View original document

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

____________ Review____________
Vol. 66, No. 4

April 1984

5 International Banking Facilities
12 A Private Central Bank: Some Olde
English Lessons
23 Money Growth Variability and GNP

The Review is published 10 times per year by the Research and Public Information Department o f
the Federal Reserve Rank o f St. Louis. Single-copy subscriptions are available to the public free o f
charge. Mail requests fo r subscriptions, back issues, or address changes to: Research and Public
Information Department, Federal Reserve Rank o f St. Louis, P.O. Rox 442, St. Louis, Missouri
The views expressed are those o f the individual authors and do not necessarily reflect official
positions o f the Federal Reserve Rank o f St. Louis or the Federal Reserve System. Articles herein may
be reprinted provided the source is credited. Please provide the Rank’s Research and Public Informa­
tion Department with a copy o f reprinted material.

Federal Reserve Bank of St. Louis
April 1984

In This Issue . . .

In the first article in this Review, “International Banking Facilities,” K. Alec Chrystal
gives a brief outline o f the regulatory changes that permitted banking institutions
in the United States to establish international banking facilities (IBFs) and conduct
international banking business on terms similar to offshore financial centers.
Chrystal describes the location and growth o f IBFs and details the extent to
which business has been shifted from other accounts and other locations. In
addition, he explains the nature o f IBF business as an integral part o f the eurodol­
lar market. Chrystal concludes that, apart from a change in geographical location,
the significance o f IBFs in international banking is small.
In the second article, “A Private Central Bank: Some Olde English Lessons," G.
J. Santoni investigates the proposition that central bankers respond systematically
and predictably when faced with different incentives relating to monetary control.
The author focuses on the actions o f the Bank o f England from 1694 to 1930.
The Bank o f England provides both an interesting and an apt case study for two
reasons. First, in contrast to modern central banks, England's central bank was a
privately-owned, for-profit institution from its founding in 1694 until the early
1930s. Further, the bank was structured such that the wealth of the Bank’s owners
was inversely related to the rate o f inflation. Second, the government seized the
monetary control function from the private owners o f the bank from 1793 to 1821.
This period is important because it demonstrates clearly h ow different incentives
produce different policy consequences, holding other important institutional
factors roughly constant.
Santoni concludes that the incentives confronting England’s p rivate central
bankers led them to choose relatively low rates o f monetary growth and inflation;
in contrast, the government chose significantly faster rates o f monetary growth
and inflation when it controlled the Bank o f England.
In the third article, "Money Growth Variability and GNP,” Michael T. Belongia
explains the sequence o f events through which short-run fluctuations in M l
growth make future econom ic conditions more uncertain and, thus, cause reduc­
tions in the growth o f GNP.
Belongia investigates the impact of variable m oney growth on GNP by estimat­
ing a variant o f the St. Louis equation that incorporates directly a measure o f the
variability o f money growth. He finds that the effects o f variable M l growth were
negative, but transitory in nature, prior to 1980. Since 1980, however, the variability
o f M l growth has tripled, and its negative effects on the path o f GNP have become
permanent. In a simulation experiment, the author shows that the growth of
nominal GNP w ould have been higher and more stable between 1979 and 1982 if
M l growth had been less variable over that period.



International Banking Facilities
K. Alec Chrystal

I n TERNATIONAL Banking Facilities (IBFsl started
operation in the United States in early December 1981.
Since then, they have grown to the point where they
now represent a significant part o f the international
banking business worldwide. The purpose o f this arti­
cle is to examine IBFs and to discuss their significance
for international banking.

A substantial “offshore” international banking sec­
tor, often called the “eurocurrency” market, grew up in
the 1960s and 1970s. Its key characteristic is that bank­
ing business is transacted in a location outside the
country in whose currency the business is denom i­
nated. Thus, eurodollar transactions are conducted
outside the United States, eurosterling transactions are
conducted outside Britain, and so on. Much o f this
offshore business occurs in major financial centers like
London, though some business is literally in islands
offshore from the United States, such as the Bahamas
or Cayman Islands.
Offshore banking business is somewhat different
from that conducted onshore. Though, in both cases,
banks take deposits and make loans, offshore banks
have virtually no checking deposit liabilities. Instead,
their deposits are typically made for specific periods of
time, yield interest, and are generally in large denom ­
Offshore banking arose as a means to avoid a variety
of banking regulations. For example, offshore banks
that deal in eurodollars avoid reserve requirements on

K. Alec Chrystal, professor of economics-elect, University of Shef­
field, England, is a visiting scholar at the Federal Reserve Bank of St.
Louis. Leslie Bailis Koppel provided research assistance.

deposits, FDIC assessments and U.S.-imposed interest
rate ceilings. The first two of these regulations increase
the margin between deposit and loan rates. Avoiding
these costs enables offshore banks to operate on much
smaller margins. Interest ceilings, where binding, re­
duce the ability o f banks subject to such ceilings to
compete internationally for deposits.
Many "shell” bank branches in offshore centers,
such as the Caymans and Bahamas, exist almost solely
to avoid U.S. banking regulations. Shell branches are
offices that have little more than a name plate and a
telephone. They are used simply as addresses for book­
ing transactions set up by U.S. banks, which thereby
avoid domestic m onetaiy regulations.

IBFs do not represent new p h y sica l banking facilities;
instead, they are separate sets o f books within existing
banking institutions — a U.S.-chartered depository in­
stitution, a U.S. branch or agency o f a foreign bank, or a
U.S. office o f an Edge Act corporation.1 They can only
take deposits from and make loans to nonresidents of
the United States, other IBFs and their establishing
entities. Moreover, IBFs are not subject to the regula­
tions that apply to domestic banking activity; they
avoid reseive requirements, interest rate ceilings and
deposit insurance assessment. In effect, they are
accorded the advantages o f many offshore banking
centers without the need to be physically offshore.

'As a result of a 1919 amendment to the Federal Reserve Act initiated
by Sen. Walter Edge, U.S. banks are able to establish branches
outside their home state. These branches must be involved only in
business abroad or the finance of foreign trade. The 1978 Interna­
tional Banking Act allowed foreign banks to open Edge Act corpora­
tions which accept deposits and make loans directly related to inter­
national transactions.



The Establishment o f IBFs
Three regulatory or legislative changes have permit­
ted or encouraged the establishment and growth of
IBFs. First, the Federal Reserve Board changed its reg­
ulations in 1981 to permit the establishment o f IBFs.
Second, federal legislation enacted in late 1981 ex­
empted IBFs from the insurance coverage and assess­
ments imposed by the FDIC. Third, several states have
granted special tax status to the operating profits from
IBFs or altered other restrictions to encourage their
establishment. In at least one case, Florida, IBFs are
entirely exempt from local taxes.

Restrictions on IBF Activities
While IBFs may transact banking business with U.S.
nonresidents on more or less the same terms as banks
located offshore, they may not deal with U.S. residents
at all, apart from their parent institution or other IBFs.
Funds borrowed by a parent from its own IBF are
subject to eurocurrency reserve requirements just as
funds borrowed from an offshore branch would be.
Four other restrictions on IBFs are designed to en­
sure their separation from domestic money markets.
First, the initial maturity o f deposits taken from non­
bank foreign customers must be at least two working
days. Overnight deposits, however, may be offered to
overseas banks, other IBFs and the parent bank. This
restriction ensures that IBFs do not create a close
substitute for checking accounts.
Second, the minimum transaction with an IBF by a
nonbank customer is $100,000, except to withdraw in­
terest or close an account. This effectively limits the
activity o f IBFs to the “wholesale” money market, in
which the customers are likely to be governments,
m ajor corporations or other international banks.2
There is no restriction on the size of interbank transac­
Third, IBFs are not permitted to issue negotiable
instruments, such as certificates o f deposit (CDs), be­
cause such instruments w ould be easily marketable in
U.S. money markets, thereby breaking down the in­
tended separation between IBFs and the domestic
money market.
Finally, deposits and loans o f IBFs must not be re­
lated to a nonresident custom er’s activities in the

2Foreign governments are treated like overseas banks for purposes of
maturity and transaction size regulations.

Digitized for

APRIL 1984

United States.'* This regulation prevents IBFs from
competing directly with domestic credit sources for
finance related to domestic econom ic activity.

Where Are IBFs Located?
IBFs are chiefly located in the major financial centers
(see table 1). Almost half o f the nearly 500 IBFs are in
N ew York; California, Florida and Illinois have the bulk
o f the rest. In terms o f value o f liabilities, however, the
distribution is even more skewed. O f IBFs reporting
monthly to the Federal Reserve (those with assets or
liabilities in excess o f $300 million), 77 percent o f total
liabilities were in New York, with California (12 percent)
and Illinois (7.5 percent) a long wav behind. It is nota­
ble that Florida, which has 16.5 percent of the IBFs,
has only 2 percent o f the liabilities o f reporting banks.
While the distribution o f IBFs primarily reflects the
preexisting locations o f international banking busi­
ness, differences in tax treatment between states may
have influenced the location o f IBFs marginally. For
example, the fact that Florida exempts IBFs from state
taxes may well explain w hy it has the largest number of
Edge Act corporation IBFs and ranks second to New
York in terms o f numbers o f IBFs set up bv U.S.chartered banks.
Although Florida has the most advantageous tax
laws possible for IBFs, it is not alone in granting them
favorable tax status. Nine other states (N ew York, Cali­
fornia, Illinois, Connecticut, Delaware, Maryland,
Georgia, North Carolina and Washington) and the Dis­
trict o f Columbia have enacted special tax laws that
encourage the establishment o f IBFs.4
The reason for the favorable tax treatment for IBFs in
states like Florida is not clear. There is no doubt that
Florida has tried to encourage its development as an
international financial center.5 The benefits from en­
couragement o f IBFs per se, however, are hard to see.
For example, the employment gains are probably triv­
ial. Since IBFs are merely new accounts in existing
institutions, each IBF will involve at m o st the em ploy­
ment o f a handful of people. In many cases, there may
be no extra employment.

