View original document

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

The Review is published 10 times per year by the Research and Public Information Department o f
the Federal Reserve Bank o f St. Louis. Single-copy subscriptions are available to the public fr e e o f
charge. Mail requests f o r subscriptions, back issues, or address changes to: Research and Public
Information Department, Federal Reserve Bank o f St. Louis, P.O. Box 442, St. Louis, Missouri
63166.
Articles herein may be reprinted provided the source is credited. Please provide the Bank’s
Research and Public Information Department with a copy o f reprinted material.




Federal Reserve Bank of St. Louis
Review
June/July 1983

In This Issue . . .




The four articles in this Review address two important public policy issues. The
first article focuses on the potential usefulness to farmers of the forthcoming
agricultural options markets. The next three articles are concerned with various
aspects of monetary policy actions.
In the first article, “Commodity Options: A New Risk Management Tool for
Agricultural Markets, ” Michael T. Belongia discusses the mechanics of trading in
the agricultural options markets, scheduled to begin next year as a supplemental
phase of a Commodity Futures Trading Commission pilot program. He finds that
options trading will provide new hedging and speculative opportunities for
farmers who produce grain and businesses that purchase grain as an input.
Belongia demonstrates how options could be used as part of a farmer’s overall
marketing strategy and how their use could affect income under various assump­
tions about market supplies and prices. The article concludes with a discussion of
whether options markets would provide price insurance and liquidity to farmers
more efficiently than current price support programs.
In the second article in this Review, “Two Measures of Reserves: Why Are They
Different?” R. Alton Gilbert describes the two measures of banking system
reserves published by Federal Reserve sources, and analyzes why their growth
rates often differ substantially over periods of a few months.
He shows that differences between growth rates of the reserves series published
by the Federal Reserve Bank of St. Louis and by the Board of Governors often
reflect simply the different methods used in adjusting the reserve series for
seasonal influences. At times, however, differences in methods of adjusting for the
effects of changes in reserve requirements and differences in the treatment of vault
cash as reserves have contributed to the short-run differences in growth rates
observed in the past.
The Board of Governors revised its measure of reserves in May 1983, when it
adopted new procedures for adjusting for seasonal influences. Although this
revision did not reduce the average difference between monthly growth rates of
the two measures of reserves, it will change the seasonal pattern that had previous­
ly existed between growth rates of the two series.
In the third article of this issue, Daniel L. Thornton reviews the policy actions of
the Federal Open Market Committee (FOMC) during 1982. Because of the
uncertainties about the relative behavior of M l and M2 during the year due to
technical factors, financial innovations and deregulation, and because of the
significant decline in the velocity of M l, the FOMC shifted the relative weights
given to M l and M2 for short-run policy purposes. Eventually, it suspended the
use of M l as an intermediate policy target. Thornton shows that, despite these
uncertainties, both M l and M2 were close to the FOMC’s desired short-run target
paths.

3

In This Issue

4



In the fourth article in this issue, “M l or M2: Which Is the Better Monetary
Target?” Dallas S. Batten and Daniel L. Thornton assess the extent to which
financial innovation and deregulation of the past few years have affected the
relative importance of M1 and M2 as intermediate targets of monetary policy.
They investigate the relationship between each monetary aggregate and economic
activity over the period that includes the latest two innovations — the introduction
of money market deposit accounts (MMDAs) and super-NOW accounts. They find
that, while the relationship between M l and nominal GNP is stronger, in general,
than that of M2 and nominal GNP, recent events have had greater confounding
effects on the Ml-GNP relationship. While this result should motivate continued
scrutiny of the relative merits of M1 and M2, it provides no rationale, at present, to
conclude that M l be de-emphasized as an intermediate target of monetary policy.

Commodity Options: A New Risk
Management Tool for Agricultural
Markets
MICHAEL T. BELONGIA

TA HE trading of options on agricultural commod­
ities has been banned in the United States since 1936.
In a preliminary step to lift this ban, Congress included
a provision in the Futures Trading Act of 1982 that
authorized the Commodity Futures Trading Commis­
sion (CFTC) to establish pilot programs in the trading
of agricultural options. Although actual trading of op­
tions on domestically produced agricultural commod­
ities has not yet taken place, the CFTC expects its pilot
program to include one option contract at each of the
major exchanges.1 The pilot programs for agricultural
commodities are expected to begin sometime in late
1984 and continue for three years, at which time they
will be evaluated.
For many people, the role of options in an overall
risk-management strategy is unclear. In fact, because
options trading has been banned for many years, the
distinguishing characteristics of options are not widely
known outside the commodities profession. This arti­
cle attempts to clarify some of these issues by explain­
ing the basic features of options and drawing distinc­
tions between options and futures. The discussion also
includes some simple examples of how options can
function as a risk-management tool. Finally, because
options contracts contain some — but not all — of the
features of agricultural price support programs, the
relationship between options markets and price sup­
ports is discussed.

Options on sugar are traded currently at the New York Coffee,
Sugar and Cocoa Exchange as part of the pilot program’s first stage;
the options apply, however, only to sugar produced outside of the
United States. Options on gold, Treasury bond and stock index
futures also are being traded as experimental contracts in the pilot
program.




FORWARD CONTRACTS AND
FUTURES CONTRACTS
To define the unique characteristics of a commodity
option, it might be useful first to discuss two related
concepts: forward contracts and futures contracts.2 A
forward contract typically takes the form of an agree­
ment between a commodity producer and an in­
termediary agent like the operator of a grain elevator.
The contract typically defines an agreement in which a
producer agrees to deliver to an elevator owner a
specified quantity of grain at a stated date for a set
price; the elevator owner agrees to accept delivery of
the grain and to pay the set price.
A futures contract is a binding legal agreement be­
tween parties to sell or purchase a specified quantity of
a standardized commodity at a stated date in the future
for a set price. For example, corn contracts at the
Chicago Board ofTrade are written in 5,000 bushel lots
of No. 2 yellow corn and carry stated delivery dates of

2General references on the role of hedging include Holbrook Work­
ing, “ Hedging R econsidered,” Jo u rn al o f Farm Econom ics
(November 1953), pp. 544- 61; Ronald I. McKinnon, “Futures
Markets, Buffer Stocks and Income Stability for Primary Produc­
ers,” Journal o f Political Economy (December 1967), pp. 844-61;
and Anne E . Peck, “Hedging and Income Stability: Concepts,
Implications, and an Example,” American Jou rn al o f Agricultural
Economics (August 1975), pp. 410-19.
A general overview of trading in commodity options can be found
in Avner Wolf, “Fundamentals of Commodity Options on Fu­
tures,” Jou rn al o f Futures Markets (Winter 1982), pp. ,391—408;
Bruce L. Gardner, “Commodity Options for Agriculture,” Amer­
ican Journal o f Agricultural Economics (December 1977), pp.
986-92; and William J. Baumol, “Commodity Options: On Their
Contribution to The Economy,” mimeographed (Princeton, N .J.:
Mathematica, Inc., September 1973).

5

FEDERAL RESERVE BANK OF ST. LOUIS

up to 16 months forward.3 The set price at which the
corn can be bought or sold is determined daily in the
market where this particular futures contract is traded.
These definitions indicate at least two respects in
which forward and future contracts differ. First, fu­
tures are standardized contracts traded in highly liquid
and well-organized markets. In contrast, forward con­
tracts are individual agreements between two parties;
their unique, case-by-case nature effectively prevents
their trading and, consequently, makes them very il­
liquid. The two contracts also differ in their handling of
prices at which exchange will occur. Specifically, the
price at which grain will change hands in the forward
contract is fixed for the duration of the contract. The
price of a futures contract, on the other hand, changes
daily as new supply and demand information affects
agents’ expectations of market prices at the time the
futures contract expires. Because forward contracts are
not traded in organized markets, they are excluded
from the remainder of the discussion.4
The price of a corn futures contract depends on
expectations of future spot corn prices, and, because
these expectations change from day to day, so, too, do
contract prices. If a trader believes that com prices will
be above the overall price expected by the market (the
average contract price) in the future, he will buy a com
contract for future delivery of 5,000 bushels; this is a
“long” position, which will generate an economic profit
if the price of corn rises above the contract price.
Conversely, an agent who wants to insure against a
decline in the expected future price of corn will sell a
futures contract agreeing to deliver com at some future
date; this is a “short” position. If the agent is an agri­
cultural producer who holds an inventory of com, this
strategy will provide a hedge against price declines.
We can see, then, why futures markets might exist.5
Producers (hedgers) who wish to avoid risk sell a fu­
tures contract; although they forfeit the chance to in­
3Com contracts are dated for March, May, July, September and
December delivery. Contracts typically expire during the second
to last week of the stated delivery month.
‘‘This is not to say that forward-contracting is unimportant. Instead,
unlike futures and options, forward contracts are individual legal
agreements not traded in organized markets.
T ’lie question of why futures markets exist does not have a definitive
answer. Some economists have argued that these markets perform
an insurance function while others find value in the amount of
information on prices and price expectations that futures markets
produce; see, for example, Fischer Black, “The Pricing of Com­
modity Contracts,” Jou rn al o f Financial Econom ics (January/
March 1976), pp. 167-79.
Some economists, however, have questioned the validity of the
insurance argument; see, for example, Lester G. Telser “Why
There Are Organized Futures Markets, ’Jou rn al o f Law and Eco-


6


JUNE/JULY 1983

crease profits if prices increase, they are guaranteed a
known return. Other agents (speculators) bear this
price risk in return for the chance to profit if prices rise
above expectations. As we will see, options function in
a similar manner, except for one distinguishing feature
of futures contracts: the only way to escape the obliga­
tion of the futures contract is to sell it to another party.

WHAT IS AN OPTION?
In contrast, an option conveys the right, but not the
obligation, to buy or sell a given amount of a commod­
ity at a fixed price until some specified date when the
option expires. Unlike a futures contract, which re­
quires the purchase or sale of a commodity, the holder
of an option may elect to let the option expire without
exercising its rights. In this sense, an option is more
like a form of price insurance in which one person pays
a premium to insure against the possibility of a particu­
lar event occurring. If that event — specifically, a large
change in price — does not occur, the person who
purchased the option loses only the premium he paid
for the price insurance. In comparison, losses on short
futures positions, essentially, are unlimited; losses on
long futures positions are limited to the price of the
contract.
The two basic types of options are the “put” and the
“call. ” A call option gives the purchaser of that option
the right to purchase a given quantity of a commodity
at a stated price on or before the option’s expiration
date. Conversely, a put option gives the option pur­
chaser the right to sell a given quantity of a commodity
at a stated price on or before the expiration date; again,
with respect to the CFTC pilot program, options will
convey the right to buy or sell a particular futures
contract.6
nomics (April 1981), pp. 1-23. In particular, Telser contends that a
foward contract can provide all of the price insurance offered by a
futures contract. Instead, he argues, futures markets exist to meet
the demand for a “fungible financial instrument traded in a liquid
market” (p. 8). Or, rather, even though forward contracts and
futures both provide price insurance, the illiquidity of forward
contracts creates a demand for a more liquid instrument that holds
the attributes of money (or near money). Futures contracts exist,
Telser argues, to meet this demand for liquidity, not the demand
for price insurance.
Finally, some observers have argued that trading in futures is
little more than gambling.
frI’he management and surveillance of the pilot program have been
simplified by permitting options to apply only to trades of futures
contracts. That is, unlike an option to purchase a physical product
— a trade that would require agreements on the quality of the
product, place of delivery and other contract features — the pilot
program will permit only the trading of options on the standardized
futures contracts of specific commodities.

JUNE/JULY 1983

FEDERAL RESERVE BANK OF ST. LOUIS

Table 1

Sample Listing of Options on Sugar Futures__________
COFFEE, SUGAR & COCOA EXCHANGE Sugar Option Prices 5/31/83 C/lb.

July 83
13.36

Mar 84
14.48

Strike
Prices

Settlement
Puts
Calls

6.50
7.00
7.50
8.00
8.50
9.00
9.50
10.00
10.50
11.00
11.50
12.00
12.50
13.00

6.90
6.40
5.90
5.40
4.90
4.40
3.90
3.32
2.85
2.40
2.00
1.60
1.20
0.85

0.01
0.01
0.01
0.01
0.01
0.01
0.04
0.08
0.30
0.49
0.54
0.65
0.75
0.80

Oct 83
13.50

7.00
8.00
9.00
10.00
11.00
12.00
13.00
14.00
15.00

7.50
6.50
5.50
4.50
3.50
2.50
1.95
1.65
1.35

0.01
0.10
0.15
0.21
0.40
0.50
0.65
1.00
1.35

July 84
14.95

Strike
Prices

Settlement
Calls
Puts

7.00
7.50
8.00
8.50
9.00
9.50
10.00
10.50
11.00
11.50
12.00
12.50
13.00
13.50
14.00
7.00
8.00
9.00
10.00
11.00
12.00
13.00
14.00
15.00
16.00

6.50
6.00
5.50
5.00
4.50
4.00
3.50
3.00
2.65
2.20
1.90
1.75
1.60
1.45
1.25
6.95
5.95
4.95
3.95
2.95
2.15
1.80
1.45
1.20

0.01
0.01
0.02
0.20
0.30
0.35
0.45
0.55
0.75
1.10
1.60
0.01
0.03
0.10
0.27
0.45
0.68
0.90
1.10
1.50
2.00

Vol. 5/27/83 37 Open Int. 5/27/83 calls 1,711 puts 171
Each .01 premium = $11.20 e.g., .50 = $560.

Farmers who wish to use options as a hedge against
declining cash prices would buy a put option in com­
bination with positions in the forward, futures or cash
markets. Food processors or other businessmen that
sought a hedge against price increases in the raw com­
modities they purchase would buy a call option to
complement their positions in other markets. An agent
who has a position only in the option market is a
speculator. Speculators fulfill a desirable market func­
tion by assuming risk that other economic agents do
not wish to bear.
Each option contract has several characteristics
specified as part of the legal document itself. These
include the futures contract to be traded, the price at
which the option purchaser may buy or sell the futures
contract (called the strike or exercise price) and the
expiration date for the option. Another important fea­
ture of options — the option premium — is deter­
mined by supply and demand conditions in the option
market.



The relationships among these different option
terms are shown in table 1, which is a reprint of one
daily summary of the sugar options traded on The New
York Coffee, Sugar and Cocoa Exchange. The sum­
mary, dated May 31, 1983, applies to options on sugar
futures that expire in July 1983, October 1983, March
1984 and July 1984. Considering only the option on
July 1983 futures, the summary indicates that strike
prices cover a range from 6.5 to 13 cents per pound of
sugar; that is, a variety of option contracts are available
and each option permits the buyer to sell or purchase
sugar futures at a stated price somewhere between 6.5
and 13 cents per pound. The number immediately
below the contract date (July 83) is the current price of
July sugar futures — 13.36 cents per pound.
The second and third columns, both under the head­
ing “Settlement,” are the premiums that apply to the
different put and call options contracts. Recall that,
while the strike prices of different options are a part of
those legal contracts, the settlement premiums on call
7

FEDERAL RESERVE BANK OF ST. LOUIS

and put options are determined by price expectations
in the option market. Generally, premiums are related
to three factors: the strike price of the individual op­
tion, the length of time until the option expires and the
price variability of the underlying futures contract.
To take a specific example from these data, the pre­
mium on a call option to purchase July 1983 sugar
futures at 10 cents per pound is 3.32 cents per pound;
the total cost of guaranteeing the possibility to pur­
chase July sugar futures at 10 cents per pound is 13.32
cents per pound (10 + 3.32), compared with 13.36
cents per pound futures price. Absent from these cal­
culations are the transaction costs (brokerage fees) of
buying an option or a futures contract.

