternatively, personal incom e is defined as the
sum of w age and salary disbursem ents, o th er labor

FEDERAL RESERVE BANK OF ST. LOUIS

incom e, p ro p rietors’ incom e w ith inventoiy valua
tion and capital con su m p tion adjustm ents, rental
incom e, personal interest incom e and transfer pay
m ents less personal contributions to social insur
ance.
Transfer paym ents to p erson s are incom e pay
m ents, generally in m on etary form, for w hich the
recipients do not ren d er cu rren t services. They
consist of business an d governm ent transfer pay
m ents. Business tran sfer paym ents include liability
paym ents for personal injury, co rp o rate gifts to
nonprofit institutions and bad debts incu rred by
con su m ers. Government transfer paym ents include
paym ents u n d er the following program s: federal
old age, survivors, disability an d hospital insu ran ce;
supplem ental m edical in su ran ce; state unem ploy
m ent insu ran ce; railroad retirem ent and u n em
ploym ent in su ran ce; governm ent retirem ent; work
e rs’ co m p en sation ; veterans benefits, including
veterans life insu ran ce; food stam ps; black lung
benefits; supplem ental security in com e; an d direct
relief. Also included are governm ent paym ents to
nonprofit institutions o th er than those for work
u n d er research an d developm ent co n tracts.
The implicit p rice deflator for personal co n
sum ption expen d itures (PCE) is a cu rren t w eighted
index (1972 = 100) derived by dividing cu rren t dollar
PCE by co n stan t dollar PCE for each period. It is a
w eighted average of the detailed p rice indexes used
in the deflation of PCE w ith com p osition of the
con stan t dollar PCE in each q u arter as w eights.

Index o f Industrial Production
This series m easures m onthly ch an ges in the
physical output of the m anufacturing, mining, and
gas an d electric utility industries. F o r m an u factu r
ing and mining, p ro d u cts at all stages of fabrication
are included. The index does not cover p rodu ction
on farms o r in the co n stru ction , transportation,

Digitized for 30
FRASER

NOVEMBER 1985

trade and service industries. It does, however, in
clude p rodu ction at plants an d shipyards ow ned
and operated by the governm ent. The ind ex is co n
stru cted using d ata supplied by governm ent agen
cies and trade organizations, an d u ses 1977 as the
base year. It is b ased on 252 series that are co m
bined with value-added w eights to create the total
index of industrial produ ction .

Manufacturing and Trade Sales in
Constant (1972) Dollars
This series m easures th e m onthly volum e of sales
of m anufacturing, m erch an t w holesalers an d retail
establishm ents in co n stan t 1972 dollars. The series
is com piled from d ata collected ea ch m on th by the
Bureau of the C ensus in the shipm ents, inventories
and orders survey an d in the m erch an t w holesalers
and retail trade surveys. They are adju sted to
benchm arks from the five-year cen su ses of m an u
factures, w holesale trade, and retail trad e an d to
interim annual surveys.
Basic d ata on m an u factu rers’ sales are the value
of their shipm ents for d om estic u se o r exp ort. Ship
m ents are m easured by receip ts, billings or the
value of p ro d u cts shipped (less discounts, retu rn s
and allowances) an d generally exclu d e freight
charges an d excise taxes.
Deflated sales are co m p u ted as follows: m an u fac
tu rers’ sales are deflated by industry levels (stan
dard industrial classification as defined by the Bu
reau of the Census) prim arily using p ro d u ce r p rice
indexes com bined w ith 1977 p ro d u ct-class sh ip
m en ts w eights; w holesale sales are deflated by kind
of business using ap p rop riate p ro d u ce r p rice in
d exes along w ith 1 9 7 7 co m m o d ity -lin e sales
weights; and retail sales are deflated by kind of
business using co m p o n en ts of the co n su m e r price
index w ith 1972 ce n su s com m odity-line sales
weights.