3“The Board expects that, with respect to nonbank customers located
outside the United States, IBFs will accept only deposits that support
the customer's operations outside the United States and will extend
credit only to finance the customer's non-U.S. operations.” See
“Announcements” (1981), p. 562.
4These provisions vary from case to case. For a summary of the
position in New York and California, see Key (1982).
5See "Florida's Baffling Unitary Tax" (1983).

APRIL 1984


Table 1
Location of International Banking Facilities
Liabilities of Monthly Reporting IBFs,
Other than to Parent Entity
of total

of banks



Agencies and
branches of
foreign banks

Edge Act

of dollars)







New York
District of
New Jersey
N. Carolina
Rhode Island

















(There are too few reporting banks in other
states for a data breakdown to be made available.)

NOTE: Figures for numbers of IBFs are as of September 28,1983. Figures for liabilities are as of October 26,1983. Monthly reporting banks
are those with assets or liabilities of at least $300 million. SOURCE: Federal Reserve Board Release G.14(518)A and Federal
Reserve Board unpublished data.
'One savings and loan association has an IBF that is in the Florida figure.

What D o IBFs Do?
The assets and liabilities of IBFs on December 30,
1981, December 29, 1982, and October 20, 1983, are
recorded in table 2; as o f October 20, 1983, over 98
percent o f their liabilities w ere dollar-denominated.
The December 30, 1981, figures largely reflect busi­
ness switched from other accounts either in the parent
bank or an offshore branch. Operations o f the IBFs
themselves are reflected more clearly in the later
figures. Consider the latest available figures in the third
column o f table 2. The most important aspects o f these
figures is the proportion o f business with other banks
vs. the proportion with nonbank customers. On the
asset side, about one-sixth o f total assets are "com m er­
cial and industrial loans” (Item 5a) and one-ninth are
loans to “foreign governments and official institutions”
(Item 5c). The remainder, over 70 percent, are claims on

either other IBFs, overseas banks or an overseas branch
of the parent bank. Claims on overseas banks (Items 3a
and 5b) are largest, while claims on other IBFs (Item 2)
and overseas offices o f the parent bank (Item 1) are of
broadly similar magnitude.
The liability structure is even more heavily weighted
toward banks. Only about 16 percent o f the liabilities of
IBFs (as o f October 26, 1983) w ere due to nonbanks. Of
these, one-third was due to "foreign government and
official institutions” (Item 10c) and two-thirds were
due to “other non-U.S. addressees” (Item lOd). The
latter are mainly industrial and commercial firms.
The high proportion o f both assets and liabilities of
IBFs due to other banking institutions reinforces the
conclusion that they are an integral part o f the euro­
dollar market. A high proportion of interbank business
is characteristic o f eurocurrency business in which


Table 2
Assets and Liabilities of International Banking Facilities (millions of dollars)
December 30, 1981

December 29, 1982

October 26, 1983

















































1. Gross Claims on Non-U.S. Offices of Establishing Entity
(1) Denominated in U.S. Dollars
(2) Denominated in Other Currencies
2. Loans and Balances Due From Other IBFs
3. Gross Due From:
A. Banks in Foreign Countries
B. Foreign Governments and Official Institutions
4. Securities of Non-U.S. Addressees
5. Loans To Non-U.S. Addressees
A. Commercial and Industrial Loans
B. Banks in Foreign Countries
C. Foreign Governments and Official Institutions
D. Other Loans
6. All Other Assets in IBF Accounts
7. Total Assets Other Than Claims on U.S. and Non-U.S.
Office of Establishing Entity
(1) Denominated in U.S. Dollars (Sum of Items 2
through 6)
(2) Denominated in Other Currencies
8. Total Assets Other Than Claims on U.S. Offices of
Establishing Entity (Sum of Items 1 and 7)
(1) Denominated in U.S. Dollars
(2) Denominated in Other Currencies
9. Gross Liabilities Due To Non-U.S. Offices of Establishing
(1) Denominated in U.S. Dollars
(2) Denominated in Other Currencies
10. Liabilities Due To:
A. Other IBFs
B. Banks In Foreign Countries
C. Foreign Government and Official Institutions
D. Other Non-U.S. Addressees
E. All Other Liabilities in IBF Accounts
F. Total Liabilities Other Than Due To U.S. and Non-U.S.
Offices of Establishing Entity
(1) Denominated in U.S. Dollars
(Sum of Items 10.A Through 10.E)
(2) Denominated in Other Currencies
11. Total Liabilities Other Than Due to U.S. Offices of
Establishing Entity (Sum of Items 9 and 10.F)
(1) Denominated in U.S. Dollars
(2) Denominated in Other Currencies
12. Net Due From ( + ) / Net Due To ( -) U.S. Offices of
Establishing Entity (Item 11 Minus Item 8)
(1) Denominated in U.S. Dollars
(2) Denominated in Other Currencies
Number of Reporters

NOTE: Unless otherwise noted, figures include only amounts denominated in U.S. dollars. This report contains data only for those entities
whose IBF assets or liabilities are at least $300 million, that is, for those entities that file a monthly report of IBF accounts on form FR
2072. SOURCE: Federal Reserve Board Release G.14 (518).


there may be several interbank transactions between
ultimate borrowers and ultimate lenders.6
An important role for interbank transactions is to
provide "swaps” that reduce either exchange risk or
interest rate risk for the parties involved. Suppose, for
example, an IBF has a deposit (liability) o f $1 million
that will be withdrawn in one month, and it has made
a loan (asset) to a customer o f $1 million that will be
repaid in two months. There is a risk that when the IBF
comes to borrow $1 million to cover the second month
o f the loan, interest rates w ill have risen, and it will
incur a loss on the entire transaction. If, however, this
IBF can find a bank that has the opposite timing prob­
lem (a deposit o f $1 million for 2 months and a loan of
$1 million outstanding for one month), the two banks
could arrange a swap. The second bank would loan the
IBF $1 million in one month and get it back in two
months (with suitable interest). The interest rate in­
volved will be agreed on at the beginning, so that nei­
ther bank w ould suffer if interest rates should change
in the second month.
These swap arrangements enable banks to match
the maturity structure o f their assets and liabilities.
The existence o f such swaps explains the high levels of
both borrowing and lending between IBFs and over­
seas branches o f their parent bank.7

Chart 1 shows the growth o f total IBF liabilities since
the end o f 1981. Although the most rapid growth oc­
curred in the first six months o f their operation, IBFs
have grown considerably over a period in which inter­
n ation al banking business in gen eral has been
stagnant.8 Within two years, they have come to be a
significant part o f the international money market. The
liabilities o f IBFs as o f October 1983 (other than to
parent banks) represent about 8Vz percent o f gross
eurocu rrency liabilities (as m easured bv Morgan
Guaranty) or about 7Vi percent o f total international
banking liabilities (as measured by the Bank for Inter­
national Settlements. This includes onshore bank

6See Niehans and Hewson (1976) for an explanation of the intermedi­
ary function of euromarkets. The interbank market is also discussed
in Dufey and Giddy (1978), chapter 5.
7For a discussion of the role of swaps in foreign exchange markets,
see Chrystal (1984).
8According to B.I.S. figures, international bank assets grew 8.8 per­
cent in 1982 in nominal terms. This compares with figures typically in
excess of 20 percent throughout the 1970s. The combined assets of
overseas branches of U.S. banks declined by 0.6 percent in 1982
[see Press Release (1983)], though this partly reflects the growth of

APRIL 1984

C h a rt 1

T o ta l L ia b ilitie s o f IBFs
Log of levels

Log of levels

NOTE: Liability levels w ere $44,777 m illion on December 30, 1981, an d $173,430 m illion
on O cto b er 26, 1983. Figures exclude liab ilitie s to p arent entity.

Where did this growth com e from? Has the creation
oflBFsgenerateda large volume o f new business or has
business been shifted from elsewhere? The evidence is
that IBF business has almost entirely been shifted from
elsewhere. Terrell and Mills use regression analysis to
test the hypothesis that the creation o f IBFs has led to
greater growth o f external bank assets.9 This hypoth­
esis is decisively rejected.
Some evidence concerning the origins o f business
shifted to IBFs is available in Key.10 It is convenient to
consider separately shifts from existing institutions in
the U.S. and shifts from overseas banking centers.