DETERMINANTS OF OPTION
PREMIUMS
Option premiums are related directly to an option’s
intrinsic value and its time value.7 Intrinsic value is the
difference between an option’s strike price and the
current futures price. For example, if a call option’s
strike price — the amount at which a corn futures
contract could be purchased — were $2.50 per bushel
and the current futures price were $2.70 per bushel,
this option would have an intrinsic value of $0.20 per
bushel. Intuitively, intrinsic value exists if a profit can
be made by exercising the rights of the option. If, in the
example above, the current futures price were $2.30, a
call option with a $2.50 per bushel strike price would
be “out of the money”: that is, a loss would be incurred
i f th e o p tio n rig h ts w e r e e x e rc ise d . T y p ic a lly , h o w e v ­

7Option premiums also are influenced by the volatility of futures
prices and interest rates. As futures prices become more volatile,
the uncertainty associated with any one contract’s profitability also
increases. This greater uncertainty tends to increase the value of
price insurance and, therefore, the value of option premiums.
Conversely, high levels of interest rates tend to have negative
effects on option values. That is, as the returns on alternative,
interest-bearing investments increase, the opportunity cost of
holding an option position increases. This “competition” among
alternative investments will tend to decrease option premiums.
For a technical discussion of how option premiums are deter­
mined, see Fischer Black and Myron Scholes, “The Valuation of
Option Contracts and a Test of Market Efficiency,” Jou rn al o f
Finance (May 1972), pp. 399-418; Black and Scholes, “The Pricing
of Options and Corporate Liabilities, "Journal o f Political Economy
(May-June 1973), pp. 637-54; Robert C. Merton, “The Theory of
Rational Option Pricing,” Bell Jou rn al o f Economics and Manage­
ment Science (Spring 1973), pp. 141-83; Clifford W. Smith, “Op­
tion Pricing: A Review "Jou rn al o f Financial Economics (January/
March 1976), pp. 3-5 1 ; James MacBeth and Larry L. Melville, “An
Empirical Examination of the Black-Scholes Call Option Pricing
Model,” Journal o f Finance (December 1979), pp. 1173-86; and
Thomas J. O’Brien and William F . Kennedy, “Simultaneous Op­
tion and Stock Prices: Another Look at The Black-Scholes Model,”
The Financial Review (November 1982), pp. 219-27.

8



JUNE/JULY 1983

er, an option’s premium will exceed the implied
amount of its intrinsic value.
One reason premiums will exceed intrinsic value is a
second source of value in an option contract: time
value. Because the future is uncertain, there is always
the possibility that unexpected events will significantly
affect prices. And, because this possibility exists, some
market participants will be willing to buy or sell an
option on the chance that one such event will occur.
This explains, for example, why an “out of the money”
option still will be traded at a positive premium; that is,
some buyers are willing to take the chance that some
event will change futures prices enough to make this a
profitable option. Similarly, premiums may be greater
than intrinsic value because buyers are willing to pay
for the chance that further changes in the futures price
may make a profitable option even more profitable
before it expires.
An option’s expiration date is the key factor in deter­
mining its time value. As the length of time until
expiration decreases, there is less time for the futures
price — and, therefore, the option’s profitability — to
change markedly. Conversely, an option of long dura­
tion has more time value, ceteris paribus, because the
probability of an unexpected event changing its prof­
itability is greater.
The concepts of intrinsic value and time value are
illustrated by the data shown in table 1. For example,
the call option on March 1984 futures with a 14 cents
per pound strike price has an intrinsic value of 0.48
c e n ts p e r p o u n d ( 1 4 .4 8 — 1 4 .0 0 = 0 .4 8 ) , w h ich is
lower than its premium of 1 .6 5 cents per pound. The

1.17 cent difference between the premium and intrin­
sic value reflects this option’s time value and other
factors that tend to increase premiums. Other things
being equal, this 1.17 cent difference should decline as
the length of time until March decreases and the time
value of the option diminishes.
Time value also is shown in the premiums associated
with options that apply to futures contracts dated for
later delivery. Compare, for example, the four call
options with 11 cents per pound strike prices that
apply to each of the four listed futures. In this instance
— and in others — premiums for options on July 1984
sugar futures are the highest premiums for any of the
listed contracts. This occurs because the greater length
of time until the option expires increases the probabil­
ity that some unanticipated event will cause significant
changes in futures prices. And, with greater price un­
certainty, agents in this market will be willing to pay
more for price insurance.

FEDERAL RESERVE BANK OF ST. LOUIS

F ig u re 1

R e la tio n sh ip Betw een P re m iu m s o n Put a n d C a ll O p t io n s

JUNE/JULY 1983

premium on a put should equal the premium on a call.
The reasoning is that, if the futures price represents
the market’s best guess about actual prices at a later
date, the value of the right to buy at that price (a call)
should be equal to the value of the right to sell at that
price (a put). Or, from a different view, options with
strike prices above or below the current futures price
carry an implicit bet that the current futures price is
“wrong.” So, for example, a call option with a strike
price at point A would have a relatively low premium,
because it would give the right to buy a product in the
future at a price higher than the market’s current best
guess of that future price. Conversely, a put option
with that same strike price would have a relatively high
premium to reflect the bet that the current futures
price underestimates the level of cash prices at the
later date.

MARKETING STRATEGIES AND THE
ROLE OF OPTIONS
Pp - p re m iu m o n a p u t o p tio n w ith s trik e p r ic e - A.
Pc = p re m iu m on a c a ll o p tio n w ith s trik e p r ic e = A .

These relationships are illustrated further in figure
1. Strike prices are plotted on the horizontal axis and
option premiums are plotted on the vertical axis. The
two interior lines labelled “Puts” and “Calls” plot the
relationships between strike prices and premiums for
the two different kinds of options. In one sense, this
figure is a stylized plot of the strike price and premium
data contained in table 1.
As the foregoing discussion suggests, strike prices
and premiums for put options should be related posi­
tively. That is, the right to sell a product at a low price
should have a relatively low value. Conversely, as the
strike price at which the product can be sold rises, the
right to execute this sale also should increase in value.
These relationships are the basis for giving the “Put”
line a positive slope.
The strike price-premium tradeoff for call options is
just opposite that of puts and, consequently, its line has
a negative slope. Intuitively, this is supported by the
notion that the right to buy a product at a low price
should have a higher value than the right to buy at a
high price. Therefore, as a call option’s strike price
declines, its premium should increase.
The intersection of the “Put” and “Call” lines also
suggests a relationship not revealed in the discon­
tinuous data of table 1. Specifically, when an option’s
strike price is equal to the current futures price, the



The mechanics and terminology of options trading
may be defined further by way of an example. Consider
the case of a farmer who, at time of spring planting,
expects to produce 5,000 bushels of com, an amount
that coincides with the size of one futures contract. He
also thinks that his total cost of producing each bushel
of corn will be $2.50. Finally, he knows that the futures
contract dated for December delivery — after his har­
vest time — values corn at $2.80 per bushel. Assuming
a constant 10-cents-per-bushel basis, he can expect
local cash prices at the time of harvest to be $2.70 per
bushel.8
These prices and the effective support prices of gov­
ernment crop programs represent the core of informa­
tion on which his marketing decisions must be made.
Still unknown, however, are the quality of the growing
season weather and the effects it and other factors may
have on his yield per acre. Or, rather, because he still
is unsure of his yields and those of other producers, it is
unclear whether cash prices at harvest will be higher or
lower than $2.70.
8Basis is the per unit difference in the futures price and the local spot
(cash) price for a commodity. In this example, the current (May)
price of a December futures is $2.80; the current spot price is
$2.70. Therefore, the basis is $0.10. The returns to a person with a
position in the futures market is the change in the basis that occurs.
That is, if December futures increase to $3.00, the basis becomes
$0.30 ($3.00 - $2.70) which produces a $0.20 ($0.30 - $0.10)
change in the basis. This 20-cent change will be a gain or loss
depending upon whether a person held a long or short position in
futures. Typically, the basis reflects the spread between local spot
markets and the relevant futures market. Costs of financing, stor­
age and insurance also are part of the basis. Also, contrary to the
simplifying assumption of this example, the basis will not be con­
stant during the crop year.

9

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

Table 2

Results of Alternative Marketing Strategies
Under a Price Increase
Assum ptions:

On May 1, prior to planting, a com producer anticipates a 5,000 bushel harvest
with costs of $2.50 per bushel. Also in May, the December futures contract
prices corn at $2.80 per bushel. In November, the realized cash price is $3.10
per bushel and the December futures price is $3.20.
STRATEGY

IN MAY:

#1: No hedge
Plant com and do
nothing

IN NOVEMBER:

Sell com at $3.10
cash price

Income
Cost
Loss on Futures
Option Premium

$

PROFIT

$

15,500
-1 2 ,5 0 0

#2: Sell
Futures Contract

#3: Buy
“ Put" Option

Plant com and
sell one December
com future at
$2.80 per bushel

Plant corn and
buy one “ put”
option (right
to sell) on
December corn
futures with an
exercise price
of $2.80 and a
150 per bushel
premium

Sell com and buy
December futures
at $3.20

Sell corn
and let
option expire

$

15,500
-1 2 ,5 0 0
- 2,0001

$

15,500
-1 2 ,5 0 0
-

3,000

$

1,000

$

750
2,250

’ The increase in the price of December futures from $2.80 to $3.20 implies a $0.40 per bushel loss to the
person who sold December futures in May.

Depending on his own attitude toward risk, an indi­
vidual producer may choose several marketing
strategies. On one extreme, he may go totally un­
hedged — that is, he may just harvest his crop and
accept whatever cash price prevails at that time. At the
other extreme, a very risk-averse producer may hedge
his entire crop by selling one com futures contract. By
hedging, the producer can guarantee that the price he
receives for his com will be $2.80 per bushel, the
current price of December corn futures. Between
these extremes is a strategy in which a portion of the
crop is hedged in the futures market and the remainder
is sold at the prevailing cash price.
These strategies, however, also indicate that there is
a gap in alternatives that would be filled by a market in
commodity options. That is, a producer who is totally
unhedged has no insurance against downside price
movements. Or, following this example, a producer
who does not hedge at least part of his crop in the
Digitized for 10
FRASER


futures market might face a cash price of something
like $2.00 at time of harvest if the national crop is larger
than previously expected; this would produce a loss of
$0.50 per bushel. Conversely, a producer who hedged
all 5,000 bushels at $2.80 has no alternative but to
accept that price at harvest. While this form of price
insurance guarantees $2.80 per bushel, it also pre­
cludes the chance to sell for the higher cash prices that
could prevail if the national crop were smaller than
expected. Instead of these marketing positions, a more
flexible approach would have two characteristics: it
would provide insurance against a decline in cash
prices, while simultaneously allowing gains to be made
if cash prices increased above contract prices. Com­
modity options have these features.9
9Options have the advantage that if a farmer’s hedge became un­
covered due to, say, a crop failure, his losses would be limited to
the option premium. Losses from a futures hedge under these
conditions could be much larger if prices increased substantially.

FEDERAL RESERVE BANK OF ST. LOUIS

Continuing with the earlier example, it is clear that a
market in commodity options would expand the scope
of marketing strategies for farmers and agribusiness. In
addition to the earlier marketing strategies — no hedg­
ing versus complete hedging — a third strategy involv­
ing an option is introduced. Under the assumptions in
this example, production costs of $12,500 will be incur­
red under any marketing strategy (5,000 bushels X
$2.50 per bushel). Also assumed is a $750 (15 centsper-bushel premium) cost for buying a “put” option.
Table 2 shows the results of these strategies under an
assumed increase in futures prices to $3.20 per bushel;
assuming a constant 10-cents basis throughout this
example, cash prices at harvest would be $3.10 per
bushel.10 Each strategy is detailed in a separate col­
umn of the table. Each strategy also involves two dis­
tinct steps: first, the choice of a marketing strategy at
time of planting (May) and, second, the execution of
that strategy after harvest (November). These stages
are represented in the upper and lower halves of the
table.11
Under assumed increases in futures prices to $3.20
and in cash prices to $3.10 per bushel, strategy No. 1
yields a return of $3,000, the highest of the three
strategies. Because income and production costs are
equal in each strategy, the “no hedge” earns greater
returns because it avoids a loss of futures (strategy No.
2) and the cost of option premiums (strategy No. 3).
Therefore, a producer choosing strategy No. 1 earned a
greater profit during this year but did so without insur­
ance against price declines. Conversely, producers
choosing to hedge their crops or purchase options real­
ized smaller profits, but were protected against the
possibility of price decreases. The return to strategy
No. 2 is lower by the $2,000 loss on the sale of Decem­
ber futures [($3.20 - $2.80) x 5,000 bu. = $2,000].
Producers choosing to buy options instead of futures
earned greater profits, but this result is dependent on
the assumed values for alternative options premiums
10Each of these examples ignores a number of factors that would
complicate the analysis. For example, the output of this individual
producer does not vary with changes in aggregate production. The
returns also are dependent on assumptions regarding the elastic­
ity of demand. Rather than providing a complete analysis that
considers these complicating considerations, however, the intent
of the examples is to illustrate qualitative differences among the
various strategies.
"T h e strategies shown are the most basic approaches to grain
marketing. Much more complicated examples, which combine
the simultaneous use of differing positions in futures and options
markets, can be used to illustrate how varying levels of price
insurance and speculation can be achieved. See, for instance, the
strategies discussed in Strategies f o r Buying and Writing Options
on Treasury Bond Futures published by the Chicago Board of
Trade. These examples can be adapted with few changes to
strategies for grain marketing.




JUNE/JULY 1983

and changes in futures prices; these results merely
illustrate qualitative differences among marketing
strategies.
Returns under different risk-management strategies
might be illustrated more clearly by re-evaluating the
previous example under a decline in the futures price.
Table 3, which includes balance sheet figures for an
assumed November futures price of $2.40 and cash
price of $2.30 (constant 10-cents-basis assumption),
reports these results. As in the previous example, the
strategy involving options (No. 3) yields a return be­
tween those of the other strategies. Now, however,
after a price decline, the unhedged strategy (No. 1)
yields a loss of $1,000; or, rather, column one shows
what can occur if market prices decline and a producer
has no protection against such losses. Conversely, col­
umn two — under a strategy of complete hedging —
shows the benefits of locking in a known price at the
time of planting. Finally, the strategy that includes
options shows a profit, but one less than that for selling
futures; the difference is the amount of the option
premium. But, although the premium costs are $750,
the purchase of the corn futures yields a return of
$ 2 , 000 .

Finally, table 4 illustrates the relative returns to the
three strategies if prices do not show a net change
during the year. As the entries in the table indicate,
each strategy would result in a sale of corn in the cash
market at $2.70 per bushel. Once again ignoring the
transaction costs of futures or options contracts,
strategies No. 1 and No. 2 would yield a profit of
$1,000, whereas the cost of the option premium would
reduce returns to strategy No. 3 to $250.
In view of these differing returns to different
strategies as assumptions vary concerning end-ofseason prices, an important consideration is the ex­
pected (ex ante) return to each marketing strategy.
That is, in May, what can an individual producer ex­
pect to earn from crop marketings in November?
A comparison of these expected values is shown in
table 5.
As the table indicates — for this set of alternative
outcomes and probabilities — strategies No. 1 and No.
2 yield equal expected returns, while the strategy us­
ing options produces a lower expected return. This is
not unlikely, however, in view of the speculative ser­
vices that options offer in addition to their basic price
insurance.12 Or, rather, because options offer a chance
lzFor example, see Telser, “Why There Are Organized Futures
Markets,” for a discussion and references concerning why agents
may choose to engage in speculative strategies in which expected
returns are negative.

11

Table 3

Results of Alternative Marketing Strategies
Under a Price Decrease
STRATEGY

IN MAY:

#1: No hedge

#2: Sell
Futures Contract

#3: Buy
“ Put” Option

Same strategy as in table 2

IN NOVEMBER:

Sell com in cash
market at $2.30
cash price

Income
Cost
Futures Premium
Option Premium

$

PROFIT

$-

11,500
-1 2 ,5 0 0

1,000

Sell corn in cash
market at $2.30
and buy December
futures at $2.40

Buy December
futures for
$2.40 and
exercise option
right to sell
December futures
for $2.80;
sell harvested
corn for $2.30
cash price

$

11,500
-1 2 ,5 0 0
2,000

$

$

1,000

$

11.5001
-1 2 ,5 0 0
2,000
750
250

'Total income of $13,500 is derived from sales of harvested corn (5,000 bu. x $2.30 = $11,500) and
profit of $2,000 on the change in futures prices [($2.80 - $2.40) x 5,000 bu. = $2,000].

Table 4

Results of Alternative Marketing Strategies
Under Constant Prices




STRATEGY
IN MAY:

#1: No hedge

#2: Sell
Futures Contract

#3: Buy
“ Put” Option

Same strategy as in table 2

IN NOVEMBER:

Sell corn at $2.70
cash price

Sell corn in cash
market at $2.70

Let option
expire and sell
corn in cash
market at $2.70

Income
Cost
Futures Gain (Loss)
Option Premium

$

13,500
-1 2 ,5 0 0

$

13,500
-1 2 ,5 0 0

$

PROFIT

$

1,000

$

1,000

$

13,500
-1 2 ,5 0 0
-

750
250

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

Table 5

F ig u re 2

Expected Returns From Alternative
Marketing Strategies

Returns U n d e r A lte rna tive M a r k e t in g S tra te gie s
Returns
IS /b u .)

STRATEGY
#1: No hedge

#2: Selling Futures

#3: Buy
“ Put” Option

3

(3 ) R eturn s to a
P ut O p tio n

(2 )R e tu rn s to a
S tr a ig h t H e d g e

$

Total
Returns $

3,000

$

1,000

$

2,250

-1 ,0 0 0

1,000

250

1,000

1,000

250

3,000

$

3,000

$

Rh

2

2,750
f/

Expected returns = TOTAL RETURNS x (0.33)1

#1
E(R) = 1,000

#2

# 3

1,000

917

1For purposes of this example, it is assumed that the real-world
conditions described in tables 2 -4 all occur with equal probability
of 1/3.

''

?

(T)Re tu rn s to
C ash S ales

2

P rice
($ / b u .)

$ - 0.50

O p tion prem iu m
P* = E xe rcise (strike ) p ric e fo r a p u t o p tio n

for additional profits if prices rise to a level greater than
the sum of the market price plus the option premium,
it is expected that this additional speculative feature
can be gained only at some additional cost.
These relationships might be seen more clearly in a
graphic comparison of returns produced by the three
marketing strategies discussed earlier. Figure 2 plots
returns for unhedged (1), straight-hedge (2) and op­
tions (3) strategies.13 The dashed line shows the re­
turns to an unhedged strategy in which all grain is sold
in the cash market at the prevailing price; as might be
expected, it is a 45-degree line from the origin. Re­
turns to a straight hedge, involving the sale of a futures
contract, are shown by the horizontal line drawn at a
level denoted by RH- This line shows that the producer
can guarantee a return of RH per bushel but cannot
gain from price increases above that level.
The kinked line shows the returns to a strategy
involving options and, by inference, the role options
play in hedging — speculative strategies. In fact, the
shape of this returns line illustrates the unique features
of a put option. The horizontal segment of the line,
drawn at a level equal to $2, shows the maximum
return that can be achieved if futures prices are below
the option s exercise price. Or, rather, because the
13This figure is adapted from a similar diagram in Gardner, “Com­
modity Options for Agriculture.”




figure indicates that the option premium is $0.50, the
$2 level of the net returns line implies that the option’s
strike price is $2.50 (strike price — premium = net
return). This horizontal line segment also is the mini­
mum return the owner of this put option will earn. The
horizontal portion of the option’s return line, then,
represents the insurance characteristics of an option.
The returns line also has an upward-sloping segment
that begins at the break-even price of $2.50; this seg­
ment illustrates the speculative characteristics of op­
tions. That is, for futures prices above $2.50, the option
can be allowed to expire, and the grain can be sold in
the cash market at higher prices. Notice, however, that
this portion of the option’s return line falls below the
“returns to cash sales” line by an amount equal to the
option premium. Conversely, at all prices above $3.00,
the option yields a higher return than a straight hedge
in the futures market.
Finally, it should be noted that figure 2 implies that a
strategy involving options performs most poorly if
prices remain within a $2.00-$3.00 band; both futures
and cash market sales will produce a higher net return
for prices in this range. And, because the current
futures price in this example is $2.50, this result high­
lights again the unique feature of an option: it carries
the implicit “bet” that the futures price underesti13

FEDERAL RESERVE BANK OF ST. LOUIS

mates the eventual level of cash prices by a substantial
margin. In fact, as this example is written, the purchas­
er of a call option on this futures contract would believe
the futures price will increase by at least 20 percent (50
cents) to offset the option’s 50-cent premium.
The general result implied by these examples is that
commodity options provide insurance against price
declines without totally eliminating the potential
profits from price increases. Although total returns to a
strategy involving options tend to be lower than re­
turns to other strategies (for example, strategy No. 1 in
table 2 and strategy No. 2 in table 3) because the
additional costs of option premiums are incurred, op­
tions never produce a loss (in these examples) and yield
substantially higher returns than futures if prices in­
crease. Therefore, somewhat lower average returns
provided by a strategy involving options might be
viewed as the price paid for additional speculative
services not available in futures markets.

OPTIONS MARKETS VS. PRICE
SUPPORTS
A market in commodity options would offer grain
producers many of the hedging opportunities current­
ly available in legislated price support programs.14 For
example, a put option’s strike price would function in
much the same manner as program loan rates. And, as
with an option position, a producer is not required to
comply with program provisions but may elect to exer­
cise program privileges at his discretion. At a general
level, options and price support programs function in
similar fashion. And, in one sense, a function of the
CFTC pilot program may be to discover whether op­
tions markets can co-exist with price support programs
as they now stand.
There are at least two important differences, howev­
er, between options and current price support pro­
grams. First, unlike one specified loan rate that applies
for an entire crop year, an option purchaser may select
from a variety of contracts with different strike prices
and premiums. Second, trading in options contracts
will not have the large and direct effects on agricultural
production and resource allocation that have been
attributed to price supports. The particulars of each
distinction are discussed below.
If a grain producer is eligible to participate in a price
support program, one of his key decision variables is
wFor a discussion of these programs, see Michael T. Belongia,
“Outlook for Agriculture in 1983,” this Review (February 1983),
pp. 14-24; or Bruce L. Gardner, The Governing o f Agriculture,
(The Regents Press of Kansas, Lawrence, Kansas, 1981).

Digitized for14
FRASER


JUNE/JULY 1983

the program’s loan rate. If market prices fall below the
loan rate, which is a legislatively determined price per
bushel of grain, the producer can place his grain in
Commodity Credit Corporation (CCC) stocks and re­
ceive a loan in exchange; the loan value is determined
by multiplying the number of bushels placed in CCC
stocks by the loan rate. If, after nine months, market
prices have not risen above the loan rate, the producer
may elect to forfeit his grain to the CCC and keep the
loan. In this way, the loan rate serves as an effective
price floor for eligible producers. Also notice that,
although these producers are hedged against price
declines, they are free to sell their grain at market
prices if such prices rise above the loan rate.
This protection against large price declines, while
maintaining the possibility of profits, also is a distin­
guishing feature of a commodity option. Options,
however, differ from government price supports by
offering a range of strike prices (essentially, different
loan rates) from which a producer can choose. In other
words, options allow producers to select the level of
prices at which they wish to be hedged against further
price declines.
This point can be clarified by an example. Consider,
for instance, the 1983 com program and its loan rate of
$2.65 per bushel. While it provides this price floor for
producers, price insurance against declines below, say
$2.90, is available only by selling a futures contract at
that price. Recall, however, that one disadvantage of
this strategy is the rigidity of obligations implied by a
futures contract.
In contrast to these less flexible strategies, a viable
options market would allow producers to select the
level of price insurance they desire. For example, as a
parallel to the data in table 1, an option on com futures
might list strike prices ranging from $2.30 to $3.20 per
bushel; each option also would have its own premium.
Therefore, a producer who wanted protection against
price declines below $3.20 could buy a put option with
that strike price. Similarly, if $2.30 were an acceptable
price floor, that put option could be purchased. The
unique feature of options, however, is that individuals
are free to select the amount of price insurance they
desire and pay a competitively determined premium
for it.
The other main distinction between options and
price supports is that options are not likely to have
large direct effects on the quantity of grain produced.15
Economic theory suggests that effective support pro­
15See Gardner, “Commodity Options for Agriculture.”

FEDERAL RESERVE BANK OF ST. LOUIS

grams will increase production by increasing produc­
ers’ expected prices and decreasing the variance of
their returns.16 Under these conditions, producers can
expect to receive greater returns at less risk. If pro­
gram incentives to increase production are not offset
by output reductions effected by program acreage
limitations, price supports will allocate too many re­
sources to the production of the protected commod­
ities. These distortions in resource allocation could be
avoided if they were replaced by options trading.
But even if direct effects on output were minimized,
options will not avoid all resource allocation effects
associated with government price support programs.
For instance, this approach to risk management may
induce some producers to shift from the use of fertilizer
and pesticides to the purchase of options. Similarly,
agents who write options will likely shift some re­
sources from other investments to the purchase of
futures or physical commodities in an effort to offset
their options positions. Therefore, to the extent op­
tions becom e an attractive asset to marketing
strategies, this new market will have some effects on
resource allocations.
What options would avoid are the wealth transfers
and capitalization effects associated with the “free”
price insurance of government programs.17 That is,
current government programs transfer wealth from
taxpayers who pay for the price insurance to producers
16See, for example, Michael T. Belongia, “Agricultural Price Sup­
ports and Cost of Production: Comment,” American Jou rn al o f
Agricultural Economics (August 1983), forthcoming.
17Producers do pay — indirectly — if they are required to reduce
output to participate in the price support programs.




JUNE/JULY 1983

who receive the benefits of its protection. Wealth also
is transferred from land buyers to land owners via the
capitalization of program benefits into the value of land
eligible for those benefits. This capitalization also
raises land prices above the level they would have been
in the absence of government programs. This induced
change in land prices then affects the mix of resources
used to produce products in which land is an input.
Other secondary effects on resource allocation also
could be avoided if government programs were re­
placed by options markets.

CONCLUSIONS
The trading of options on agricultural commodities is
likely to begin sometime in 1984 under a pilot program
supervised by the CFTC. Options fill a gap between
futures markets and the price insurance of government
programs by offering market participants the oppor­
tunity to select the amount of price insurance they
desire while, simultaneously, not precluding the
opportunity for profits if prices change appreciably.
Although options will never provide the highest
level of income that could have been earned under an
assumption of perfect foresight, marketing strategies
that include options establish a minimum price for
producers without eliminating the opportunity for
gains if market prices increase. Finally, although op­
tions and price support programs are alike in many
respects, options would provide greater flexibility in
choosing a level of price insurance. Further, they are
less likely to increase agricultural production or pro­
duce the distortions in resource allocation associated
with price support programs.

15

Two Measures of Reserves:
Why Are They Different?
R. ALTON GILBERT

J S OTH the Board of Governors of the Federal Re­
serve System (ROG) and the Federal Reserve Rank of
St. Louis (St. Louis) publish series on the reserves of
depository institutions that have similar descriptions.
Each reserves series is adjusted both for the effects of
reserve requirem ent changes and for seasonal
influences.1
Though these series have similar descriptions, their
growth rates often differ substantially, especially over
periods as short as a month (table 1). For instance, in
the three years ending in December 1982, the differ­
ence between monthly growth rates of the two re­
serves series (absolute value) averaged 8.6 percentage
points. Average differences in growth rates were much
smaller over periods longer than a month (chart 1).
Absolute value of differences in quarterly growth rates,
for example, averaged 3.2 percentage points over the
years 1980-82, while the differences in growth rates
over four-quarter periods averaged 1.3 percentage
points.
The ROG revised its total reserves series in May of
this year. That revision primarily reflects new methods
of seasonal adjustment. The revision to total reserves
(ROG) essentially has no effect on the average differ­
ence between growth rates of the two reserves series.
For instance, the average difference in monthly
'Each of these institutions also publishes a measure of the monetary
base. This paper analyzes the reserves series from these two Fed­
eral Reserve sources, since financial analysts generally focus on the
reserves series in monitoring Federal Reserve actions that in­
fluence money growth. For earlier discussions of monetary base
and reserves measures, see R. Alton Gilbert, “Revision of the St.
Louis Federal Reserve’s Adjusted Monetary Base,” this Review
(December 1980), pp. 3 -1 0 ; John A. Tatom, “Issues in Measuring
an Adjusted Monetary Base,” this Review (December 1980), pp.
11-29; and Albert E . Burger and Robert H. Rasche, “Revision of
the Monetary Base,” this Review (July 1977), pp. 13-28.

Digitized for16
FRASER


growth rates over the 36 months ending December
1982 was 8.4 percentage points based on data available
just before the recent revision in total reserves (ROG),
compared with 8.6 percentage points with the revised
data.
Large differences in the growth rates of these re­
serves series make things difficult for those who
attempt to monitor the influence of Federal Reserve
actions on money growth. Large differences in these
growth rates also create public concern that data from
one, or possibly both, of the sources have been mea­
sured incorrectly; this is especially troublesome during
periods when there are major changes in reserve re­
quirements or the activities of depository institutions.
The purpose of this paper is to describe the two
methods of measuring the monetary base and reserves
and to analyze the effects of differences in the methods
of measurement on the growth rates of the reserves
series.2

CALCULATING THE RESERVES
SERIES
The St. Louis Series
Basic data: the source base — The basic series used
in calculating the St. Louis monetary base and reserves
series is the source base; it equals reserve balances of
depository institutions with Federal Reserve Ranks
plus total currency in circulation, whether held by
depository institutions or the public. The source base
^The text discusses the procedures for producing these measures of
reserves, while the appendix provides a more detailed description.
This paper does not select one series as a more appropriate or
useful measure of reserves.

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

Table 1

Growth Rates of the Reserves Series
(compounded annual rates of change, seasonally adjusted)
Board of Governors

Period
1980

1981

1982

1983

1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4

St. Louis
-1 7 .5 %
18.0
14.6
- 2 .7
- 7 .8
14.5
2.7
14.3
20.3
5.4
36.2
-2 4 .6
8.1
-5 .1
2.6
8.1
8.0
2.6
0.0
5.2
- 9 .7
- 2 .5
5.3
-2 .5
22.6
25.2
0.0
2.5
7.7
15.8
- 9 .3
7.6
2.5
7.5
18.2
0.0
12.6
26.2
28.6
4.6

Prior to
May 1983
revision

After
May 1983
revision

7.2%
0.0
3.5
0.0
- 9 .9
7.2
3.5
14.8
18.6
3.4
30.6
3.3
0.0
0.0
14.0
3.3
6.7
0.0
3.3
0.0
13.7
-3 .1
0.0
13.5
24.5
-1 1 .7
3.2
0.0
3.2
3.2
-3 .1
9.8
27.8
9.5
19.6
12.5
3.0
-1 3 .7
6.1

2.3%
1.1
2.3
- 0 .2
-0 .1
3.2
5.0
16.8
10.8
2.5
40.0
- 6 .9
- 3 .8
11.7
21.1
4.2
16.5
- 0 .5
5.5
1.3
1.1
- 4 .5
- 0 .6
7.6
8.8
3.0
7.4
4.4
3.7
5.8
1.6
6.3
14.7
8.7
15.3
11.7
-1 7 .9
6.8
21.6
9.1

is derived from the combined balance sheets of the
Federal Reserve Banks and the U.S. Treasury. It
reflects the combined actions of the Federal Reserve
and the U.S. Treasury that affect the amount of cur­
rency held by the public and reserves of depository
institutions.