Shifts fro m Banks in the United States
Up until January 27,1982, about $34 billion o f claims
on overseas residents w ere shifted to IBF books from
other U.S. banking institutions. The bulk o f this (85
percent) came from U.S. branches o f foreign banks —
especially Japanese and Italian. Foreign banks typical­
ly w ould have had a higher proportion o f assets eligible
for shifting to IBFs, w hile Japanese and Italian banks
generally had not established shell branches in Carib­
bean offshore centers.
9See Terrell and Mills (1983).
10See Key (1982).



in the same period, shifts of liabilities (due to parties
other than overseas branches o f the parent bank) from
books o f parent entities w ere much smaller. These
amounted to about $6 billion, o f which 90 percent
came from branches o f foreign banks . The small shift of
liabilities relative to assets was affected by several fac­
tors: the negotiable nature o f some deposits (CDs); the
existence o f penalties for renegotiations before matu­
rity; the delay in passing N ew York tax relief for IBFs
until March 1981; the small proportions o f short-term
deposits unrelated to trade with the United States; and
the availability o f accounts with similar returns yet
fewer restrictions as to maturity and denomination
(such as repurchase agreements).
If only the domestic books o f U.S.-chartered banks
are considered, the shift to IBFs is extremely small. Key
reports a shift o f $4.3 billion (through January 27,1982)
o f claims on unrelated foreigners and only $0.1 billion
o f liabilities to unrelated foreigners. An alternative
figure for claims shifted to IBFs is obtainable by looking
at the change in commercial and industrial loans to
non-U.S. addressees plus loans to foreign banks (F e d e r­
al Reserve Bulletin, table A18, for large weekly reporting
banks with assets o f $750 million or more). This indi­
cates a decline o f $3.3 billion in the same period.

APRIL 1984

shift reflected the redundancy o f shell branches, at
least for business with non-U.S. residents, once IBFs
were permitted.
W hile m uch o f the r a is o n d ’e tr e o f Caribbean
branches for business with foreigners has been re­
moved by the establishment o f IBFs, these branches
continue to be important for business with U.S. resi­
dents. Terrell and Mills report that the proportion o f
the liabilities o f Caribbean branches due to U.S. resi­
dents rose from less than half in mid-1981 to about 70
percent by the end o f 1982. However, the attraction o f
offshore deposits to U.S. residents is likely to decrease
as interest regulations on domestic U.S. banks are re­
laxed, thereby narrowing the gap between domestic
and offshore deposit rates.
Based on the figures o f the Bank for International
Settlements, Terrell and Mills estimate that the propor­
tion o f total international banking assets and liabilities
due to U.S. banks’ offshore branches declined by 4
percent in the first year o f IBF operation. Another 3'/2
percent was lost by other overseas banking centers to

Shifts from Other Offshore Centers


Whereas foreign banks were mainly responsible for
shifts to IBFs from banks located in the United States,
banks chartered in the United States w ere mainly re­
sponsible for shifts o f business from offshore centers
and other overseas banking locations. Key estimates
that U.S.-chartered banks shifted about $25 billion in
claims on unrelated foreigners and about $6 billion in
liabilities due to unrelated foreigners (through Januaiy
27,1982) to IBFs from overseas branches. The compara­
ble figures for foreign banks w ere $5Vz billion and $9
billion, respectively.

The primary significance o f the experience with IBFs
is that it enables us to better understand the forces that
led to the growth o f eurocurrency markets. In partic­
ular, the significant decline in business in Caribbean
branches following the creation o f IBFs suggests that
the growth of business in this area was almost entirely
intended to bypass U.S. monetary regulations. Dereg­
ulation o f domestic banking in the United States will
presumably have further effects, since much o f the
remaining business in Caribbean branches o f U.S.
banks is with U.S. residents.

This difference in the propensity to shift assets to
IBFs is probably explained by the differential tax incen­
tives o f U.S. and foreign banks. U.S. banks pay taxes on
w orldw ide income and may benefit from tax advan­
tages o f IBFs. Foreign banks may increase their tax
liability to the United States by establishing an IBF
instead o f operating in an offshore center.

The regulatory changes that permitted the establish­
ment o f IBFs w ere intended to ease the burden o f
domestic monetary restrictions on U.S. banks in the
conduct o f international banking business.11 The ex­
tent to which this aim has been achieved is probably
very limited. This is because IBFs play no role in financ­
ing either activities o f U.S. residents or the U.S. activities
o f nonresidents.

The bulk o f business shifted by U.S. banks from their
overseas branches has com e from the Bahamas and
Cayman Islands (collectively called Caribbean). In the
first two months o f operation o f IBFs (11/30/81-1/29/821,
liabilities to unrelated foreigners o f branches o f U.S.
banks located there fell by $6.8 billion, w hile claims on
unrelated foreigners fell by $23.3 billion. Much o f this

Digitized for

Major U.S. banks that w ere involved in international
finance to a significant degree had already found ways
around U.S. banking regulations and w ere not re­
stricted in their ability to com pete internationally. The

11Ibid., p. 566.

APRIL 1984


fact that major U.S. banks have shifted business to IBFs
from offshore centers means, o f course, that there must
be some benefit from having an IBF. This may result
from low er transaction costs, some tax advantages or
the greater attraction, from a risk perspective, o f de­
posits located in the United States. However, the big­
gest gainers among U.S. banks may be medium-sized
banks that were big enough to have some international
business but not big enough to have an offshore
Other major beneficiaries from IBFs have been the
U.S. branches and agencies o f foreign banks. It is no
accident that w ell over half o f all IBFs have been estab­
lished by these banks. The benefit to them arises from
the high proportion o f their existing business that is
IBF-eligible, that is, the portion with nonresidents. Not
the least o f this w ould be transactions with their parent
banks overseas.

creased the total volume o f international banking busi­
ness. Indeed, IBFs have grown at a time when interna­
tional banking growth has been at its slowest for over
two decades. This growth has been largely at the ex­
pense o f banking offices in other locations.
For the U.S. and w orld economies, however, IBFs are
not o f great significance. There may be efficiency gains
resulting from the relaxation o f U.S. regulations that led
to the establishment o f IBFs. But such gains are small.
Interest rates in w orld capital markets are unlikely to
have been affected. Benefits that accrue to banks lo ­
cated in the United States from their IBF facilities are
largely offset by losses in offshore banks, though in
many cases the gainers and losers are both branches of
the same parent bank.

“Announcements." Federal Reserve Bulletin (July 1981), pp. 561-63.

The establishment o f IBFs in the United States repre­
sents a change in the geographical pattern o f interna­
tional banking. It facilitates the conduct in the United
States o f some business that was previously conducted
offshore. It also increases the ease with which foreign
banks can operate branches in the United States. The
creation o f IBFs, however, does not seem to have in­

12lt is true that the largest banks have the largest IBFs. However, the
cost saving at the margin from IBFs for a bank that had, say, a
Caribbean shell operation is much smaller than for a bank that had
no offshore booking location.

Chrystal, K. Alec. “A Guide to Foreign Exchange Markets,” this
Review (March 1984), pp. 5-18.
Dufey, G., and Ian H. Giddy.
(Prentice-Hall, 1978).

The International Money Market

“Florida's Baffling Unitary Tax: What Is It, Whom Does It Hurt?”
American Banker, December 28, 1983.
Key, Sydney J. “International Banking Facilities,” Federal Reserve
Bulletin (October 1982), pp. 565-77.
Niehans, Jurg, and John Hewson. “The Euro-dollar Market and
Monetary Theory,” Journal of Money, Credit and Banking (Febru­
ary 1976), pp. 1-27.
Press Release.

Federal Reserve Board of Governors, August 22,

Terrell, Henry S., and Rodney H. Mills. International Banking Facili­
ties and the Eurodollar Market, Staff Studies No. 126 (Board of
Governors of the Federal Reserve System, August 1983).


A Private Central Bank:
Some Olde English Lessons
G. J. Santoni


ISSATISFACTION with persistent and volatile in­
flation since the mid-1960s has led to numerous calls
for a different approach to monetary policy. In some
cases, people have suggested that monetary policy de­
cisions be made more explicitly political, for exam­
ple, subject to greater control by Congress via congressionally mandated monetary growth targets. In the
same vein, Milton Friedman has proposed that m one­
tary policy be set by the Treasury, thus making the
President o f the United States ultimately responsible
for its conduct.
In contrast, some critics o f the current system have
argued for a return to the constraints o f Bretton Woods
or the even earlier classical gold standard. Some have
suggested that the only lasting solution to the problem
entails the private production o f money.
Behind these different suggestions is the implication
that central bankers will respond systematically and,
hence, predictably to the different incentives em ­
bodied in these alternative programs of monetary con­
trol. If the incentives are changed, so the theory goes,
better policy decisions will be made.
This paper focuses directly on the theoretical and
empirical support for the claim that different incen­
tives induce policymakers to choose different m one­
tary growth rates. This paper does not advocate a par­
ticular set o f incentives or form o f organization for the
central bank. Rather, it merely points out that the
choices o f m onetary policym akers depend, as all
choices do, upon the set o f incentives the individual

G. J. Santoni is a senior economist at the Federal Reserve Bank of
St. Louis. Thomas A. Pollmann provided research assistance.


A verification o f this proposition is sought by ex­
amining the behavior o f the Bank o f England over the
period from, roughly, 1700 to 1930. This period, which
encompasses two significant changes in the incentives
facing England’s central bankers, provides support for
the view that policymakers, like other individuals, re­
spond predictably to changes in the cost-reward cir­
cumstances facing them.

The central banker is the person (or group) holding
the enforceable right to control the quantity o f nominal
money balances in circulation.1 This right is valuable.
Whoever holds it can, among other things, materially
influence the rate o f inflation and the flow o f profits
from money creation (seigniorage), as well as the pres­
ent value o f the right itself.

The Bank o f England presents an interesting case in
studying the effect o f different incentives on the be­
havior o f central bankers. There are two reasons for
this. First, the original organization of the Bank differed
from its modern counterparts in one fundamental re­
spect: the Bank o f England was a privately-owned-forprofit central bank from its inception in 1694 until the
'More precisely, they control the issuance of “high-powered money”
or base money. Since the long-run link between base money and the
transaction balances of the public (money) is so close, the paper
treats the right to control base money as synonymous with the right to
control money. See Balbach (1981); Johannes and Rasche (1979).


early 1930s. Further, the Bank was immersed in a set of
institutional arrangements that related the wealth of
the Bank’s owners inversely to the rate o f inflation. This
paper shows that the costs and benefits o f varying
monetary growth rates were different for England's
private for-profit central bankers than those typically
taken into account by modern central bankersr Conse­
quently, a different monetary growth rate emerged.
A second reason for studying this particular case is
that control o f the money supply by the Bank's owners
was interrupted from 1793 to 1821, when the govern­
ment seized the Bank’s monetary control function. For
the purposes o f this paper, the interruption is im por­
tant because it allows a contrast o f monetary growth
outcomes produced bv certain identifiable changes in
the incentive structure, while other important institu­
tional factors remained roughly constant.

ENGLAND: 1694-1832
Prior to 1694, England's m oney supply consisted
mainly o f coins. These coins w ere controlled by the
government through regulation o f the mint.
The coins were in a continually bad state because
the populace persisted in clipping, sweating, filing,
washing and boring them. Further, the government
resorted to progressive debasement in the form o f fre­
quent recoinages, a practice that was particularly pro­
nounced during the reigns o f Henry VIII and Edward
In some cases, the government expropriated m one­
tary wealth outright. In 1640, Charles I closed the Lon­
don Mint and confiscated the funds o f private citizens
that had been stored there for safekeeping. Later, in
1672, Charles II expropriated funds deposited with the
Treasury by London goldsmiths.4 This irresponsible
behavior had important consequences when, in 1692
and 1693, William III floated long-term loans to finance
a war with France. Because o f the earlier debasements

2Alchian (1977), pp. 127-50.
3Kemmerer (1944), pp. 34-36.
“Charles finally acknowledged one-half the debt, but the promise to
pay was never kept. Payments, at the rate of 6 percent, were made
only during the period 1677-83. Charles explained the reason for his
action as follows:
Whereas since the time of our happy Restoration We have been involved
in great Forreigne Warrs as well for the Safety of our Government as the
vindication of the Rights and Privileges of our Subjects, In the prosecu­
tion whereof we have been constreyned for some years past, contrary to
our Inclinations, to postpone the payment of the moneys due from Us to

APRIL 1984

and expropriations, the interest rates dem anded for
long-term loans to the Crown contained a substantial
premium :’

Establishment o f the Bank
W illiam ’s war with France was a costly affair. When
additional funds w ere required in 1694, a proposal that
twice previously had been put forward by William
Paterson was adopted in the Ways and Means Act of
that year. The Act provided that those w ho subscribed
for and towards the raising and paying into the receipt o f
the Exchequer the said sum o f twelve hundred thousand
pounds part o f the sum o f fifteen hundred thousand
pounds w ere to constitute jointly the Com pany o f the
Bank o f England.*’

The loan was a perpetuity, paying interest at the rate of
8 percent. This was considerably below the interest
rates that previously had been charged the Crown. The
subscribers, however, received additional rights to 1)
form a joint stock banking company, 2) deal in bills of
exchange, gold and silver, 3) grant advances on secu­
rity, and 4) issue promissory notes transferable by en­
dorsement in an amount not exceeding the Bank’s
These terms apparently were very attractive. The
entire loan was subscribed within 12 days. Every sub­

several Goldsmiths and other upon Tallys s tru ck. . . , And although the
present Posture of Our affaires cannot reasonably spare so greate a sum
as must be applied to the satisfaction of those debts, Yet considering the
great difficulty which very many of our Loving Subjects (who putt their
moneys into the hands of those Goldsmiths and others from whom we
received it) doe at present Lye under, almost to their utter ruine for want
of their said moneys. We have rather chose out of our princely care and
compassion towards Our people, to suffer in Our owne Affaires than that
our loving subjects should want soe reasonable a Reliefe.

Bisschop (1967), p. 48. Incidentally, tallys were pieces of wood
upon which government indebtedness was recorded. The govern­
ment issued tallys when it borrowed from individuals. Goldsmiths
regularly accepted these tallys and credited the accounts of the
sHomer (1977), p. 126. Both loans were of 1 million pounds. The first
was a life annuity paying interest of 10 percent until 1700 and 7
percent thereafter on a semitontine basis (surviving subscribers split
one-half the proceeds due decedent subscribers). The second loan
paid interest at the rate of 14 percent.
6Bisschop, p. 74, and Clapham (1958), vol. 1, pp. 16-20.
7Bisschop, pp. 70-71, and Macleod (1897), pp. 773 and 776. The
Bank employed three methods of accounting for the transaction
accounts (“running cash") of its depositors and these methods de­
fine how the balances were transferred in the exchange process. The
methods were by “Notes payable to Bearer, to be endorsed,” by
“Books or Sheets of Paper, wherein their Account to be entered,” or
by “Notes to persons to be accomptable." The first method was the
forerunner of central bank notes. The third was essentially equivalent
to a present-day checking account. The second was much like
modern passbook accounts. See Clapham, vol. 1, p. 21.



scriber became a shareholder of the Bank to the extent
o f his subscription, and all or any fraction o f his share
could be sold to others.8
The Bank opened for business on July 27, 1694, in
Mercers’ Chappell. From its inception, eveiy effort was
made by the Governor and Court o f Directors (Board of
Directors o f the Bank) to attract depositors and to
promote the circulation o f its “running cash notes.’’9
These notes were convertible into legal tender money,
gold coins, at a fixed exchange rate upon demand at
the Bank."1
Subsequent legislation strengthened the Bank's
position. The Bank was granted a m onopoly in joint
stock banking in early 1697.” In 1708, the Bank
obtained a m onopoly in the issue o f joint stock bank
notes.1' Later, in 1742, Act 10 and 11 George II., C. 13
(par. 5), reaffirmed the earlier rights granted to the
Bank. Each o f these pieces of legislation was accompa­
nied by an additional Bank loan to the government.
The Bank subscribed an additional 1,001,017 pounds
for loan to the government at 8.0 percent in 1697. In
1707, it extended a 1,500,000 pound loan at 4.5 percent
to the government and, in 1742, another 1,600,000
pounds at 3.0 percent.
No further significant legislative changes regarding
the Bank's position occurred until 1826. In that year,
the Bank's m onopoly on joint stock banking was lim­
ited to within a 65-mile radius o f London.1:1Seven years
later, in 1833, its m onopoly o f joint stock bank note

8 Bank shares exchanged hands regularly and, in 1747, Gentleman’s
Magazine began publishing daily price quotes for Bank of England
shares of stock. In 1773 New Jonathan's Coffee House printed the
words, “The Stock Exchange" over its door and admittance was
permitted only by fee.
9Clapham, vol. 1, pp. 20-23.
,0lt is interesting to note that an official (established by law) gold
standard was not enforced in England at this time. The Bank's
commitment to redeem its notes at a fixed price in terms of gold was
not foisted upon the Bank by the government. Rather, this was a
voluntary contract established by the Bank with its customers. En­
gland’s official gold standard was not established until 1821 (more
than 100 years later) and then only as a result of the Bank’s
11Clapham, vol. 1, pp. 46-50.
12Bisschop, pp. 82-83. Act 7 Anne, C.7, provides “that during the
continuance of the said corporation of the Governor and Company
of the Bank of England, it shall not be lawful for any body politic or
corporate whatsoever, created or to be created (other than said
Governor and Company of the Bank of England), or for any other
persons whatsoever, united or to be united in covenants, or part­
nership, exceeding the number of six persons, in that part of Great
Britain called England, to borrow, owe, or take up any sum or sums
of money on their bills or notes payable at demand, or at a less time
then six months from borrowing thereof.”


APRIL 1984

issue was also limited to the same area. However, Bank
of England notes were made legal tender at this time.14
This legislation provided legal force to the practice that
had already been adopted by other banks o f maintain­
ing their reserves in the form o f Bank o f England notes.