Growth rate of the
St. Louis series less
the growth rate of
the revised BOG
series
-1 9 .8
16.9
12.3
-2 .5
- 7 .7
11.3
- 2 .3
-2 .5
9.5
2.9
- 3 .8
-1 7 .7
11.9
-1 6 .8
-1 8 .5
3.9
- 8 .5
3.1
- 5 .5
3.9
-1 0 .8
2.0
5.9
-1 0 .1
13.8
22.2
- 7 .4
- 1 .9
4.0
10.0
-1 0 .9
1.3
-1 2 .2
- 1 .2
2.9
-1 1 .7
30.5
19.4
7.0
-4 .5

Method o f reserve adjustment — The source base,
by itself, does not take into account another Federal
Reserve action that affects the money stock: changes in
reserve requirements. When reserve requirement
ratios are lowered (raised), a given amount of source
base can support a higher (lower) level of the money
17

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

Board o f Governors Series
Chart 1

R e s e r v e s S e rie s 11

Basic data: required reserves plus excess reserves —
The staff of the BOG does not use the source base in
deriving its monetary base and reserves series. In­
stead, it calculates the reserves of depository institu­
tions and adds currency to derive a monetary base
measure that is similar to one derived directly from the
source base.
Total reserves (BOG) include reserve balances of
depository institutions at Federal Reserve Banks. Be­
fore December 1980, only the vault cash of member
banks was included in the BOG reserves series. As
noted previously, the St. Louis reserves series in­
cludes the vault cash of nonmember commercial
banks, even before those institutions became subject
to reserve requirements of the Federal Reserve.

stock, holding constant all other factors that influence
the relationship between the money stock and the
source base.
To reflect the effects of changes in reserve require­
ments, the source base is adjusted for the amount of
reserves released or absorbed by these changes. This
adjustment involves adding a reserve adjustment mag­
nitude (RAM) to the source base, which produces a
measure called the adjusted monetary base. RAM is
simply the difference between what required reserves
would have been (given current deposit liabilities) if a
base period’s reserve requirements were still in effect
and the reserves that are actually required (given cur­
rent reserve requirements). Adding RAM to the
source base produces a series that indicates what the
source base would have been if the reserve require­
ment ratios had always been those of the base period.
The adjusted reserves series — The St. Louis Fed
derives its adjusted reserves series by subtracting cur­
rency held by the public from the adjusted monetary
base. Since only currency held by the public is sub­
tracted, vault cash of all depository institutions is in­
cluded in adjusted reserves.
Seasonal adjustment — The monetary base (source
base plus RAM) is seasonally adjusted directly. The
adjusted reserves series is adjusted for seasonal in­
fluences by subtracting seasonally adjusted currency of
the public from the seasonally adjusted monetary base.
18FRASER
Digitized for


Since December 1980, when the reserve require­
ment provisions of the Monetary Control Act of 1980
were implemented, all depository institutions have
been subject to reserve requirements of the Federal
Reserve. Since then, the BOG reserves measure in­
cludes all reserves (vault cash and reserve balances at
Federal Reserve Banks) of depository institutions
whose required reserves exceed their vault cash, plus
the required reserves of those institutions that hold
vault cash in excess of their required reserves. A large
proportion of nonmember depository institutions have
held vault cash in excess of their required reserves
since December 1980, because their reserve require­
ments are being increased gradually over eight years to
those specified in the Monetary Control Act. The dif­
ference between their vault cash and required reserves
is excluded from the BOG measure of reserves.
Method o f reserve adjustment — The BOG staff
revises its total reserves series after changes in reserve
requirements. For the period since the most recent
change in reserve requirements, the reserves series
equals the sum of required reserves and excess reserve
balances. The total reserves series for the period be­
fore the most recent reserve requirement change
equals excess reserve balances plus the sum of four
required reserves series, each for a different type of
institution and type of deposit.
Each of these four series is adjusted for breaks due to
changes in reserve requirements by use of a ratio
method. When reserve requirements are changed,
required reserves for each of the four series affected
are calculated using both the new and old reserve
requirements. The levels of required reserves prior to
the change are multiplied by the ratio of required
reserves under the new requirements to required re­

FEDERAL RESERVE BANK OF ST. LOUIS

serves under the old requirements.3 This procedure
indicates what reserves would have been if the current
structure of reserve requirements had been in effect
throughout the entire period, and if the deposit mix
within each of the four categories was the same as the
present deposit mix.
Prior to the revision in May 1983, the seasonally
adjusted total reserves series (BOG) was derived as the
sum of two series that each were seasonally adjusted
plus three other series that were not seasonally ad­
justed. The two seasonally adjusted components of
total reserves (BOG) were (1) required reserves on the
net transaction deposits of member banks, and (2) re­
quired reserves on the time and savings deposits of
member banks. Required reserves on transaction de­
posits of member banks were calculated by multiplying
seasonally adjusted transaction deposits of member
banks by the seasonally adjusted average reserve re­
quirement on those deposits. The same method was
used to derive seasonally adjusted required reserves
on the time and savings deposits of member banks:
seasonally adjusted deposits multiplied by their sea­
sonally adjusted average reserve requirements. The
following components were included in seasonally ad­
justed total reserves (BOG) on a not seasonally ad­
justed basis:
1. Total required reserves o f nonm em ber depository
institutions,
2. Required reserves o f Edge Act corporations, and
3. Excess reserves.

In the revised series, seasonal factors have been
derived directly for required reserves on net transac­
tion deposits of member banks and for required re­
serves on the time and savings deposits of member
banks. The method of multiplying seasonally adjusted
deposits by seasonally adjusted average reserve re­
quirements has been discontinued. Also, total re­
quired reserves of nonmember institutions are now
seasonally adjusted. The required reserves of Edge Act
corporations and excess reserves are still included in
seasonally adjusted total reserves (BOG) on a not sea­
sonally adjusted basis.

THE EFFEC T S OF D IFFER EN C ES IN
SEASONAL ADJUSTMENT
Financial analysts who compare trends in the St.
Louis and BOG reserves series typically focus on their
®Total reserves (BOG) are adjusted for changes in reserve require­
ments on liabilities of depository institutions other than transaction




JUNE/JULY 1983

seasonally adjusted growth rates. Since the methods
used to seasonally adjust the two reserves series differ
considerably, their growth rates can vary widely over
periods of a few months due to this difference alone.
The effects of differences in methods of seasonal
adjustment can be analyzed by comparing the differ­
ences in growth rates of the two seasonally adjusted
reserves series to similar differences in growth rates of
the data that are not seasonally adjusted. If some of the
variation in differences between growth rates of the not
seasonally adjusted data reflects differences in seasonal
patterns of the two reserves series, which are factored
out through seasonal adjustment, the differences be­
tween growth rates of the seasonally adjusted reserves
series would be less variable. Differences in the
methods of seasonal adjustment, however, may am­
plify rather than dampen the variation in the differ­
ences between the growth rates of these series. In this
case, variation in differences of seasonally adjusted
growth rates would be greater.
The results in table 2 indicate that, using data avail­
able before the recent revision in total reserves (BOG),
the differences between growth rates of the data that
are not seasonally adjusted are less variable. Some of
the differences in seasonally adjusted growth rates of
the two reserves series observed in recent years,
therefore, must be attributed to differences in
methods of seasonal adjustment. Table 2 also indicates
that the variability of differences between growth rates
of the two seasonally adjusted reserves series is smaller
in 1980 and 1982 for the revised data, and the same in
1981 for the old and revised data. Thus, the new
method of calculating seasonally adjusted total re­
serves (BOG) tends to reduce variation in the differ­
ence between growth rates of the two reserves series.
Therefore, differences between monthly growth rates
of the two reserves series may be less variable in the
future than those differences observed in the past.4
or time and savings deposits (such as commercial paper, Eurodollar
borrowings and ineligible acceptances) by subtracting from total
required reserves the sum of required reserves against these non­
deposit liabilities.
4As noted in the introduction, the average difference (in absolute
value) between monthly growth rates of the two reserves series
essentially was unaffected by the revision of total reserves (BOG).
The results in table 2, however, indicate that in two of the three
years, the standard deviations of the difference between growth
rates of the reserves series were reduced by the revision of total
reserves (BOG). The revision reduced some of the more extreme
differences between monthly growth rates, while increasing differ­
ences in other months. During the 36 months ending in December
1982, the number of months in which the growth rates differed (in
absolute value) by more than 20 percentage points declined from 5
to 1, but the number of months in which growth rates differed by
between 10 and 20 percentage points rose from 6 to 15.

19

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

Table 2

Standard Deviation of the Differences between Monthly
Growth Rates of the Reserves Series
Using data for total
reserves (BOG) available
in April 1983

Year

Not
seasonally
adjusted

Seasonally
adjusted

(1)
13.0

(2)

(3)

(4)

1980

13.3

13.0

11.5

Not
seasonally
adjusted

Seasonally
adjusted

1981

8.9

9.7

8.4

9.7

1982

11.2

14.8

10.6

10.7

The differences between the average seasonal fac­
tors of the two reserves series have a seasonal pattern.5
With data available before the recent revision in total
reserves (BOG), the average seasonal factors for ad­
justed reserves (St. Louis) were higher than those for
total reserves (BOG) in January, August, September
and December (chart 2). The BOG series tended to
grow faster (slower) than the St. Louis series during
those months in which the line in chart 2 rises (falls).
With the revised data for total reserves (BOG), the
difference between average seasonal factors still has a
seasonal pattern, but the seasonal pattern has been
changed. The average seasonal factor for adjusted re­
serves (St. Louis) is especially low relative to that for
the revised total reserves (BOG) in January and Febru­
ary. From the relatively low levels for those months,
the difference between the average seasonal factors
rises to a peak level in August. The pattern of these
differences between average seasonal factors in the
revised data indicates that total reserves (BOG) will
tend to grow faster than adjusted reserves (St. Louis)
from around January or February of each year through
August, if the differences between seasonal factors
continue to have the same pattern as in the past five
years.

THE EFFEC T S OF D IFFER EN C ES IN
TREATMENT OF VAULT CASH
Vault Cash o f Bound Institutions
Institutions with required reserves in excess of their
vault cash are referred to as “bound institutions. ” Their
5The seasonal factors are calculated for each reserves series by
dividing the not seasonally adjusted data by the seasonally adjusted
data.

20FRASER
Digitized for


Using revised data for
total reserves (BOG)
available since May 1983

vault cash is included in the BOG series with a twoweek lag, since it is included as reserves available for
meeting reserve requirements two weeks after the
vault cash is actually held. In contrast, all currency
held by the public and depository institutions is in­
cluded contemporaneously in the St. Louis series.
This difference in accounting for vault cash between
the St. Louis and BOG reserves series tends to pro­
duce differences in their monthly growth rates. The
magnitude of this effect can be estimated by generating
reserves series that treat vault cash identically for
bound institutions and calculating the extent to which
growth rates of the modified series differ. The easiest
way to do this is to subtract the monthly average levels
of vault cash held contemporaneously by bound in­
stitutions from adjusted reserves (St. Louis), not sea­
sonally adjusted, and add their vault cash held two
weeks earlier. When adjusted reserves (St. Louis) are
modified in this manner, the standard deviation of the
difference between the growth rates of the two re­
serves series over the 24 months ending in December
1982 is reduced by about 22 percent. Thus, the differ­
ence in treatment of vault cash at bound institutions
produces sizable differences in monthly growth rates of
the two reserves series. This effect is reduced, of
course, when periods longer than one month are ana­
lyzed; the two series differ by only two weeks in their
treatment of the timing of vault cash.

Vault Cash o f Nonbound Institutions
Some relatively large differences between the
monthly growth rates of the two reserves series in the
past three years have occurred when the reserve re­
quirements of nonmember institutions were in­
creased: D ecem ber 1980, Septem ber 1981 and

FEDERAL RESERVE BANK OF ST. LOUIS

Chart 2

A v e r a g e D ifferen ce b e t w e e n M o n t h ly S e a s o n a l Factors
for the R e se rv e S e rie s a

JUNE/JULY 1983

This explains why relatively large differences be­
tween monthly growth rates of the two reserves series
have occurred when nonmember reserve require­
ments have been increased. These differences are
directly attributable to the difference in methods of
adjusting for the effects of changes in reserve require­
ments.

A SPECIAL E F F E C T : GROWTH IN
MONEY MARKET DEPOSIT
ACCOUNTS

September 1982.6 In these months, total reserves
(BOG) grew faster than adjusted reserves (St. Louis),
with differences in annual growth rates ranging from
about 11 to 18 percentage points (table 1).
These large differences in monthly growth rates re­
sult from differences in the way the two reserves series
are adjusted for the effects of changes in reserve re­
quirements of nonbound institutions (those with vault
cash that exceeds their required reserves). The RAM
component of adjusted reserves (St. Louis) treats the
increase in reserve requirements of nonmembers as a
policy action that absorbs reserves by the full amount
of the increase in required reserves. The BOG method
of calculating total reserves treats only the increase in
required reserves of nonmembers above their holdings
o f vault cash as a policy action that absorbs reserves.
This difference reduces adjusted reserves (St. Louis)
relative to total reserves (BOG) by approximately the
amount that excess vault cash declines when nonmem­
ber reserve requirements are increased.7

6December 1980 was the first full month in which nonmember
depository institutions were subject to reserve requirements of the
Federal Reserve.
7See the appendix for an algebraic derivation of this result. From
August to September 1981, the rise in total reserves (BOG) less the
change in adjusted reserves (St. Louis), on a not seasonally ad­
justed basis, was about $600 million. Excess vault cash declined by
about $600 million in September 1981. From August to September
1982, the BOG series rose by about $600 million more than the St.
Louis series, approximately equal to the decline in excess vault
cash in September 1982. Data on excess vault cash prior to Decem­
ber 1980 are not available.




The rapid growth of money market deposit accounts
(MMDAs) also has produced large differences between
the growth rates of the St. Louis and BOG reserves
series, since those accounts were introduced in midDecember of last year.8 While MMDAs rose to $320.3
billion in March, time and savings deposits of all de­
pository institutions (not seasonally adjusted), exclud­
ing MMDAs, declined by $218.5 billion between last
November and March. Although the basic reserve re­
quirements against MMDAs are the same as those
against time and savings deposits, the actual required
reserves on personal MMDAs at member banks are
lower because reserve requirements against MMDAs
are not subject to the phase-in that applies to time and
savings deposits under the Monetary Control Act of
1980.9 With a high proportion of MMDAs being in the
personal category, and thus immediately exempt from
reserve requirements even at member banks, the shift
of funds into MMDAs from other categories of time
and savings deposits has reduced the average reserve
requirement on total time and savings deposits, inclu­
sive of MMDAs.
No adjustment has been made to total reserves
(BOG) for the reserves released as a result of growth of
personal MMDAs. In contrast, the growth of personal
MMDAs has caused the St. Louis RAM to rise. The

SM M DAs are exempt from maximum interest rates that may be paid
by depository institutions. The two major regulations on MMDAs
are a minimum denomination of $2,500 per account and a restric­
tion on guaranteeing an interest rate to a depositor for longer than
one month. Personal MMDAs (those held by individuals or non­
profit organizations) are exempt from reserve requirements, while
nonpersonal MMDAs are subject to a 3 percent reserve require­
ment.
®The basic reserve requirements against MMDAs (currently in
effect) and against time and savings deposits (after completion of
the phase-in) are 3 percent against nonpersonal deposits and zero
against personal deposits. At member banks, time and savings
deposits were subject to an average reserve requirement of about
3.3 percent prior to the start of the phase-in, which is now 75
percent complete at such institutions.

21

FEDERAL RESERVE BANK OF ST. LOUIS

category of time and savings deposits used in calculat­
ing required reserves with base period reserve re­
quirements includes personal and nonpersonal
MMDAs. Deposits in this category have risen since
MMDAs were authorized, and required reserves have
declined. Both changes have contributed to the growth
of RAM. Thus, the shift of deposits from those with
positive reserve requirements to personal MMDAs
generates a release of reserves. The growth of the St.
Louis reserves series since mid-December of last year
reflects that release of reserves through a rise in the
RAM component. The authorization of personal
MMDAs — subject immediately to a full phased-in
reserve requirement ratio equal to zero — is viewed as
a policy action that has resulted in a release of reserves.
This treatment reflects the principle that, if the reserve
requirements of the base period currently were in
effect, a higher source base would be needed to sup­
port the existing levels of checkable deposits and time
and savings deposits.
This difference in treatment of MMDAs in reserve
adjustment has produced substantially different
growth rates in the two reserve series in recent
months. From November 1982 through March of this
year, adjusted reserves (St. Louis) rose at about a 15
percent annual rate, compared with a 4.5 percent
growth rate for total reserves (BOG). Data available
prior to the May 1983 revision indicated that total
reserves (BOG) rose at a 1.5 percent rate from Novem­
ber of last year through March of this year. Thus, the
May 1983 revision raised the growth rate of total re­
serves (BOG), but not enough to narrow substantially
the gap between growth rates of the two reserves series
over this period of rapid MMDA growth.
The effect of this difference in treatment of MMDAs
on the growth of the reserves series can be gauged
indirectly by comparing the rise in the RAM since
November of last year to the adjustment to total re­
serves (BOG) for changes in reserve requirements.
Total reserves (BOG), not adjusted for changes in re­
serve requirements and not seasonally adjusted, de­
clined by $3,164 million from November 1982 to
March 1983. Total reserves, adjusted for changes in
reserve requirements, but not seasonally adjusted, de­
clined by only $449 million over the same period. The

22FRASER
Digitized for


JUNE/JULY 1983

adjustment for changes in reserve requirements,
therefore, is $2,715 million. This adjustment reflects
primarily the exemption from reserve requirements of
the first $2.1 million of reservable liabilities at each
depository institution, which became effective in late
December, and a phased reduction of member bank
reserve requirements in early March.
From November 1982 to March, RAM rose by
$4,333 million. Much of the difference between this
figure and the $2,715 million figure reflects reserves
released through shifts of deposits to personal
MMDAs.

CONCLUSIONS
Although the reserves series of the Federal Reserve
Bank of St. Louis and the Board of Governors have
similar descriptions, they often display substantially
different growth rates from month to month and over
periods of several months. There are several reasons
for these differences. First, different methods are used
to seasonally adjust the series. Growth rates of these
series can differ substantially over periods of a few
months solely for this reason.
Second, some of the largest monthly differences in
growth rates of these reserves series during the past
three years have occurred when reserve requirements
of nonmembers were increased. At those times, the St.
Louis measure of reserves declined relative to that of
the BOG. This effect is due to differences in the
methods used to account for vault cash and to adjust for
the effects of changes in reserve requirements.
Finally, the growth of money market deposit
accounts, which have been available at depository in­
stitutions since mid-December of last year, has raised
the growth of the monetary base and reserves series of
the St. Louis Fed relative to the BOG series. The
growth of MMDAs has produced a release of reserves
which has been reflected in the growth of the St. Louis
series in recent months. In contrast, no adjustment has
been made to the BOG series for the release of re­
serves that has resulted from the policy action of au­
thorizing MMDAs with a zero reserve requirement.

Appendix
PROCEDURES FOR CALCULATING
THE TWO MEASURES OF RESERVES
St. Louis Series
A reserve adjustment magnitude (RAM) is added to
the source base (currency in circulation plus reserve
balances of all depository institutions at Federal Re­
serve Banks).1 The base period for calculating RAM is
January 1976 through August 1980. Base period re­
serve requirements are average reserve requirement
ratios of member banks over that period on transaction
deposits (12.664 percent) and on total time and savings
deposits (3.1964 percent). RAM is calculated each
week by multiplying member bank transaction de­
posits held two weeks earlier by 0.12664, adding
0.031964 multiplied by member bank time and savings
deposits held two weeks earlier, and subtracting re­
quired reserves of all depository institutions for the
current week. Deposit liabilities held two weeks ear­
lier are used in calculating RAM because of lagged
reserve requirements, under which required reserves
for the current week are based on deposit liabilities
held two weeks earlier.
Seasonal factors are calculated for the adjusted
monetary base (source base plus RAM). Adjusted re­
serves on a seasonally adjusted basis are calculated by
subtracting currency held by the public, seasonally
adjusted, from the adjusted monetary base, seasonally
adjusted.

(1) R equired reserves on net transaction deposits of
m em ber banks, SA,
(2) Required reserves on tim e and savings deposits of
m em ber banks, SA,
(3) Total required reserves o f nonm em ber depository
institutions, SA,
(4) Total required reserves o f E d ge Act corporations,
not seasonally adjusted (NSA), and
(5) Excess reserve balances held at Federal Reserve
Banks, NSA (which excludes required clearing bal­
ances).

Each of the four series on required reserves is ad­
justed for the effects of changes in reserve require­
ments using a ratio method. At the time of a change in
the reserve requirements that apply to one of the four
series, required reserves are calculated using the new
and the old requirements. Data for that series of re­
quired reserves prior to the most recent change in
reserves requirements are multiplied by the ratio of
required reserves under the new requirements to re­
quired reserves under the old requirements.
The seasonally adjusted monetary base is derived as
follows:
(1) Total reserves, adjusted for the effects o f changes in
reserves requirem ents, SA,
(2) Plus vault cash o f nonm em ber com m ercial banks,
SA,
(3) Minus required reserves o f all nonm em ber institu­
tions held as vault cash, NSA,

Board o f Governors Series

(4) Plus excess vault cash o f nonm em ber depository
institutions other than com m ercial banks, NSA,

The following description of the BOG total reserves
and monetary base series reflects the methods used to
derive the series as revised in May 1983. Total re­
serves, seasonally adjusted (SA), equal the sum of the
following series:

(5) Plus currency held by th e public, SA,

d ep ository institutions maintain clearing balances at Federal Re­
serve Banks as a means of paying for the fees Federal Reserve
Banks now charge for services. Depository institutions receive
implicit interest on their clearing balances at the federal funds rate,
which may be used to pay the fees on services. Required clearing
balances are subtracted in computing the source base because
clearing balances are part of total reserve balances held by deposi­
tory institutions at Federal Reserve Banks, but are not related to
the levels of deposit liabilities.




(6) Plus required clearing balances, NSA.

Steps 2 through 5 involve adding components of
currency in circulation that are not included in total
reserves (BOG). Thus, this measure of the monetary
base includes reserve balances plus total currency in
circulation.

METHODS OF RESERVE
ADJUSTMENT
This section of the appendix describes the basic
differences between the two methods of adjusting the
23

FEDERAL RESERVE BANK OF ST. LOUIS

series on the monetary base and reserves of depository
institutions for the effects of changes in reserve re­
quirements. To simplify this illustration, the only dif­
ferences between the St. Louis and BOG series are
assumed to be differences in methods of adjusting for
the effects of changes in reserve requirements.
The illustration is based on the following assump­
tions:

JUNE/JULY 1983
(3) M B S L 2 = S B 2 + RAM 2
= C P 2 4- r2 D 2 + E 2 + (rx D 2 — r2 D 2)
= C P 2 + r x D 2 + E 2.

Adjusted reserves are calculated by subtracting cur­
rency in the hands of the public from the monetary
base:
(4) AR2 = M B SL a -

CP2

= r j D 2 + E 2.
1. The liabilities o f all depository institutions are sub­
je c t to one reserve requirem ent ratio.
2. The base period for th e reserve requirem ent ratio
used in calculating RAM (St. Louis series) is desig­
nated as period 1; it im m ediately precedes the cu r­
ren t period, designated as period 2.
3. T h e reserve requ irem ent ratio is different in the
cu rrent period (period 2) from what it was in the
prior period (period 1).

The following symbols are used in describing the
methods of reserve adjustment:
r 1; r2

— the reserve requirem ent ratio in periods 1
and 2

Calculation o f the BOG Series
Total reserves, not adjusted for the effects of changes
in reserve requirements, are specified as2
(5) TRN A! = ri D , + E ,
(6) TRNA2 = r2 D 2 + E 2.

For period 2, total reserves not adjusted for the effects
of changes in reserve requirements are the same as
total reserves adjusted. The method of adjusting for
the effects of changes in reserve requirements is ap­
plied to total reserves in period 1 as follows:

D [, D 2 — total deposit liabilities o f all institutions in
periods 1 and 2

(7) TR A , = n D , • ( - ! ^ - ) + E ,

E j, E 2 — excess reserves o f all depository institutions
in periods 1 and 2

= r2 D , + E j.

RAM

— reserve adjustm ent magnitude (St. Louis
series)

SB

— source base

CP

— currency in the hands o f the public

M B S L — monetary base (St. Louis series)
AR

— adjusted reserves (St. Louis series)

TRNA — total reserves, not adjusted for reserve re­
quirem ent changes (B O G series)
TRA

— total reserves, adjusted for the effects of
changes in reserv e requ irem en ts (B O G
series)

Calculation o f the St. Louis Series
In period 1, RAM is zero, since that is the base
period. RAM in period 2 is specified as
(1) RAM2 = r j D 2 — r2 D 2.

The source base can be specified as the sum of currency
held by the public, required reserves and excess re­
serves. For period 1,
(2) S B ! = C P , + rx

+ E j.

Since RAM is zero in period 1, the source base is the
same as the monetary base. In period 2,
Digitized for
24FRASER


rl

U 2

Table A1 summarizes the basic differences between
the St. Louis and BOG series in the methods of adjust­
ing for the effects of changes in reserve requirements.
For the St. Louis series, adjusted reserves are calcu­
lated for period 2 as though the reserve requirement
ratio of period 1 was in effect. For the BOG series, total
reserves are calculated for period 1 as though the re­
serve requirement ratio of period 2 was in effect.

IMPLICATIONS OF D IFFER EN CES IN
TREATMENT OF THE VAULT CASH
OF NONBOUND INSTITUTIONS FOR
THE RESERVES MEASURES
Depository institutions with vault cash in excess of
their required reserves are called nonbound institu­
tions. Their vault cash in excess of required reserves is
excluded from the BOG reserves series, but included
in the St. Louis series. The difference in treatment of

^The excess reserves in total reserves (BOG) includes only excess
reserve balances at Federal Reserve Banks. Excess vault cash at
institutions with greater vault cash than required reserves is ex­
cluded from the BOG measure of total reserves, but is included in
the St. Louis measure of adjusted reserves. This difference is
ignored in this section to simplify the illustration.

JUNE/JULY 1983

FEDERAL RESERVE BANK OF ST. LOUIS

V j1, V 2 — vault cash o f nonbound institutions in
periods 1 and 2

Table A1

Methods of Calculating Reserves of
Depository Institutions Adjusted for
Changes in Reserve Requirements
Period

St. Louis series

BOG series

1

r, D, + E 1

r2 D, + E,

2

fi D2 + E2

r2 D2 + E2

Adjusted reserves (St. Louis) in period 1 are given
by
(8) ARj = r B D f + r ? D ? + E f + E f .

Because nonbound institutions hold all of their re­
quired and excess reserves as vault cash, their excess
reserves can be expressed as
(9) E ? = V ? -

vault cash of nonbound institutions causes the St.
Louis reserves series to decline relative to the BOG
series when the reserve requirements of nonbound
institutions increase.
This effect is illustrated under the following assump­
tions, using some additional terms. Assumptions:
1. The base period for calculating RAM is period 1.
2. Required reserves o f nonbound institutions are in­
creased in period 2, but vault cash still exceeds
required reserves at all o f the previously nonbound
institutions.
3. This increase in required reserves o f nonbound in­
stitutions has no effect on their demand for reserves
relative to their deposit liabilities. They have the
same level o f deposit liabilities and hold the same
amount o f vault cash in period 2 as in period 1. W ith
higher reserve requ irem en ts, they simply have
higher required reserves and lower excess vault
cash.
4. Bound institutions have the same deposit liabilities
and excess reserves in periods 1 and 2.

Additional terms:
rB

— reserve requirem ent ratio for bound in­
stitutions

r f , r2

— reserve requ irem ent ratio for nonbound
institutions in periods 1 and 2

D f , D f — deposit liabilities o f bound institutions in
periods 1 and 2
D j1, D 2 — deposit liabilities o f nonbound institu­
tions in periods 1 and 2
E f , E f — excess reserves o f bound institutions in
periods 1 and 2




r? D ? .

Substituting this expression for their excess reserves
into the equation for adjusted reserves yields
(10) ARi = rB D f + E f + V ?.

Using the same terms for period 2:
(11) AR2 = rB D f + E f 4- V ?
+ (rB D f + r j1 D 2 — rB D f -

r£ D 2 )

= rB D f + E f + V £ + (rj1 -

r£) D £ .

By the assumptions above,
(12) AR2 -

ARi = (r? -

r£) D £ ,

which is negative, because r2 is larger than rf.
Because total reserves (BOG) exclude excess vault
cash, values of that series in periods 1 and 2 can be
written as
(13) TRA2 = rB D f -I- r£ D £ + E f
(14) TR A t = rB D f + r? D ? ( ^ j ^ ) + E f
«i D 2
= rB D f + r£ D f + E f .

Given our assumptions,
(15) TR A j -

TRA2 = 0.

Thus, under our assumptions, a rise in the reserve
requirement ratio of nonbound institutions causes ad­
justed reserves (St. Louis) to decline and total reserves
(BOG) to remain unchanged. Since most of the non­
bound institutions are nonmembers, this analysis indi­
cates why a rise in reserve requirements of nonmem­
bers tends to reduce adjusted reserves (St. Louis) rela­
tive to total reserves (BOG).

25

The FOMC in 1982: De-emphasizing Ml
DANIEL L. THORNTON

JL

HE year 1982 was marked by rapid and variable
growth of the monetary aggregates. The growth of Ml,
the narrow monetary aggregate, was up sharply from
1981, while M2 growth was slightly above the previous
year’s rate. Of the three targeted aggregates, only M3
growth was lower in 1982 than in 1981. Moreover,
1982 marked the first time since the Federal Open
Market Committee (hereafter FOMC or Committee)
adopted its new procedures in October 1979 that the
fourth-quarter-to-fourth-quarter growth rate of M l
accelerated.1
As was the case in 1981, the Committee faced un­
usual uncertainties regarding the relative behavior of
M l and M2 during the year associated with various
technical factors, regulatory changes and financial in­
novations. Furthermore, the income velocities of the
monetary aggregates, especially that of M l, declined
relative to their historical norms.2 Because of these
difficulties, the Committee had considerable discus­
sion about the weight that should be assigned to M l
Note: Citations referred to as “Record” are to the “Record of Policy
Actions of the Federal Open Market Committee” found in various
issues of the Federal Reserve Bulletin.
'F o r a description of the operating procedure, see R. Alton Gilbert
and Michael Trebing, “The FOMC in 1980: A Year of Reserve
Targeting,” this Review (August/September 1981), pp. 2-22; and
Richard W. Lang, “The FOMC in 1979: Introducing Reserve
Targeting,” this Review (March 1980), pp. 2-25.
"The income velocity of a monetary aggregate is given by the ratio of
nominal GNP to the aggregate. It indicates the number of times
each unit of nominal money “turns over” in producing this year’s
final output. This conclusion about the record decline in velocity
was based on the fact that M l growth had been rapid in the first
quarter compared with what would have been predicted on the
basis of the actual behavior of nominal GNP and interest rates. This
interpretation was supported by the growth in relatively lowinterest-yielding savings deposits. See “Record” (June 1982), pp.
366-67.

26




and M2 as a guide to policy. Ultimately, it decided to
suspend setting explicit growth objectives for M1 dur­
ing the fourth quarter of the year. This article will
review the factors affecting the long- and short-run
policy decisions of the Committee during 1982, includ­
ing those leading up to the decision to suspend setting
an explicit target for Ml.

ANNUAL TARGETS FOR 1982
The Full Employment and Balanced Growth Act of
1978 (the Humphrey-Hawkins Act) requires the Board
of Governors to transmit to Congress, each February
and July, reports on the objectives for growth rate
ranges for monetary and credit aggregates over the
current calendar year and, in the case of the July
report, over the following calendar year as well.3 The
Committee has chosen to establish ranges from the
fourth quarter of the previous year to the fourth quar­
ter of the current year.4
These ranges must be reported to Congress each February and
July, although the Act provides that the Board and the Committee
may reconsider the annual ranges at any time. The period to which
the annual ranges apply, however, may not be changed. The base
period that the Committee has chosen (the fourth quarter of the
previous year) would remain the same even if the Committee
decided to change the desired growth rates of the aggregates for the
year.
4Before 1979, the Committee adopted one-year growth rates each
quarter, and the base period for the annual targets announced each
quarter was brought forward to the most recent quarter. This
method resulted in a problem referred to as “base drift. ” Growth in
aggregates above (below) an annual growth range in a quarter
would raise (lower) the base level for calculating the next annual
growth path. The specification of annual objectives in terms of
calendar year growth rates, which eliminates the base drift prob­
lem within a calendar year, does not solve this problem from one
calendar year to the next, since new ranges are established from the
end of each calendar year.