The Government Steps In: 1793—1821
The Napoleonic Wars between England and France
began in 1793. With the exception o f a m inor truce, the
war continued until N apoleon’s abdication on April 6,
1814.15Government demands from the Bank for financ­
ing rose substantially during the war. O f course, the
Bank's contract with its depositors to redeem its notes
at a fixed price in terms o f gold got in the w ay o f the
government’s interest and, on February 26, 1797, the
King and Privy Council ordered the Bank to suspend
specie payments, a suspension that was to last for
more than 20 years."’
During the suspension, control o f the money supply,
which had rested with the Bank’s owners, was largely
usurped by the government. Clapham notes that
The minutes o f the Court and those o f the Committee o f
Treasury are full o f ........ requests for help from Perceval,
a n d o f th e B an k's re lu c ta n t bu t in v a r ia b le a c­

Th e Bank app a ren tly acq u iesced because o f an
“understanding, a gentleman's understanding... to do
this business and to do it in the way most convenient to
the Treasury."18
Given the transfer o f monetary control to the govern­
ment, it is, perhaps, not surprising that the Bank's
Board o f Directors became unusually lackadaisical in
their attention to duties. As a result o f absenteeism, the
Bank Court experienced difficulty in maintaining a
quorum. Letters w ere sent to a number o f directors
that “pointedly asked 'when their attendance could be
depended upon’ " for "too much o f the business had

13Bisschop, p. 198.
14Andreades (1924), p. 261; Bisschop, p. 198.
15There was, of course, Napoleon’s “Campaign of 100 Days” be­
tween his escape from Elba on March 1, 1815, and his defeat at
Waterloo on June 18, 1815. Due to its brevity, it is ignored in this
,6The wording of the order ran as follows: The Bank will “forbear
issuing any Cash in Payment until the Sense of Parliament can be
taken on that Subject.” Clapham, vol. 1, p. 272.
17Clapham, vol. 2, p. 33; see, as well, Viner (1937), p. 122; Cannan
(1919), p. xi.
18Clapham, vol. 2, p. 11.
19Clapham, vol. 2, p. 31.

APRIL 1984


been done 'by a Single Director with the assistance of
the Head o f the Discount Office’.” 19

Return to Private Control
The suspension played an important role in the
governm ent’s effort to wrest control o f the money
supply from the Bank. In the absence o f suspension,
"control” of the stock o f money would have meant little
to the government since the requirement to redeem
notes at a fixed price in terms of specie eventually
w ould have (and did during 1790—96) placed an effec­
tive constraint on note issue.
The public was never keen on the suspension, and
the Bank made this the political issue in its fight to
regain control of the m oney supply. In October o f 1797,
six months after it was ordered to suspend payments,
the Bank indicated that it could “with safety resume its
accustomed functions (payment o f specie), if the polit­
ical circumstances o f the country do not render it
inexpedient.”211The Bank’s report was virtually ignored
by government. In June o f 1810, the "Beport from the
Select Committee on the High Price o f Bullion” recom ­
m ended to Parliament that the resumption o f specie
payments (at the old par) begin within two years. The
issue was not even taken up for discussion until July of
the following year.21 A vote on the recommendation
was taken in the House o f Commons in 1811. The
House voted 180 to 45 against the issue.
On its own initiative, the Bank began partial resump­
tion o f specie payments for notes of 5 pounds or less in
Januaiy o f 1817. Early in 1819, however, Parliament
required the Bank to discontinue the practice.22 Parlia­
ment had promised on five different occasions to even­
tually return to specie payments, but continued to
drag its feet on fixing a date. Finally, on July 2,1819, the
House o f Commons passed an act permitting the re­
sumption o f cash payments (bullion a n d coin) after
May 1, 1822. At the request o f the Bank, this date even­
tually was moved forward to May 1, 1821 23

There are a number o f important points to draw
from the previous discussion in analyzing the incen­

20Clapham, vol. 1, p. 272. See, as well, Cannan, p. xi.
2'Viner, p. 171.
22Viner, p. 172.
23Viner, pp. 172-73.

tives faced by the central bankers. First, given the one
exception noted, the right to control the money supply
was held privately. This right, in the form of ownei'ship
shares in the Bank, was traded in an organized market.
Any expected changes in the future profits o f the Bank
w ould be reflected by changes in the price o f Bank
shares and w ould immediately affect the wealth of
Bank owners.
Second, the owners of the Bank had loaned consid­
erable sums to the government in perpetuity at fixed
rates o f interest. By 1743, the sum was w ell in excess of
nine million pounds.
Finally, the Bank’s contract with its customers to
redeem its notes at a fixed price in terms o f gold was a
voluntary arrangement. An official (established by law)
gold standard did not exist in England until 1821. In
fact, the gold standard came about largely as a result of
the Bank’s continuous prods to an unwilling Parlia­
The following discusses how this unique incentive
structure faced by E ngland’s central bankers in­
fluenced the monetary growth rate.

Like the right to control the production o f any com ­
modity, the right to control the production of money is
valuable. The central bank, at the cost o f a few cents
worth o f paper and ink, can produce a $100 bill (or a
100-pound note) that can be exchanged in the market
for $100 worth o f resources.

The Flow o f Profits
The central bank introduces m oney into circulation
by exchanging units o f m oney (which it prints) for
commodities. These comm odities may be either real or
financial assets. Since the bank buys these assets at
market prices, the expected flow o f nominal profits
generated by the purchase o f the assets is equal to the
nominal interest rate times the price o f the assets pur­
chased. This is equivalent to the nominal interest rate
multiplied by the quantity o f money exchanged for the
assets. The flow of real revenue is simply the nominal
flow divided by the price level (i X M/P).
Since w e are interested in relating the bank's real
revenue flow to the rate of money production, account
must be taken o f the fact that, at higher rates o f money
production (higher rates o f inflation), people will want


APRIL 1984


Table 1
The Revenue From Money Production


i = r+ir

R = i(M/P)







= the rate of inflation (rate of monetary growth)

M/P = the stock of real purchasing power demanded
i(M/P) = the profit from monetary production (inflation)

population are stationary, that the real interest rate is
10 percent, and that the nominal interest rate is equal
to the sum o f the real interest rate and rate o f inflation
( i = r + t t ).
The numbers in the first two columns o f table 1
indicate that, as the rate o f inflation rises (falls), the
public's demand for real purchasing pow er falls (rises).
The third column indicates the nominal rate o f interest
at the various rates o f inflation. The fourth column
indicates the profit stream at the different rates of
As the rate o f inflation falls from very high rates, the
bank’s profit from inflation initially rises because peo­
ple choose to hold a greater amount of real purchasing
power. Reducing the rate o f inflation increases total
profits up to a point (1.80 real goods per unit o f time in
this example), after which further reductions in the
rate o f inflation cause profits to fall. In this example, the
profit-maximizing rate o f inflation (monetary growth
rate), which is the one the central bankers will choose,
is 20 percent.24

‘ indicates maximum profit

to hold less o f their wealth in the form o f real purchas­
ing pow er (M/P). Other things unchanged, the flow o f
real revenue w ould decline at higher rates o f money
growth because M/P declines. Even though M is rising,
the price level rises faster. However, other things are
not unchanged. Faster money growth increases the
rate o f inflation and this raises the nominal interest
rate (i).
Faster m oney prod u ction exerts tw o opposing
forces on the bank’s real revenue. One force tends to
reduce revenue, w hile the other tends to increase
revenue. In general, there is a unique rate o f money
growth (and rate o f inflation) that will maximize the
flow o f real revenue.

The Flow o f Profits
Among other things, the above result depends upon
the particular set o f operating rules the bank faces.
Apart from the particular assum ptions expressed
above, the foregoing example does not constrain the
bank in any way. If additional rules w ere imposed, the
profit function m ay change. As a result, the central
banker’s would be confronted with different incentives,
causing them to select a different monetary growth
The Bank o f England was founded on the condition
that the stockholders grant a substantial loan to the

A Simple Example
Table 1 presents a hypothetical example relating
different rates o f inflation (or rates o f m onetary
growth), 'ir, to the public’s demand for real purchasing
power, M/P. In order to facilitate the calculation o f the
rate o f inflation that maximizes the bank’s revenue
flow, suppose that the public knows the rate o f infla­
tion w ith certainty (extreme rational expectations),
that changes in the monetaxy growth rate affect only
the rate o f inflation and the public’s desire for real
purchasing pow er but no other real variables, that the
cost o f producing nominal units o f money is zero (so
revenue and profits are identical I, that real output and

Digitized for

24Since R = i x (M/P) = (r + tt)(M/P), real profits are maximized





(r + ir) =

Hence, nm(— + 1) = -1,

where nm = the elasticity of demand for real purchasing power with
respect to the rate of inflation. When r = 0, this result reduces to
nm = -1 which is the familiar result obtained by others.
See Friedman (1953), pp. 251-62; Friedman (1971), pp. 846-56;
and Bailey (1956), pp. 93-110.


government at a fixed rate o f interest. By 1743, that loan
amounted to almost 10 million pounds. At the point
when these loans w ere made to the government, the
interest rate charged was below the market rate. (Becall
the 1G94 loan at 8 percent when the market rate on
long-term loans to the Crown was 14 percent.) This
subsidized loan rate is a payment made by the Bank to
the government for the lease rights to the production of
money. The right was never granted to the Bank in
perpetuity. Bather, as indicated above, the Bank's char­
ter came up for review periodically.
The cost to the Bank o f its government loan depends
upon the market rate o f interest. If the coupon rate is c
and the amount loaned is L, the nominal value o f the
lease payment per unit o f time is (i —c)L. The real value
is the nominal amount divided by the price level,
(i —c)L/P. The higher the nominal rate o f interest, i,
relative to the coupon rate, c, the larger the cost o f the
lease to the Bank.
An additional constraint is relevant. The quantity of
notes the Bank could issue was restricted by law to an
amount less than or equal to the capital invested by
stockholders. Since the capital represented the loan to
the government, M must be less than or equal to L.
Given this constraint, the Bank’s owners will choose
M = L because the flow o f real profit is highest in this
case, other things the same (see insert). Consequently,
the Bank's profit is simply the coupon rate earned on
the loan, c, times the loan (which is equal to the quanti­
ty o f notes issued, M), divided by the price level, P.
Table 2 illustrates the effect o f this set o f rules on the
profit-maximizing rate o f inflation for the Bank. The
first three columns o f table 2 simply reproduce the first
three columns o f table 1. Column 4 calculates the real
profits o f the Bank under the new set o f rules where the
coupon rate, c, is assumed to be 10 percent. Note that
profits are maximized at a zero rate o f inflation rather
than the 20 percent rate obtained previously.-5

25This result is completely general as long as the demand for real
purchasing power is inversely related to the rate of inflation. In this



Since the derivative of the profit function with respect to the rate of
inflation is negative, it does not pay the Bank to generate an inflation
by continuously expanding M and, of course, the constraint that
M=sL will eventually become binding. Given that profit maximization
requires M = L from expression 1, a deflation would not benefit the
Bank because it would require M to fall below L. As a result, the Bank
will choose a zero rate of inflation.
In addition, the Bank’s owners tended to be net monetary credi­
tors as a class and this further reduced their incentive to inflate.