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

Organization of the Committee in 1982
The Federal Open M arket C om m ittee (FO M C ) con­
sists o f 12 m em bers: the seven m em bers of the Federal
R eserve Board o f Governors and five of the 12 Federal
R eserve Bank presidents. The Chairman of the Board of
Governors is, by tradition, also chairman o f the Com m it­
tee. The president o f the New York Federal Reserve Bank
is, also by tradition, its vice chairman. All Federal R e­
serve Bank presidents attend C om m ittee m eetings and
present their views, but only those presidents who are
mem bers o f the C om m ittee may cast votes. Fou r m em ­
berships rotate among Bank presidents and are held for
one-year term s beginning March 1 o f each year. The
president of the New York Federal Reserve Bank is a
perm anent voting m em ber o f the Com m ittee.
M em bers o f the Board of Governors at the beginning of
1982 included Chairm an Paul A. Volcker, Preston M ar­
tin, Henry C. W allich, J. Charles Partee, Nancy H. T e e ­
ters, E m m ett J. Rice and Lyle E . G ram ley.1 The follow­
ing presidents served on the C om m ittee during January
and February 1982: Anthony M. Solomon (New York),
Edward G. Boehne (Philadelphia), R obert H. Boykin
(Dallas), E . G erald Corrigan (M inneapolis) and Silas
Keehn (Chicago). The C om m ittee was reorganized in
March and the four rotating positions w ere filled by: John
J. Balles (San Francisco), R obert P. Black (Richmond),
William F . Ford (Atlanta) and Karen Horn (Cleveland).2
The Com m ittee m et eight tim es during 1982 to dis­
cuss, among other things, econom ic trends and to decide
upon the future course o f open market operations.3 As in
previous years, however, telephone or telegram consul­
tations were held occasionally betw een scheduled m eet­
ings. During each regularly scheduled m eeting, a d irec­
tive was issued to the Federal Reserve Bank o f New York.
Each directive contained a short review o f econom ic de­
velopments, the general econom ic goals sought by the
Com m ittee, and instructions to the M anager o f the Sys­
tem Open M arket Account at the New York Bank for the
conduct o f open m arket operations. T hese instructions
were stated in term s o f short-term rates o f growth o f M l
and M 2 that w ere considered to b e consistent with de­
sired longer-run growth rates o f the monetary aggre­
g a te s.4 T h e C o m m itte e also specified in te rm e etin g
ranges for the federal funds rate. T hese ranges provide a
mechanism for initiating consultations betw een m eetings
w henever it appears that the constraint on the federal

'Governor Frederick H. Schultz’s term expired January 1982.
He was replaced by Preston Martin.
2Karen Horn took office as President of the Cleveland Bank May
1, 1982, and subsequently became a voting member of the
FOMC. Mr. Winn voted as an alternate member in March.
3No formal meetings were held in January, April, June or
September of 1982.
4In October 1982, short-run objectives for M1 were dropped and
short-run objectives for M3 were adopted.




funds rate is proving inconsistent with the objectives for
the behavior o f the monetary aggregates.
The Account M anager has the m ajor responsibility for
formulating plans regarding the timing, types and amount
of daily buying and selling o f securities in fulfilling the
Com m ittee’s directive. Each morning the Manager and
his staff plan the open market operations for that day. This
plan is developed on th e basis o f the C om m ittee’s direc­
tive and the latest developm ents affecting money and
credit market conditions, growth o f the monetary aggre­
gates and bank reserve conditions. T h e M anager then
informs staff m em bers o f the Board o f Governors and one
voting president about present m arket conditions and
open market operations that he proposes to execute that
day. O ther m em bers o f the C om m ittee are informed of
the daily plan by wire.
T he directives issued by the C om m ittee and a sum­
mary o f the reasons for the C om m ittee actions are pub­
lished in the “Record o f Policy Actions o f the Federal
Open M arket Com m ittee. ” T he “R ecord” for each m eet­
ing is released a few days after the following Com m ittee
meeting. Soon after its release, the “R ecord” appears in
the Federal Reserve Bulletin. In addition, “Records” for
the entire year are published in the Annual Report of the
Board o f Governors. T he “R ecord” for each m eeting dur­
ing 1982 included:
1) A staff summary o f recen t econom ic developm ents
— such as changes in prices, em ploym ent, indus­
trial production and com ponents o f the national in­
com e accounts — and projections o f general price,
output and em ploym ent developm ents for the year
ahead;
2) A summary o f recen t international financial de­
velopm ents and the U .S. foreign trade balance;
3) A summary o f recen t credit m arket conditions and
recent interest rate movem ents;
4) A summary o f open m arket operations, growth o f
monetary aggregates and bank reserves, and money
market conditions since the previous m eeting;
5) A summary o f the C om m ittee’s discussion o f current
and prospective econom ic and financial conditions
and the cu rrent policy considerations, including
m oney market conditions and the m ovem ent of
monetary aggregates;
6) Conclusions o f the Com m ittee;
7) A policy directive issued by the C om m ittee to the
Federal Reserve Bank o f New York;
8) A list o f the m em bers’ voting positions and any
dissenting com m ents;
9) A description o f any actions and consultations that
may have occurred betw een the regularly sched­
uled meetings.

27

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

Table 1

FOMC Operating Ranges — 1982
Short-Run O perating Ranges

Date of
meeting
February 1-2, 1982
March 29-30a
May 18b

Periods to
which monetary

Federal funds
rate range

M1

growth paths apply

M2

M3

January-March

no further
growth1

around 8%

no range set

no change

March-\June

about 3%

about 8%

no range set

10-15%

March-June

reaffirmed targets
about 9%

no range set

12-16%

no change

June-September

about 5%

August 24d

7-11%

June-September

reaffirmed targets

October 5®

7 -1 0 1/2%

September-December

no specific
objective

around 8V2-91/2%

around 81/2-91/2%

November 16'

6-10%

September-December

no specific
objective

around 91/2%

around 91/2%

December-March

no specific
objective

around 914%

about 8%

June 30-0uly 1c

December 20-219

no change

Long-Run O perating Ranges
Date of
meeting

Target Period

M2

M1

M3

February 1-2, 1982h

IV/81— IV/82

2y^5'/2%

6-9%

61/2-91/2%

July 15
(telephone meeting)'

IV/81 — IV/82

reaffirmed above
range

reaffirmed above
range

reaffirmed above
range

1At its December 1981 meeting the Committee set an objective for M1 of "around 4 to 5 percent,” however, the surge in M1 growth in January
prompted the Committee to set a “ no further growth” objective at its February meeting.

At its February meeting, the Committee completed
the review, begun at its December 1981 meeting, of
the annual targets for the monetary and credit aggre­
gates for 1982. It remained committed to its long­
standing goal of restraining the growth of money and
credit to reduce further the rate of inflation. Neverthe­
less, Committee members disagreed about the precise
ranges to set for the various monetary aggregates. Most
members favored reaffirming the ranges for M l that
had been tentatively adopted at the July 1981 meeting.
A substantial number, however, favored a somewhat
higher range for M2 based on the belief that various
developments during the year would likely boost the
growth of M2 relative to M l.5 Also, it was generally
agreed to give considerable weight to M2 in interpret­
ing developments during the year.6

In setting its growth range for M l, the Committee
argued that the growth in “other checkable deposits,”
which had accelerated during January and which was
in large part responsible for the rapid January growth
of M l, was likely to be temporary, and that the rela­
tionship between the M1 growth and the nominal GNP
growth likely would be closer to its historical pattern
during 1982. On this assumption, the Committee
argued that it would be acceptable for M l to grow at a
rate near the upper end of its annual range during
1982. The Committee also expected that the growth of
M2 would be high in its range, although somewhat
below that of 1981.7 At the end of the discussion, the
Committee reaffirmed its tentative ranges for M l and
M2. These ranges are presented in table 1.

5At its midyear review of the annual ranges, the Committee estab­
lishes tentative ranges for the monetary aggregates for the next
year — measured from the fourth quarter of the current year to the
fourth quarter of the following year.

7Indeed, the Committee believed that the growth in M2 might
meet or exceed the upper end of its range if the personal savings
rate grew more rapidly than anticipated, or if depository institu­
tions attracted an exceptionally large flow of funds into IRAs from
sources outside of M2. See “Record” (April 1982), p. 233.

6See “Record” (April 1982), pp. 232-33.

28



FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

Table 1 (continued)

Footnotes — Dissents to FOMC Actions_____________________________________
aMessrs. Black and Wallich dissented from this action because they favored specification of somewhat lower rates for monetary growth from
March to June than those adopted by the Committee, which would be associated with a relatively prompt return of M1 growth to its range for
the year.
Mr. Black believed that continued growth of M1 above its longer-run range for any extended period would adversely affect economic
activity by exacerbating inflationary expectations and weakening markets for longer-term securities; for that reason, he felt that it was
particularly important to resist any surge in growth of M1 that might develop in April.
In Mr. Wallich’s opinion, it would be desirable to restrain the pace of prospective recovery in economic activity consistent with some
reduction in the unemployment rate to sustain a degree of pressure for the continuation of the reduction in the underlying rate of inflation.
bMrs. Teeters dissented from this action because she favored specification of somewhat higher rates of monetary growth from March to June
with the objective of improving liquidity and easing financial pressures. In her opinion, the time had come to foster lower and less variable
interest rates in order to enhance prospects for significant recovery in output and employment.
cMessrs. Black, Ford and Wallich dissented from this action because they favored a policy for the period immediately ahead that was firmly
directed toward bringing the growth of M1 down to its range for 1982 by the end of the year. They were concerned that accommodation of
relatively rapid growth over the summer months might jeopardize achievement of the monetary objectives for the year and thus would risk
exacerbating inflationary expectations. Accordingly, they believed that tendencies toward rapid monetary expansion in the months
immediately ahead should be met by greater pressures on bank reserve positions and in the money market.
Mrs. Teeters dissented from this action because she favored specification of somewhat higher rates for monetary growth during the third
quarter along with an approach to operations early in the period that would clearly signal an easing in policy. In her opinion, policy at this
point should be directed toward exerting downward pressure on short-term interest rates in order to promote recovery in output and
employment.
dMr. Wallich dissented from this action because he favored an approach to operations early in the period that would lessen the chances of
short-term interest rates remaining below the prevailing discount rate or falling further below it. He was concerned that such interest rate
behavior would tend to accelerate monetary expansion and that the necessary restraint of reserve growth to curb such expansion might
lead to a sizable rebound in short-term rates with adverse implications for business and consumer confidence.
eMr. Black dissented from this action because he preferred to direct operations in the period immediately ahead toward restraining monetary
growth. Although he was mindful of the current difficulties of interpreting the behavior of M1, he was concerned that the recent strength in
M1 might be followed by still more rapid growth in lagged response to the substantial decline in short-term interest rates that had occurred in
the summer, which could require even more restrictive operations later.
Mr. Ford dissented from this action because he preferred a policy for the period immediately ahead that was more firmly directed toward
restraining monetary growth, although he recognized that the behavior of M1 in particular would be difficult to interpret. He was concerned
that the Committee’s policy directive might be misinterpreted in ways that could adversely affect pursuit of the System’s longer-run
anti-inflationary objectives, particularly in the context of a highly expansive fiscal policy program.
Mrs. Horn dissented from this action because she preferred to continue setting a specific objective for growth of M1, as well as for M2,
over the current quarter, notwithstanding the problems of interpreting its behavior. In setting a target for M1, she would tolerate faster growth
early in the period, owing to the uncertain impact of the proceeds from maturing all-savers certificates, and would give greater weight to the
behavior of M2 for some weeks after the introduction of the new instrument at depository institutions.
' Mr. Ford dissented from this action because he believed that it ran the risk of complementing very large budget deficits with substantial
increases in the supply of money. In his view, the result would be an overly stimulative combination of policies that could rekindle inflation
and drive up interest rates during 1983.
9Mr. Black dissented because he preferred to direct policy in the weeks immediately ahead toward ensuring that the growth of M1,
abstracting from temporary effects of the introduction of new money market deposit accounts, would moderate from the extremely rapid
rate of recent months. While recognizing the difficulties in interpreting M1 currently, he was concerned that excessive underlying growth in
that aggregate might reverse the progress achieved in reducing inflation and inflationary expectations and lead to substantially weaker
markets for long-term securities.
Mr. Ford dissented from this action because he continued to prefer a policy for the current period that was more firmly directed toward
restraining monetary growth, after allowance for the short-run impact of the introduction of new money market deposit accounts. He
remained concerned that rapid expansion in the supply of money together with very large budget deficits would produce an overly
stimulative combination of policies that could rekindle inflation and inflationary expectations and lead to higher interest rates during 1983
and 1984.
hMrs. Teeters dissented from this action because she believed that somewhat higher monetary growth over the year ahead was needed to
promote adequate expansion in economic activity and a reduction in the rate of unemployment. Specifically, she favored a range for M1 that
was at least Vi percentage point higher than that adopted by the Committee and a range for M2 that provided for somewhat greater growth
in the broader aggregate relative to that in M1.
1Mrs. Teeters dissented from this action because she favored an explicit statement that growth of M1 above the upper end of the
Committee’s range for 1982 by 1 percentage point, or even as much as 1V4 percentage points, might be acceptable. In her opinion, it was
important to indicate the acceptable degree of growth of M1 above the range in order to foster market behavior that would lower interest
rates and enhance the prospects for sustaining recovery in output and employment.




29

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

Table 2

Planned Growth of Monetary Aggregates for 1982
(percentage changes, fourth-quarter-to-fourth-quarter)
Aggregate1

Actual 1981
growth rate2

Proposed range
for 1982

Actual 1982
growth rate2

M1

5.0%

2.5-5.5%

M2

9.5

6.0-9.0

9.9

M3

11.4

6.5-9.5

10.4

8.5%

1M1 is defined as currency plus private demand deposits and other checkable deposits at depository
institutions exclusive of deposits due to foreign commercial banks and official institutions, plus travelers
checks of non-bank issuers.
M2 is M1 plus savings and small-denomination time deposits at all depository institutions, shares in
money market mutual funds, overnight repurchase agreements issued by commercial banks and
overnight Eurodollar deposits held by U.S. residents at Caribbean branches of U.S. banks.
M3 is M2 plus large time deposits at all depository institutions and term repurchase agreements
issued by commercial banks and savings and loan associations.
2Data as revised by Board of Governors in January 1983.

Actual Money Growth in 1982
As shown in table 2, all three of the monetary aggre­
gates exceeded their target ranges during 1982.8 Their
patterns of growth relative to their ranges, however,
were considerably different, as can be seen in chart 1.
Both M l and M2 were above their targeted ranges
nearly all of the year. In contrast, M3 growth was
within its range during the first half of 1982 and above it
during the second half.
Although both M1 and M2 were above their target
ranges throughout the year, their growth rates dis­
played different patterns. While the quarter-toquarter growth of M2 during 1982 was less stable than
that of 1981, it was stable compared with the quarterto-quarter growth of M l. M l grew rapidly in January
and was fairly flat until July, when it began a growth
spurt that accelerated markedly in October. This pat­
tern of M l growth was basically consistent with the
Committee’s short-run objectives for the year.

SHORT-RUN POLICY DIRECTIVES
FOR 1982
The announced annual target ranges for the mone­
tary aggregates provide a basis on which the FOMC
®The definition of M2 was changed effective February 14, 1983, to
include tax-exempt money market funds and to exclude all IRA/
Keogh balances at depository institutions and money market
mutual funds. These changes also affected M3. Thus, data available
January 20, 1983, were used. The growth rates of M l, M2 and M3
will differ from those reported from revisions after February 14,
1983.

30



chooses its short-run policy objectives during the year.
The short-run policy directives, however, are the ones
that influence the day-to-day implementation of
monetary policy. The Committee issues these direc­
tives for implementation by the Manager of the Open
Market Account at the Federal Reserve Bank of New
York.
During 1982, the Committee specified short-run
growth rates for M l, M2 and M3.9 It also specified
intermeeting ranges for the federal funds rate as a
mechanism for initiating further consultations in
periods between regularly scheduled meetings.10
These intermeeting ranges and the actual federal funds
rate are presented in chart 2. The growth rate targets
for the monetary aggregates and the intermeeting
ranges for the federal funds rate that the Committee
specified during 1982 appear in table 1.
As in the previous year, discussions pertaining to
short-run policy decisions in 1982 were marked by
considerable uncertainty about both the effect of var­
ious regulatory changes and financial innovations on
the growth rates of the monetary aggregates and the

‘The short-run growth rate target for M l was dropped at the Octo­
ber meeting and a short-run target for M3 was introduced.
10If movements of the federal funds rate within the range appear to
be inconsistent with the short-run objectives for the monetary
aggregates and related reserve paths during the intermeeting
period, the Manager of Domestic Operations at the Federal Re­
serve Bank of New York is to notify the Chairman, who in turn
decides whether the situation calls for supplementary instructions
from the Committee.