APRIL 1984

Table 2
An Example of the Bank’s Profit


i = r+w






'indicates maximum profit

The Role o f Specie Payments:
A Contract f o r Price Level Stability
It is in the interest o f Bank owners to inform the
public o f their intention to maintain a relatively low
inflation rate. Demand for the Bank's product does not
rise until the anticipated rate o f inflation declines. This
eventually will result from the Bank’s policy o f main­
taining a relatively low monetary growth rate. The ow n­
ers o f the Bank, however, chose to hurry the adjust­
ment o f expectations by "marketing” their bank notes
in a particular way.
In marketing its notes, the Bank guaranteed its cus­
tomers a low rate o f price inflation. This guarantee took
the form o f a contract to redeem Bank notes at a fixed
price (a fixed weight o f gold). The contract can be
thought o f as insurance against the overissue o f Bank
notes, because it pledged the original investment o f the
Bank’s stockholders as surety for meeting the con­
tract.26 If bank notes w ere issued in such quantity as to

There were about 1,300 original subscribers to the Bank stock.
Many of them were London businessmen who were “linked
sneeringly with the rather ill-famed money-lending scriveners.”
Others were Gentlemen and Esquires, “people who . .. live idly as
gentlemen'.” See Clapham, vol. 1, pp. 273-89.
26“That double event, (1) a low identification cost to everyone about
the intermediate commodity and (2) specialist-experts who provide
quality assurance and information more cheaply than novices can
provide for themselves, explains the use of a low identification cost
commodity as a general intermediary medium of exchange-money.
It permits purchase of information from lower cost sources, a cost
reduction that exceeds the added cost of using an intermediary
good for indirect exchange.” Alchian, pp. 117-18.



APRIL 1984

The Profit Function and Sharing Arrangement
The Bank’s flow o f real profit is i(M/P), as before,
minus the opportunity cost o f the government loan,
(i-c)IV P .

R' =



(1 —

c )







•L . L
l— + c—

Given the constraint that M < L, it is clear from
expression 1 that the Bank’s owners will choose
M = L. The Bank’s profits are highest in this case,
other things the same. As a result, the profit function
reduces to

R = c— = c—

The capitalized value o f the nominal profit stream

Since the above expression takes account o f the
opportunity cost o f issuing the government loan,
expression 3 is the net increase in stockholder
wealth that derives from the right to issue money.
The increase is proportional to the loan IL) where
the proportion is determined by the ratio c/i. The
higher is c relative to i, the greater is the stockhold­
ers' share and the low er is the government’s share.
For the initial loan o f 1,500,000 pounds, the stock­
holders' share was 857,142 pounds (1,500,000 X
.08/.14) in terms o f the prices that existed in 1694.
The government’s share was 642,857 pounds.
Long-term interest rates fell dramatically after the
formation o f the Bank.1 In part, this was due to the
eventual effect o f the Bank’s choice o f a zero rate o f
inflation on price expectations. O f course, the de­
cline in the nominal interest rate had the effect o f
raising the Bank’s share o f the net wealth derived
from issuing money (c/i in expression 3 rises as i
falls). As noted above, the government renegotiated c
downward each time the Bank’s charter was re­
newed in an effort to maintain the relative shares.
’ See Homer, pp. 131 and 161; Martin (1865), pp. 24-25; and
Rogers (1887), p. xiv.

cause their market price in terms o f gold to fall below
the price promised by the Bank, people w ould arbi­
trage the difference by trading gold for notes in the
market at the low price and exchanging the notes for
gold at the Bank for the higher price. In the process,
wealth would be transferred awav from stockholders
to those engaging in the arbitrage. The guarantee was
believable because customers knew that stockholders
would lose wealth if the Bank overissued its notes
relative to the supply of goods in general and gold in

270f course, the guarantee is not perfect. New gold discoveries or
improvements in mining technology would cause the price of gold
and Bank notes to fall in terms of, say, a standard commodity
basket. However, the guarantee, while imperfect, was operational. It
provided a relatively low-cost method of metering the Bank’s rate of
note production and policing the guarantee.

Digitized for18

When, for all practical purposes, the government
took control o f the money supply in 1793, the con­
straints facing the decisionmakers changed substan­
tially. Recall that the government did not expropriate
ownership rights in the Bank outright. Had they done
so, it w ould have been a clear land, possibly, politically
unsavory) transfer o f wealth from stockholders to the
government. The government, however, did the next
best thing from its point o f view. It expropriated the
wealth o f the stockholders by a m ore circuitous
When the government took over, the Bank held a
loan which, while an asset to the Bank, was a liability to
the government. In terms of the example used here, an
increase in the rate o f inflation increases the nominal

APRIL 1984


interest rate, i, and increases the Bank's opportunity
cost o f the governm ent loan, (i —c)L/P. In effect,
accelerating the rate o f inflation raised the lease pay­
ment the Bank made to the government.
Further, during its period o f control, the government
continuously violated the constraint that the quantity
o f notes in circulation not exceed the capital o f the
Bank. The government did not wish to be bound by the
same rule that it believed appropriate in regulating the
behavior o f the Bank's stockholders.
As a result of the different constraints faced by Bank
ow ners vs. the governm ent, w e should expect to
observe relatively low rates o f monetary growth and
inflation during periods when the money supply is
controlled bv the private owners o f the Bank o f En­
gland and more rapid rates o f m onetaiy growth and
inflation during the periods o f government control. In
addition, the dem and for real purchasing pow er
should be low er during the period of government con­
trol and the price o f Bank stock should decline.


The Behavior o f English Prices
One o f the more interesting pieces o f evidence con­
cerning the effect o f different incentives is England’s
history o f inflation during the period o f private m one­
tary control. England’s m oney supply was under
private control for almost 200 years, and the rate of
inflation during this period was statistically indistin­
guishable from zero.
From the establishment of the Bank in 1694 until the
beginning o f the Napoleonic War's in 1793 when the
government usurped control o f the money supply, the
annual average rate o f inflation in England was .01
percent. In 1821, after the Napoleonic Wars, the govern­
ment returned control o f the m oney supply to the Bank
and, at the Bank’s insistence, established an official
gold standard. Private control continued until 1913.
During this period, 1822-1913, England's annual aver­
age rate o f inflation was .42 percent which, again, is
statistically indistinguishable from zero. In contrast, by
1931, the English government had taken complete con­
trol o f the m oney supply. Since that date, the annual
average rate o f inflation has been significantly positive
at 6.47 percent.28

T able 3

The Real Price of Bank Stock,
Ps/P = 1.375 ,232r + 005RW - ,141D0
Rho =


R2 = .72
DW = 2.02
where: Ps/P = the real price of Bank stock.

= the yield on 3 percent consols.


= real weekly earnings.


= a dummy variable for the period subsequent to
government's seizure control. D0 = 1 for the
years 1793-1832 and zero otherwise.

'The price of Bank stock is the annual average calculated from the
daily price quotes reported in Gentleman's Magazine for the
years 1780-1832.
2Corrected for first-order autocorrelation, t-values in parentheses.
3The E. W. Gilboy series for average weekly earnings is spliced to
the A. L. Bowley and G. H. Wood series by a factor which is the
ratio of the average levels of the two indexes over the years
1791-93 and then divided by the price level. See R. B. Mitchell,
Abstract for British Historical Statistics (Cambridge University
Press, 1962), pp. 347-48.
‘ Significantly different from zero at the 5 percent level.

The Napoleonic Wars:
Additional Evidence
The above data regarding the history o f English infla­
tion are consistent with one o f the implications o f the
theory. Other aspects o f the theory can be examined by
considering data from the period immediately before,
during and after- the Napoleonic Wars. The war is im­
portant because the transfer o f control o f the money
supply during the war was accomplished through a
“gentleman’s understanding” rather than an outright
government expropriation o f Bank ownership. As a
result, ownership shares in the Bank continued to be
exchanged by private individuals, and changes in the
value o f these shares along with changes in the de­
mand for real purchasing pow er provide further evi­
dence in regard to the theory.

The Market Price o f Bank Stock
28This period includes the Depression and World War 11. If these years
are excluded and the inflation rate is calculated over the period
1946-82, the mean rate of inflation is 6.87 percent (t-score = 9.10).
The t-scores for the periods 1694-1793, 1822-1913 and 1931-82
are .145, .489 and 6.62, respectively.

Table 3 presents an estimate o f the effect o f the
government takeover and suspension o f specie pay­
ments on the real price o f Bank stock. The estimate



controls for the effect o f business cycles (proxied by
average annual real weekly earnings o f men in full-time
employment) and the interest rate, and includes a
dummy variable for the period subsequent to the sus­
pension. The coefficients o f the business cycle proxy
and the interest rate are significant and have the ex­
pected sign. The coefficient o f D„ is negative and sig­
nificant. Its magnitude implies that the government
takeover caused the real price o f Bank stock to fall
''permanently" by about 11 percent.21’ In short, the
government chose a rate o f inflation that was inconsis­
tent with maximizing the real flow o f Bank profits (from
all sources) and this was reflected in the price o f Bank
The price o f the stock did not return to its original
level w hen monetary control was returned to the
stockholders in 1819. “’ This was tested by including a
second dummy variable that assumes the value of 1 for
the period 1793-1818 and zero otherwise, along with
D„ that assumes a value o f 1 for the period 1793-1832
and zero otherwise. The coefficient o f the second dum­
my is insignificant, indicating that the variable is re­
dundant. That is, singling out the 1793-1818 period
adds nothing to the explanatory pow er o f the equation.
In addition to this evidence, monthly data for the
price o f Bank and India Company stock are available
for the period 1780-1801.31 There appeal's to be a break
in the ratio o f the price o f Bank stock to India Company
stock in 1793. Before then, the mean o f the ratio was .93
with a standard deviation o f .11. After 1793, the mean
fell to .83 with a standard deviation o f .03. The decline
in the ratio is statistically significant.32 The price of
Bank stock apparently declined relative to India Com ­
pany stock by about 11 percent, virtually identical to
the estimated decline produced by the regression in
table 3.

Disgruntled Stockholders
Understandably, the stockholders were restive dur­
ing the suspension. In 1801, Alexander Allardvce,

29The estimates are adjusted for first-order autocorrelation. They
were checked for second-order autocorrelation with the result that
Rho 2 was insignificant.
30During the later part of the suspension, various moves to resume
specie payments were afoot. As early as 1810, the Bullion Report
advocated a return to specie payments in 1812. In addition, the
Bank had begun a partial resumption in 1817, and in 1819 Parlia­
ment finally committed itself to a specific date for resumption. For
purposes of the following test, I terminate the period in 1818, the
year prior to Parliament's decision to return control to the Bank.
31See Sinclair (1803), pp. 22-48.
32The t-score = 10.78.


APRIL 1984

Table 4
Growth of Bank Money and Inflation,










1B. R. Mitchell (1962), pp. 442-43.
2lbid., pp. 469-70. The Schumpeter index for consumer goods is

spliced to the Gayer, Rostow, Schwartz index of domestic and
imported commodities by a factor which is the ratio of the average
levels of the two indexes over the years 1821-23.
3The period of the Peace of Amiens (the year 1802) is excluded
from this period.
'Significantly different from zero at the 95 percent confidence

spokesman for the critics, m oved that a complete
accounting o f the Bank’s financial condition be pre­
sented to the stockholders so that the Court might
“ declare a dividend o f the w hole profits, the Charges of
Management only excepted, as the Law directs.”33
Beal dividend payments, inclusive o f bonuses, did
not increase during the suspension and the real value
o f Bank stock declined. These two factors along with
interest-free loans made by the Bank to the govern­
ment must have appeared to critics as a thinly veiled
expropriation o f wealth.34 They no doubt recognized
the spirit o f Charles II lurking in the government.

As was the case for the more extended period dis­
cussed above, the rate o f inflation is indistinguishable
from zero (see table 4) in the years im mediately preced­
ing the government takeover and those following re­
sumption o f specie payments in 1821. In contrast, the
price level rose significantly (at an average annual rate

33Clapham, vol. 2, p. 40.
34ln 1799, when the market rate of interest stood at 5.07 percent on
long-term securities, the Bank made a "loan” to the government of 3
million pounds interest-free for six years. The present value of this
gift was, roughly, 770,000 pounds. In addition, when the loan came
due in March of 1806, the government asked that the loan be
renewed until a point in time six months after a “Definitive Peace.”
The government offered to pay 3 percent interest. At the time, the
long-term interest rate was considerably higher and 3 percent con­
sols were selling at about a 40 percent discount.
35The years 1814-21 are treated separately in table 4. During this
period, various steps were being taken to return to specie payment
(see footnote 30).


Table 5

The Demand for Real Purchasing
Power, 1780-1832_____________________
M/P = -1.535 + 2.002t + .573RW - 18.678D0
Rho =


R2 = .70
DW = 1.97
where: M/P = the stock of real purchasing power,

= time in years.

RW = real weekly earnings.

= a dummy variable for the period subsequent to
government’s seizure of monetary control. D0 = 1
for the years 1793-1832 and zero otherwise.2

APRIL 1984

Table 5 presents a regression that estimates the
effect o f the government’s seizure o f control over the
money supply on the demand for real purchasing
pow er by controlling for the effect o f business cycles,
population and the alternative cost o f holding money.
Annual population data are not available back to 1780,
so time is used as a rough proxy to control for popula­
tion growth. The interest rate on 3 percent consols was
included as a measure o f the alternative cost o f holding
wealth in the form o f money. In addition, a dummy
variable is included to test for a shift in the relationship
in the period subsequent to the government takeover.
The coefficients o f the proxies for the business cycle
and population are significant and have the expected
signs. The interest rate proved insignificant and was
excluded from the estimate. The coefficient o f D„ is
negative and significant. Its magnitude implies that the
demand for real purchasing pow er fell by about 12
percent when the government seized control o f the
Bank 37

'Corrected for first-order autocorrelation, t-values in parentheses.
2Since the hypothesis excludes positive values for D0, a one-tailed
t-test is employed.
"Significantly different from zero at the 5 percent level.

o f 4.85 percent) during the years 1793-1813.35 This was
a result o f a significant increase in the m onetary
growth rate. Note that the rate o f inflation closely cor­
responds to the rate o f growth in the money supply
during this period.30

The Real Value o f Bank Money
The theory implies that the demand for real pur­
chasing pow er will fall if the guarantee regarding the
low rate of inflation is broken.30

36The issue of private vs. government control of the money stock
might seem to be a red herring since the money supply and price
level always rise during wars. This, however, was not the case in two
previous instances. During the Seven Years' War (1755-63), the
government did not tamper with the Bank's control over the money
supply, and the mean rates of growth in money and prices were 2.94
and 1.18 percent, respectively. Neither of these magnitudes differs
significantly from zero. Similarly, during the War of Jenkin's Ear
(1739-43), the mean rates of growth in money and prices were
-1.62 and .66 percent, respectively. Again, neither of these differs
significantly from zero.
37Exactly when the public became aware of a break in the trust is
problematical. They certainly were aware of it by 1797 when the
government ordered the Bank to suspend specie payments. The
data, however, suggest an earlier date. Prices began rising rapidly
in 1790 and, shortly afterwards, the public began arbitraging the
difference between the price of gold in terms of notes at the Bank
and its price in the foreign market. The Bank’s bullion account began
to decline in 1791, then fell substantially in 1792. The following

The above analysis suggests that decisions regarding
the control o f m oney depend more on the incentives
individuals face in making choices than on the partic­
u lar in d ivid u a ls w h o m ake the ch oice. Various
methods o f organizing monetary control produce dis­
tinct policy outcomes insofar as they confront policy­
makers with different incentives. Since it is unclear, for
example, that the incentives confronted by the policy­
maker w ould be much different if monetary control
were placed in the hands o f Congress or the Treasury
instead o f the Board o f Governors, it is unclear that the
adoption o f either o f these alternatives would cause a
noticeable change in policy. Additional research along
these lines may prove helpful in suggesting a system of
incentives that will induce the present-day equivalents
o f the Court o f Directors to assign the desired weights
(whatever they happen to be) to present and future
consequences in reaching decisions regarding m one­
tary control.

assumes the public became aware of the break in 1793 when the
war with France began.
38As was the case with the price of Bank stock, the demand for real
purchasing power did not return to its original level when the govern­
ment returned monetary control to the Bank and the guarantee was
reinstated. This was tested by the same procedure as that employed
in the case of the price of Bank stock. The results were the same.
The coefficient of the second dummy was insignificant, indicating
that it is redundant.


APRIL 1984


Alchian, Armen A.


________ “Government Revenue from Inflation,” Journal of Politi­
cal Economy (July/August 1971), pp. 846-56.

"Some Economics of Property Rights,” Eco­

Homer, Sidney. A History of Interest Rates (Rutgers University
Press, 1977).

nomic Forces at Work (Liberty Press, 1977).

Andreades, A.

History of the Bank of England (P. S. King and Son,

Bailey, Martin L.

“The Welfare Cost of Inflationary Finance," The

Journal of Political Economy (April 1956), pp. 93-110.

Balbach, Anatol B. "How Controllable is Money Growth?” this Re­
view (April 1981), pp. 3-12.
Bisschop, W. R. The Rise of the London Money Market: 1640-1826
(Burt Franklin, 1967).
Cannan, Edwin.
Son, 1919).

The Paper Pound of 1797-1821 (P. S. King and

Johannes, James M., and Robert H. Rasche. “Predicting the Money
Multiplier,” Journal of Monetary Economics (July 1979), pp. 30125.
Kemmerer, Edwin W.