C h a rt 1

Long-Run O perating Ranges for the Period IV /8 1 - IV /8 2
B ill io n s of d o l l a r s

B il lio n s of d o l l a r s

480

480

M l range

470

470

460

460

Actual monthly levels
450

450

440

440

430

430
B ill io n s of d o l la r s

B il lio n s of d o l la r s

2200

2200

*12 rangi j
-"9 %

1770

1970

\
\

*
1920

«•>»

6%

1920

\
\
\

\
\
i

\

\
\
\

-

„•

Actual monthly levels

1870

\
\
\

1870

1820

1820

1970

1770

B illio n s of d o l l a r s

B ill io n s of d o l la r s

2420

2420

\13 rang t
''m

2370

2370

Actual monthly levels
2320

’

^

m

2320

*

2270

2270

2220

2220

2170

2170

2120

2120
O CT.

NOV.

DEC.

JAN .

1981
N O T E : D a ta r e v is e d J a n u a r y 198 3.




FEB.

M AR.

APR.

MAY

JUNE

JULY

1982

AUG.

SEPT.

O CT.

NOV.

DEC.

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

F O M C Ranges for Federal Funds Rate

JA N .

FEB.

M AR.

APR.

M AY

JU N E

JULY

AUG.

SEPT.

O C T.

NOV.

DEC.

JAN .

FEB.

M AR.

APR.

M AY

1981

JU N E

JULY

AUG.

SEPT.

OCT.

NOV.

DEC.

1982

N O T E : R a te s a r e c a lc u la te d a s w e e k ly a v e r a g e s o f d a i ly r a te s . A t e a c h m e e tin g th e c o m m itte e s p e c ifie d a r a n g e f o r th e fe d e r a l fu n d s ra te . T h e se r a n g e s a r e in d ic a t e d
f o r t h e f ir s t f u ll w e e k t h e y w e r e in e ffe c t.

relative weight that should be given to M l and M2 in
implementing the Committee’s short-run policy
decisions.11 Indeed, the relative importance of M2 and
M l for short-run policy purposes shifted during the
year.
Nevertheless, just as in 1981, short-run movements
in the aggregates during 1982 followed their short-run
target paths. This correspondence between the target
paths and actual growth of the aggregates is illustrated
in chart 3, which shows the short-run target ranges and
actual levels of M l and M2, respectively, based on
first-published data. First-published data give a more
n See Daniel L. Thornton, “The FOMC in 1981: Monetary Control
in a Changing Financial Environment,” this Review (April 1982),
pp. 3-22.
12Because of a definitional change, data for M2 prior to February 5,
1982, are not first-published. Prior to that date, M2 included
repurchase agreements and isntitution-only money market
mutual funds.

32



accurate representation of the Committee’s short-run
policy decisions based on information available at the
time.12 Chart 3 shows that short-run targets for Ml
were specified only for the first three quarters of the
year. During its October meeting, the Committee de­
cided to place much less weight than usual on the
narrow aggregate and not set a specific objective for its
growth. At this time, the Committee began setting
short-run targets for M3.

First Quarter
The short-run targets for the first quarter of 1982
were made against a backdrop of rapid expansion in M2
and M l from November 1981. The growth of both
monetary aggregates accelerated during January 1982,
especially that of M l. The Committee believed that
the rapid growth in the demand for components of Ml
would abate during the ensuing months. It noted that if

C h a rt 3

Short-Term and Long-Term G row th Objectives Based on First Published D ata
B illio n s of d o l la r s

B i l l i o n s of do lla r s

2025

* 2 rangi

2025

\

91
/

200 0

200 0
*

1975

1975

1950

1950

Actual monthly levels
1925

5^

"

'

—-

6%

1900

1925

1900

8%
1875

1875

8%..
1850

***

1850

1825

1825

2^
1800

1800

1775
O CT.

NOV.

1981

DEC.

JAN .

FEB.

M AR.

APR.

M AY

JUNE

JULY

AUG.

SEPT.

O CT.

NOV.

DEC.

1982

N O T E : L o n g d a s h e d lin e s r e p r e s e n t th e lo n g te rm g r o w th o b je c tiv e ( o r th e p e r io d I V / 81 - 1V / 8 2 . S h o rt d a s h e d lin e s r e p r e s e n t th e
c u r re n t s h o r t- te r m g r o w th o b je c tiv e s . D a ta , e x c e p t ( o r M 2 p u b lis h e d b e fo r e 2 / 5 / 8 2 , a r e f ir s t p u b lis h e d fro m th e B o a rd 's
H -6 r e le a s e . P r io r to 2 / 5 / 8 2 , M 2 in c lu d e d RPs a n d i n s t it u t io n - o n ly M M M F s .




1775

FEDERAL RESERVE BANK OF ST. LOUIS

such a decline were not forthcoming, the income
velocity of M l would decline at a postwar record rate,
based on the then-projected growth of nominal GNP
for the first quarter. Thus, the Committee established
growth paths for M l and M2 that, if achieved, would
move these aggregates closer to the upper limit of their
annual target ranges. Specifically, the Committee
sought no further growth in M1 from January to March
and growth of M2 at an annual rate of around 8 percent.
It was agreed that some decline in M1 would be accept­
able in the context of reduced pressures in the money
market.13

Second Quarter
Continued uncertainty about the relative behavior
of M1 and M2 marked the short-run policy decisions
for the second quarter. Staff analysis continued to sug­
gest that the demand for money might be expected to
moderate significantly in the second quarter. Further­
more, the Committee was concerned that technical
problems associated with the federal income tax dead­
line in April might result in an April bulge in M l
growth. It was understood that most, if not all, of the
M l growth for the second quarter might occur during
April.14
Given these technical factors and given uncertain­
ties about near-term economic prospects and other
factors affecting the monetary aggregates, most mem­
bers of the Committee favored actions that would per­
mit modest growth in M l over the second quarter.
Thus, the Committee set a short-run target for M l of
about 3 percent, while maintaining the short-run
target growth rate for M2 at its first-quarter rate. Fur­
thermore, it noted that deviations from these targets
should be evaluated in the light that M2 was less likely
than M1 to be affected by deposit shifts and technical
factors over the second quarter.15

Third Quarter
In setting its short-run objectives for the third quar­
ter, the Committee noted that the growth of M1 and

13See “Record” (April 1982), p. 234.
I4See “Record” (June 1982), p. 368.
15The Committee reevaluated its position for the second quarter at
its May meeting. Most members agreed that somewhat more
rapid growth of M l might be acceptable if it appeared to be
associated with a continued desire of the public to build up liquid­
ity, and if the growth of M2 was near its specified range. See
“Record” (July 1982), p. 420.

34



JUNE/JULY 1983

M2 for the whole period from March to June appeared
to be in line with its objectives for that period (see chart
3). The Committee was increasingly pessimistic, how­
ever, about the outlook for the economy, and it con­
tinued to be concerned about the uncertainty over the
public’s demand for liquidity and precautionary bal­
ances. Additionally, it was concerned that the midyear
reduction in withholding rates for federal income taxes
and the scheduled cost-of-living increase in social
security payments would lead to a bulge in M l during
July. After a discussion of these factors, most of the
Committee members agreed that they would accept
somewhat faster monetary growth in the third quarter
if the demand for liquidity and precautionary balances
did not ease as anticipated. Thus, the Committee
voted for faster growth for both M l and M2 from the
second to the third quarter, increasing the M l target
from about 3 percent to about 5 percent and increasing
the M2 target from about 8 percent to about 9
percent.16

De-emphasizing M l
At the October meeting, when the short-run objec­
tives for the fourth quarter were first considered, a
number of new considerations concerning the state of
the economy and financial markets emerged. The
Committee was concerned that the general worsening
of the world economy and financial problems associ­
ated with large accumulated external debts of develop­
ing countries in recent years had contributed to an
atmosphere of uncertainty that was reflected in the
exchange value of the dollar, among other things. This,
in turn, had serious implications for U.S. export indus­
tries and for the ability of foreign governments to pur­
sue flexible monetary policies. Also, the Committee
was concerned that the U.S. banking system had been
subjected to pressures associated with the general un­
easiness about further credit problems both domesti­
cally and internationally. The result was a general
widening of risk premiums, with interest rates on pri­
vate securities generally falling less than the rates on
Treasury issues from July to September. It noted that
short-term interest rates had tended to move up in the
weeks just before the meeting. Furthermore, the com­
mittee noted that the widely held expectations of
a spring or summer recovery had been disappointed,
and there were no signs of a strengthening in the
economy.17
16Three members dissented from this action because they favored a
policy of bringing growth of M l down to its range for 1982 by
year-end. See table 1 or “Record” (September 1982), p. 548.
17See “Record” (December 1982), pp. 763-64.

FEDERAL RESERVE BANK OF ST. LOUIS

With respect to the monetary aggregates, the Com­
mittee faced two new concerns: First, a large volume of
all-savers certificates would mature in early October.
Second, later in the quarter, the Depository Institu­
tion Deregulation Committee (DIDC) would imple­
ment the Garn-St Germain Depository Institutions
Act of 1982 and create an account that would be equiva­
lent to and competitive with money market mutual
funds. While the exact nature of this new account and
the timing of its implementation were unknown in
October, it was known that the new account would be
free of interest rate ceilings and would have some
degree of usefulness for transaction purposes.18
It was believed that the maturing all-savers certifi­
cates would induce a temporary increase in M l, while
the new money market deposit accounts (MMDAs)
would depress M l growth upon their introduction.
Because of these conflicting effects, the Committee
believed it would be difficult to interpret movements
in M l during the months ahead.19 It acknowledged
that the new accounts also would affect the growth of
M2; however, it believed that M2 and the broader
aggregates would be affected to a much smaller extent
than M l. Therefore, it decided to set no specific
objectives for M l growth for the fourth quarter, to
increase the weight given to M2 and to set short-run
policy objectives for M3.
At the November meeting, the Committee acknowl­
edged that the bulge in M l growth, which it had
anticipated, had persisted longer than some members
expected, but staff analysis suggested M l growth could
be expected to decelerate over the remainder of the
fourth quarter. It was noted, however, that growth of
both M l and M2 could accelerate in the near term due
to a buildup of balances for eventual placement in the
new MMDAs.20 The Committee concluded that somelsAt the time of this meeting (October 5), the Act had not been
enacted (October 15). The Act required implementation of the
new account no later than 60 days after taking effect.
19There was, however, some reason to believe that the effect of the
new money market deposit accounts (MMDAs) on M l would be
minimal. See John A. Tatom, “Money Market Deposit Accounts,
Super-NOWs and Monetary Policy,” this Review (March 1983),
pp. 5-16.
“ By this time, the Committee knew that MMDAs would become
effective on December 14, 1982. See “Record” (January 1983), p.
19.




JUNE/JULY 1983

what slower growth in M2 for the fourth quarter would
be desirable if such growth were associated with a
decline in market interest rates, and that somewhat
faster growth would be tolerated if exceptional liquid­
ity demands persisted. Once again, the Committee
decided not to set specific policy objectives for Ml.

The growth of M2 during the fourth quarter was very
near the Committee’s short-run objective (see chart 3).
The growth of M l, however, was extremely rapid,
growing at an annual rate of nearly 14 percent. This
rapid fourth-quarter growth of M1 resulted in a fourthquarter-to-fourth-quarter growth rate of 8.5 percent,
well above the upper end of the long-run target range
for the year.

SUMMARY
As in 1981, the FOMC argued that a number of
financial developments and innovations continued to
make it difficult to interpret movements in the two
principal monetary aggregates, M l and M2, during
1982. From the beginning of the year, the Committee
believed that M2 was less likely to be affected by these
factors than M l. This opinion was bolstered by unusual
declines in the income velocity of M l during the first
and fourth quarters of 1982. It was generally felt that
considerable weight should be given to M2 in inter­
preting developments during the year. The Commit­
tee increased the weight given to M2 during the year,
ultimately dropping M l as an explicit intermediate
policy target for the fourth quarter.
Nevertheless, the growth of both M l and M2 fol­
lowed the short-run growth objectives of the Commit­
tee fairly closely during the year. Growth of M1 was
near the Committee’s short-run path until the fourth
quarter, when short-run growth objectives for the
aggregate were dropped. Actual growth of M2 was
near the Committee’s desired short-run path for the
entire year. Rapid fourth-quarter growth of M l, how­
ever, pushed its growth well above the Committee’s
long-run range.

35

Ml or M2: Which Is the Better
Monetary Target?
DALLAS S. BATTEN and DANIEL L. THORNTON

i . HE past few years have been marked by financial
innovation and deregulation: the rapid growth of
money market mutual funds (MMMFs), the nation­
wide introduction of NOW accounts 0anuary 1, 1981),
the introduction of tax-exempt, all-savers certificates
(October 1, 1981) and, most recently, the introduction
of the Garn-St Germain money market deposit ac­
counts (December 14, 1982) and super-NOW accounts
(January 5, 1983).1These changes have led the Federal
Open Market Committee (FOMC) to alter the relative
weight given to M l and M2 in its policy deliberations
during the past two years.
In 1981, the rapid growth of all-savers certificates
prompted the FOMC to lessen the weight assigned to
the M l target relative to the broader monetary
aggregate.2 More recently, the large volume of matur­
ing all-savers certificates and the anticipated introduc­
tion of the new money market deposit accounts
(MMDAs) prompted the FOMC to give much less
weight to M l at its October 1982 meeting.3 Many
believe that these regulatory changes and financial
innovations have increased the substitutability be­
tween M l and non-M l financial assets, thereby
weakening the link between the narrow monetary
aggregate and economic activity.

'F o r a discussion of these developments, see Daniel L. Thornton,
“The FOMC in 1981: Monetary Control in a Changing Financial
Environm ent,” this Review (April 1982), pp. 3 -2 2 ; John A.
Tatom, “Recent Financial Innovations: Have They Distorted the
Meaning of M l?” this Review (April 1982), pp. 23-35; and John A.
Tatom, “ Money Market Deposit Accounts, Super-NOWs and
Monetary Policy,” this Review (March 1983), pp. 5-16.
2See Thornton, “The FOMC in 1981,” p. 15.
3See “Record of Policy Actions of the Federal Open Market Com­
mittee,” F ederal Reserve Bulletin (December 1982), pp. 761-66;
and Daniel L. Thornton, “The FOMC in 1982: De-emphasizing
M l,” this Review 0une/July 1983), pp. 26-35.

36



The purpose of this article is to investigate whether
the relationship between M l and nominal GNP has
deteriorated and to examine the relative performance
of M l and M2 over recent years.4 While considerable
research effort has been devoted to these questions
already, we extend these efforts by (1) using a modified
St. Louis-type equation that has performed well based
on both in-sample and out-of-sample criteria, (2) con­
sidering both in-sample and out-of-sample perfor­
mances of M l and M2, (3) examining the role of the
non-Ml components of M2 separately, and (4) extend­
ing the sample period to include the two most recent
financial innovations.5

MONETARY AGGREGATES AS
INTERMEDIATE POLICY TARGETS
In order for a monetary aggregate to be an appropri­
ate intermediate policy target, there must be a predict­
able relationship between it and income.6 To see this,
note that the chain of causality for monetary policy runs
from the instruments of monetary control to the in-

4W e should note at the outset that we do not see this as a theoretical
debate. The innovations of the past three years could have affected
the income and interest elasticities of various financial assets so as
to alter the usual relationships between these assets (or simple sum
aggregates of these assets, such as M l and M2) and nominal in­
come. Thus, we believe that the issue is essentially empirical.
5For the specification of this modified St. Louis equation, see Dallas
S. Batten and Daniel L. Thornton, “Polynomial Distributed Lags
and the Estimation of the St. Louis Equation,” this Review (April
1983), pp. 13-25.
6It is argued at times that this link must be stable as well as predict­
able. As a general rule, however, the less stable the relationship,
the less predictable it is as well. Moreover, a stable relationship
need not be a numerical constant as is often argued in the context of
the money-GNP relationship.

JUNE/JULY 1983

FEDERAL RESERVE BANK OF ST. LOUIS

The Relationship Between Money and GNP

Figure 1

Chain of Causality for Monetary
Control

The relationship between a monetary aggregate and
economic activity can be summarized by the following
equation:
MV = Y,

where M is a monetary aggregate, V is the income
velocity of money (that is, the rate at which money
changes hands in the purchase of final goods and ser­
vices) and Y is nominal GNP.
'Open Market Operations, changes in reserve requirements and
the discount mechanism — the discount rate and the administra­
tion of the discount window.
^ h e two main goals of policy, full employment and price level
stability, are directly linked to nominal GNP growth.

termediate monetary target to the final goal, nominal
GNP growth, as illustrated in figure 1. It is usually
conceded that M2 is more difficult to control and,
hence, the first link in the chain is stronger for an M1
target.7 Furthermore, there is evidence that the rela­
tionship between the growth of the narrow aggregate
and nominal GNP growth has been more stable
historically.8
Recently, however, some have argued that the rela­
tionship between M l and nominal income has become
weaker than that between M2 and income.9 In the
context of figure 1, those who now claim that M2 is
preferable to M l must be arguing implicitly that the
relationship between M2 and nominal GNP has
strengthened sufficiently to offset any policy problems
that may result from the difficulty of controlling M2.
7Forexample, seeR. W. Hafer, “Much Ado About M 2,” this Review
(October 1981), pp. 13-18; and Patrick J. Lawler, “The Large
Monetary Aggregates as Intermediate Policy Targets,” Voice o f the
Federal Reserve Bank o f Dallas (November 1981), pp. 1-13.
8See Hafer, “Much Ado About M2;” Keith M. Carlson and Scott E.
Hein, “Monetary Aggregates as Monetary Indicators,” this Review
(November 1980), pp. 12-21; and Mack Ott, “Money, Credit and
Velocity,” this Review (May 1982), pp. 21-34.
9See, for example, Edward P. Foldessy, “New Bank Accounts May
Force Fed To End Experiment in Monetarism,” The Wall Street
Journal, December 2 8 ,1 9 8 2 ; “The Money Muddle that Clouds the
Recovery,” Business W eek (May 16, 1983), pp. 120-21; Vincent G.
Salvo, “The Increasing Irrelevance of M l,” International Finance,
The Chase Manhattan Bank (June 6,1983), pp. 4 -5 ; and Aubrey G.
Lanston & Co. Inc., Newsletter (October 4, 12 and 18, 1982).
Similar arguments had been made prior to the fourth quarter of
1982. See, for example, Edward Yardeni, “Unlocking The Secrets
ofThe Federal Reserve,” E . F . Hutton Economics Alert (June 26,
1981); Irwin L. Kellner, “Breakingthe Gridlock,’ The Manufactur­
ers Hanover Econom ic R eport (September 1981); William N.
Griggs and Leonard J. Santow, The S chroder R eport (August 17,
1981); and Irving Kristol, “The Trouble with Money,” The Wall
Street Journal, August 26, 1981.




This relationship is viewed frequently in terms of
growth rates. That is,
M + V = Y,

where the dots over each variable indicate compound­
ed annual growth rates. From this representation, it is
clear that the predictability of the relationship be­
tween a change in money growth and a subsequent
change in GNP growth depends crucially on the pre­
dictability of the rate of growth of velocity.
For the past two decades, M l velocity has been
growing at an average rate of approximately 3 percent
while, on average, M2 velocity has not grown at all.
This is illustrated by chart 1, which contains the fourquarter growth rates of M l and M2 velocities. The
time path of M l velocity growth oscillates around 3
percent, and the path of M2 velocity growth fluctuates
around zero. During the past year and a half, however,
the growth of each of these velocities has declined
dramatically. As a result, the link between these aggre­
gates and GNP appears to have become weaker. Be­
cause the behavior of both velocities have been so
similar, however, casual observation is insufficient to
determine which of these relationships has deterio­
rated more.

AN ECONOMETRIC INVESTIGATION
An econometric analysis of the relationship between
money growth and economic activity involves the use
of a version of the St. Louis equation. The St. Louis
equation was developed to investigate the impact of
monetary and fiscal actions on nominal economic activ­
ity (measured by the growth of nominal GNP). The
equation usually is written as:
(1) Yt = «0 +

J

K

2
Mt_j + 2 7i G,_i + et,
i= 0
i= 0
where Y, Mand Gare the compounded annual growth
rates of GNP, a monetary aggregate and highemployment government expenditures, respectively.
37

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

^ n a ri i

G row th Rates of M l Velocity a n d M 2 Velocity
Percent

Percent

-4

-6

In this article, the appropriate lag lengths (J, K) are
s e le c te d u sin g an o rth o g o n a l re g re ss io n p r o c e d u r e .10

Table 1 contains the results of estimating equation 1
over three sample periods — 11/1962 to III/1982, II/
1962 to IV/1982 and 11/1962 to 1/1983 — using either
M l or M2 as the monetary aggregate. Because the
observed velocity behavior of both M l and M2 have
been unusual during the past two quarters (IV/1982
and 1/1983), this stepwise augmentation of the sample
period was employed to isolate the impact of these
occurrences on the explanatory power of equation l . 11
Several points of comparison are of interest. First,
the M l equation explains 48 percent of the variation in

nominal GNP growth in the II/1962-III/1982 period,
while the M2 equation explains only 26 percent. The
explanatory power of each equation, however, deterio­
rates substantially when the last two quarters of data
are added. In relative terms, the decline in explanatory
power is about the same for each aggregate; conse­
quently, the absolute explanatory power of the Ml
equation remains greater than that of the M2 equation
when the last two quarters are included. Second, a 1
percentage-point change in the growth of either M l or
M2 ultimately leads to a 1 percentage-point change in
GNP growth, regardless of the sample period. Finally,
the cumulative impact of a change in high-employment
government spending is not statistically significant in
either equation for any sample period.

10See Batten and Thornton, “Polynomial Distributed Lags. ” The lag
lengths chosen are lOfor M l and 9 for G in the M l equation, and
11 for M2 and 2 for G in the M2 equation.
''Furtherm ore, an iterative analysis of several subsample periods
was conducted beginning with the subsample period II/1962-IV/
1979 and iterating (adding one quarter at each iteration) until the
full sample period, 11/1962—1/1983, was reached. The only indica­
tion of any deterioration in the explanatory power of either equa­
tion occurred when IV/1982 was added to the sample.




In-Sample and Out-of-Sample Forecasts
To investigate the possible impact of financial in­
novations and regulatory changes in-sample root mean
square errors (RMSEs) are calculated for two sub­

JUNE/JULY 1983

FEDERAL RESERVE BANK OF ST. LOUIS

Table 1

Ordinary Least Squares Estimates of the St. Louis-Type
Equation
Sample Period
11/1962-111/1982

11/1962-IV/1982

11/1962-1/1983

1.150*
(4.52)

1.096*
(3.92)

0.952*
(3.43)

M1 EQUATION
Summed Coefficients

Lags

M

10

G

9

-0 .0 4 2
(0.31)

-0 .0 9 0
(0.61)

-0 .0 4 7
(0.31)

R2

0.48

0.38

0.34

SE

3.16

3.48

3.56

DW

2.12

1.97

1.89

Summary Statistics

M2 EQUATION
Summed Coefficients
M

11

1.310*
(4.64)

1.291*
(4.35)

1.281*
(4.38)

G

2

0.066
(0.76)

0.042
(0.46)

0.041
(0.46)

Summary Statistics
R2

0.26

0.19

0.19

SE

3.77

3.97

3.94

DW

1.91

1.79

1.85

Note: Absolute values of t-statistics in parentheses.
'Statistically significant at the 5 percent level.

periods, before and after 1/1980.12 (These RMSEs are
computed for each of the three estimations of the M1
and M2 equations presented in table 1.) The latter
period was chosen as the period within which the most
important financial innovations and regulatory changes
have occurred. These results are reported in table 2.
They reveal two important facts: First, the in-sample
explanatory performance of Ml during the 1/1980—III/
1982 period actually improved somewhat compared
with the period II/1962-IV/1979, while that of M2
deteriorated. Second, when the last two quarters are
12The in-sample RMSE is defined as:

where ef is the ith residual and rij is the number of observations in
the jth subsample.




added to the second period, the performance of each
aggregate deteriorates. The performance of M l,
although still better than that of M2, does degenerate
relative to that of M2. For example, the RMSE of the
M l equation for the 1/1980—1/1983 period is 66 percent
larger than that for the 1/1980-111/1982 period, while
the same comparison for the M2 equation yields only a
9 percent increase in the RMSE.
A comparison of out-of-sample forecasts of the equa­
tions yielded results similar to those cited above.13 The
13Since the imposition of polynomial restrictions tends to smooth
the distributed lag weights and, thus, tends to improve the accura­
cy of out-of-sample forecasts, these restrictions are imposed in
both of the out-of-sample experiments. The appropriate polyno­
mial degrees are chosen using the methodology presented in
Batten and Thornton, “Polynomial Distributed Lags.” The degrees_ selected are 6 for M l and 3 for G in the M l equation, and 5
for M2 in the M2 equation; no polynomial restrictions are imposed
on G in the M2 equation.

39

JUNE/JULY 1983

FEDERAL RESERVE BANK OF ST. LOUIS

Table 2

Table 3

In-Sample Root Mean Square Errors

Out-of-Sample Root Mean Square
Errors

Equation
Period

Equation

M1

M2

11/1962-IV/1979

2.73

3.09

Forecast Period

M1

1/1980-111/1982

2.50

4.86

1/1980-111/1982

4.57

5.85

11/1962-IV/1979

2.82

3.12

1/1980— IV/1982

7.06

6.22

1/1980-IV/1982

3.77

5.53

1/1980-1/1983

6.93

5.99

11/1962-IV/1979

2.81

3.12

1/1980-1/1983

4.16

5.32

experiment conducted was to estimate each equation
over the period II/1962-IV/1979 and to forecast GNP
growth to the end of the sample period. The out-ofsample RMSEs were calculated for three forecast
periods — 1/1980 to III/1982, 1/1980 to IV/1982 and
1/1980 to 1/1983 — to demonstrate the impact that the
last two quarters have on the forecasting accuracy of
each equation. These results are reported in table 3,
and the individual errors are presented in chart 2. The
evidence indicates that, until the last two quarters, the
M l equation was more accurate in out-of-sample fore­
casting. When the last two quarters are included, how­
ever, the performance of M l deteriorates significantly
while that of M2 remains essentially unchanged. In
fact, the initial relative success of the M l equation
vanishes completely when the last two quarters are
considered.
These results reveal that the link between M l
growth and GNP growth remained strong up to the
fourth quarter of 1982. Thus, the contention that this
relationship had deteriorated prior to the unusual oc­
currence of IV/1982 appears to be without substance.14
Both the in-sample and out-of-sample performances of
the M l equation are considerably better than those of
the M2 equation. Thus, there is no evidence to support
the contention that the relationship between M2 and
income became stronger relative to that of M l and
income before IV/1982. The performance of M l, how­
ever, appears to be more adversely affected by the
developments of the last two quarters. Even though
there is evidence to indicate a recent deterioration in
the Ml-GNP relationship relative to the M2-GNP re14Toida and Gavin also find that M l is preferable to M2 as an
intermediate target. See Mitsuru Toida and William T. Gavin,
“Non-nested Specification Tests and the Intermediate Target For
Monetary Policy, ” Federal Reserve Bank of Cleveland Working
Paper No. 8301 (June 1983).

40



M2

lationship, this period is too short to ascertain whether
this change is temporary or permanent.

Analysis o f the Non-Ml
Components o f M2
By definition, M2 contains M l plus certain other
financial assets.15 Thus, implicit in the argument that
M2 is preferable to M l is the assumption that the
non-Ml components of M2 (NM1) provide additional
explanatory power over that of M l alone. Further­
more, the non-M 1 components of M2 have characteris­
tics which differ, in some cases markedly, from those of
M l. Consequently, the marginal impacts of the Ml
and the non-Ml components of M2 upon economic
activity may vary significantly.16 In order to capture
the possibility of this differential impact, the growth of
the non-Ml components of M2 is included separately
with the growth of M l in equation 1. Estimates from
this augmented equation are given in table 4 for the
three sample periods used previously.17
The inclusion of the non-Ml components has little
effect on the performance of the equation: the standard
errors and adjusted R2s are about the same for compa­
rable sample periods. More importantly, neither the
hypothesis that the cumulative impact of the growth of
the non-Ml components is zero nor the joint hypoth­
esis that all of these coefficients are zero can be re15These other assets are savings (including MMDAs) and small
denomination time deposits at all depository institutions, over­
night repurchase agreements at commercial banks, overnight
Eurodollars held by U. S. residents other than banks at Caribbean
branches of member banks and balances of money market mutual
funds (general purpose and broker/dealer).
16The marginal influences of both sets of components are assumed
implicitly to be the same in the estimation of the M2 equation
because the coefficients of both sets are constrained to be iden­
tical.

17The lag lengths selected for the augmented equation are 10 for
M l, 9 for G and 11 for NM1.

FEDERAL RESERVE BANK OF ST. LOUIS

JUNE/JULY 1983

O u t-o f-S a m p le Forecast Errors of A lte rn a tiv e
St. Louis-Type E qu atio n S pecifications

In -S a m p le R esiduals o f A lte rn a tiv e
St. Louis-Type E qu atio n S pecifications

Actual m in us Predicted V alu es

Actual m inus Predicted V alu e s

1980

1981

1982

1983

1980

1981

1982

1983

Table 4

Ordinary Least Squares Estimates of the Augmented
St. Louis-Type Equation
Sample Period
11/1962-111/1982

11/1962-IV/1982

11/1962-1/1983

Ml

1.050*
(3.79)

1.004*
(3.29)

0.955*
(3.22)

NM1

0.316
(1.31)

0.339
(1.28)

0.356
(136)

-0 .0 4 7
(0.32)

-0 .0 8 2
(0.51)

-0 .0 7 4
(0.46)

R2

0.46

0.35

0.35

SE

3.23

3.56

3.55

DW

2.20

2.03

2.05

SUMMED COEFFICIENTS

G

SUMMARY STATISTICS

Note: Absolute values of t-statistics in parentheses.
'Statistically significant at the 5 percent level.




41

FEDERAL RESERVE BANK OF ST. LOUIS

jected at conventional significance levels during any of
the three periods.18 Thus, the non-Ml components of
M2 provide no additional power over M l alone in
explaining the variation of nominal GNP.
A closely related issue concerns whether the ex­
planatory power of the non-Ml components of M2 has
improved as financial innovation has progressed. Chart
3 contains the in-sample residuals of the M l equation
and the augmented M l equation for the period 1/19801/1983. If the additional explanatory power of the nonM l components has improved during this period, one
would expect to see the residuals of the augmented M l
equation becoming smaller relative to those of the
initial M l equation. The residuals of the augmented
M l equation do appear to be smaller than those of the
M l equation for the last three quarters. In other
words, while these results provide only preliminary
evidence, they do indicate that the performance of the
non-Ml components may have improved during the
past two or three quarters.

SUMMARY AND CONCLUSIONS
Financial innovation in the 1980s has led many to
believe that the relationship between M l growth and
GNP growth has deteriorated relative to that between
M2 growth and GNP growth. Although this is a conlsThe F-statistics calculated to test the hypothesis that all of the
coefficients of NM1 are zero in each of the three periods are 0.77,
0.76 and 1.06, respectively, well below the critical value of 1.95 at
the 5 percent significance level.

42



JUNE/JULY 1983

ceptual possibility, an empirical investigation provides
mixed support for this contention. It is clear that,
within the framework of the version of the St. Louis
equation presented here, M l growth explains more of
the variation of nominal GNP growth than M2 growth
and that there was no marked deterioration in the
Ml-GNP relationship prior to the fourth quarter of
1982.
Drawing conclusions from summary statistics of ex­
planatory power, however, confuses past with present
performance. An analysis of in-sample and out-ofsample forecasting errors reveals that the relative suc­
cess of M l has been due primarily to its past perfor­
mance, not its present one. In particular, the occur­
rences of the past two quarters have had a substantially
larger impact on the relationship between M l and
nominal GNP than that between M2 and GNP.19
While this evidence should promote continued re­
view of the relative merits of M l and M2, it does not
seem sufficient, at present, to conclude that M l should
be de-emphasized as an intermediate target of mone­
tary policy. If subsequent empirical studies provide
more conclusive evidence to support this tentative
finding, then policymakers should consider changes
that will enhance their ability to control M2.
19It should be noted that even though recent financial innovations
and deregulation have motivated this study, the findings do not
necessarily indicate that these innovations and regulatory changes
have been the cause of the results obtained. In fact, much of the
innovation and deregulation that has occurred predated the time
period during which the changes in explanatory power have been
identified.