Gold and the Gold Standard (McGraw-Hill,

Macleod, M. A. The Theory of Credit, 2nd ed. (Longmans, Green
and Co., 1897), vol. 2, part 2.
Martin, Frederick.
Co., 1865).

Stories of Banks and Bankers (Macmillan and

Rogers, Thorold. The First Nine Years of the Bank of England
(Clarendon Press, 1887).

Clapham, Sir John. The Bank of England a History, vols. 1 and 2
(Cambridge University Press, 1958).

Sinclair, Sir John. The History of Public Revenue of the British
Empire (A. M. Kelly: New York, 1803).

Friedman, Milton.

Viner, Jacob. Studies in the Theory of International Trade (Harper
and Brothers, 1937).

“Discussion of the Inflationary Gap," Essays In
Positive Economics (University of Chicago Press, 1953).


Money Growth Variability
and GNP
Michael T. Belongia

MjiCENTLY, a number o f economists have argued
that sharp fluctuations in the short-run growth rate of
M l since 1979 have reduced GNP growth, raised in­
terest rates and generated expectations of higher fu­
ture inflation. Milton Friedman, for one, has concluded
that variable money growth — bv producing these
conditions — was responsible for the shorter and more
abrupt cycles in real income experienced over that
period.1 Based on slightly different analyses, Bomhoff,
and Mascaro and Meltzer also have concluded that
variable money growth has tended to lower the level of
output.2 Finally, a recent conference sponsored by The
Cato Institute was devoted entirely to the adverse
effects o f variable m oney growth and methods by
which money growth could be made more stable. ’
E con om ic th eory im p lies that variable m oney
growth could low er the level o f GNP by reducing its
short-run growth rate, if this variability were associated
with certain changes in money demand and velocity.
This article reviews the theoretical case for such a link
and provides empirical evidence on the existence of
this relationship. The results support the notion that
variable money growth — by increasing money d e­
mand and reducing velocity — has had significant
negative effects on both the level and the growth rate of
nominal GNP in recent years.

Michael T. Belongia is an economist at the Federal Reserve Bank of
St. Louis. John G. Schulte provided research assistance.

’ Friedman (1983).
2Bomhoff (1983); Mascaro and Meltzer (1983).
3See The Search for Stable Money (1983).

The most common approach to constructing a link
between variable money growth and GNP is based on
intermediate relationships involving money demand.
Although the theory behind these relationships sug­
gests that more variable m oney growth will increase
uncertainty about future econom ic conditions and in­
crease the demand for money, the empirical evidence
on this hypothesis has been mixed.4 The discussion
that follows, however, proceeds with a standard model
of m oney dem and and shows how more variable
money growth — by increasing uncertainty — can be
linked to a decline in the level o f income and, possibly,
the long-run growth rate o f GNP. Since the expected
effects o f variable m oney grow th on inflation are
assumed to be small, the conclusions that follow apply
to real GNP as well."’

The Basic Tobin Model
A money demand m odel derived by Tobin suggests
that there is an explicit relationship between uncer-

40ne statement of uncertainty's effect on money demand and interest
rates is found in Friedman and Schwartz (1982), p. 39:
Another variable that is likely to be important empirically is the degree of
economic stability expected to prevail in the future. Wealth holders are
likely to attach considerably more value to liquidity when they expect
economic conditions to be unstable than when they expect them to be
highly stable. This variable is likely to be difficult to express quantitatively
even though the direction of the change may be clear from qualitative
information. For example, the outbreak of war clearly produces expecta­
tions of instability, which is one reason war is often accompanied by a
notable increase in real balances — that is, a notable decline in velocity.

5For one argument to support this assumption, see Friedman.



APRIL 1984

Figure 1
The Links Between Variable Money Growth and GNP

taintv about future values o f interest rates and money
demand.1’ In its most basic form, the model assumes
that an individual can hold both money and govern­
ment bonds in his portfolio. Moreover, if the yield on
money is zero, both the expected return o f the portfolio
and its variance depend on only the bond yield and
the proportion o f the total portfolio held in bonds.
Therefore, in this simple world, an individual who
seeks to maximize utility bv holding some combination
o f cash balances and bonds in his portfolio faces a
tradeoff between return and risk. That is to say, he can
hold more bonds and increase the return on his port­
folio only at the cost o f increased risk: if the interest rate
rises, the value of his bonds will fall. He can reduce risk,
however, only by holding more cash balances, which
reduces earnings.
This m odel implies that risk and money demand are
negatively related.' If more variable money growth in­
creases uncertainty about future values o f interest
rates, greater money growth variability will result in an
increase in money demand. This inverse relationship
has been supported empirically in several studies.K
What remains to be seen, however, is whether this type
of shift in money demand can be linked to a decrease in
the level o f GNP.

6Tobin (1958).
7Some economists disagree with this conclusion. For discussions of
the theoretical indeterminacy of a sign relating uncertainty to money
demand and supporting evidence, see Blejer (1979), Levi and Makin
(1979), Smirlock (1982), Fieleke (1982), and Berson (1983).
8Klein (1977), Slovin and Sushka (1983), and Mascaro and Meltzer.

Digitized for 24

Money Demand, Velocity and GNP
The sequence of events depicted in figure 1 illus­
trates the first-round effects o f greater money growth
variability on uncertainty, m oney demand, velocity
and GNP.!1 Reading from the figure’s left side, more
variable m oney grow th is h yp oth esized to cause
greater uncertainty about future econom ic conditions.
Increased uncertainty increases the precautionary de­
mand for money. A higher level o f money demand
implies lower velocity (V). From the equation o f ex­
change, MV = Y, low er velocity clearly implies a lower
level for GNP (Y). Because GNP will shift to a low er level
with some lag, this level shift will be observed as a
tem poraiy decline in the growth rate o f GNP. After the
adjustment process is complete, the growth o f GNP
should return to its long-run equilibrium path unless
further changes in uncertainty and risk premia (or
other exogenous shocks) set off another round o f shifts
in the levels o f money demand and velocity.

Theoretical Indeterminacy:
Several Paths f o r GNP Are Possible
Whether increased uncertainty about future money
growth has any effect on GNP, however, is an empirical
issue. Moreover, if increased uncertainty does have
some effect on these variables, the nature o f its effect
could cause GNP to follow one o f several different
paths. For example, if the effect o f greater uncertainty is
a once-and-for-all shift in m oney demand, the level of

9This figure is adopted from a similar figure in Bomhoff, p. 98.


APRIL 1984

Table 1
Estimates of a Reduced-Form GNP
Equation Adding a Measure of Money
Growth Variability__________________
Y, = aQ + 2 b,Mt , + 2 Cj(EP - P), ,
i =0
j =0


2 gk VARMt_k + dQSt + et

affected permanently. A third possibility is that greater
uncertainty will alter investment decisions in a man­
ner that also changes the econom y’s long-run capitallabor ratio: in this case, both the level and growth rate
of GNP would be permanently lower. Finally, money
growth variability may have no observable effect on
uncertainty, money demand and velocity; in this event,
neither the level nor the growth rate o f GNP w ould be
affected. Hypotheses concerning the impact o f in­
creased money growth variability and these alternative
paths for GNP are tested in the next section.

k =0















- 0.047























'Absolute values of t-statistics in parentheses.
2This t-statistic applies to the null hypothesis 2 b, = 1.
i =0
3The F-statistic for the null hypothesis gD = g, = ... = g5 = 0 is
10.09, which is greater than the critical value of F6 69 — 2.22.

GNP will be permanently lower, but its growth rate
eventually will return to its former path. If the shift in
money demand is transitory, however, there w ill be a
short-run decline in the growth rate o f GNP, but nei­
ther the level nor the growth rate o f income will be

The effects o f variable money growth on GNP can be
tested by adding a measure o f m oney growth variability
to a basic reduced-form monetarist m odel o f nominal
GNP growth. The general reduced-form GNP equation
to be estimated is shown at the top o f table 1. This
equation expresses nominal GNP growth (Y) as a func­
tion o f the growth rate o f M l (M), the relative price of
energy (EP — P), the variability o f m oney growth
(VARM) and S, a variable that denotes periods o f major
strikes: the strike variable is defined as the change in
the quarterly average o f days lost due to strikes, de­
flated by the size o f the civilian labor force.10
The measure o f money growth variability chosen is
the square root o f a four-quarter moving average of
squared errors o f m oney growth forecasts over the
I/1950-IV/1983 sample period.11 The errors then were
used to construct a measure o f error variability meant
to represent changes in the risk or uncertainty faced by
econom ic agents as the pattern o f m oney growth
changes. Intuitively, one might conclude that risk has

10The model chosen is discussed in Tatom (1981). The initial speci­
fication of the equation in table 1 also includes high-employment
government expenditures as a right-hand-side variable. Pre-test
statistics, however, indicated no significant marginal contribution to
the model’s explanatory power from this variable. This pre-test
result is consistent with earlier studies that have found no long-run
effect of government spending on GNP growth. See, for example,
Andersen and Jordan (1968); Carlson (1978); and Hafer (1982). For
these reasons, the variable was omitted from the equation esti­
mated in this paper.
1'See Berson on the construction of a similar measure. The transfor­
mation is defined as:
[(UM?_, + UMf_2 + UMf-3 + UM?_4) - 4]1/2,
where UM represents unanticipated money growth, i.e., the re­
siduals from an autoregressive model of money growth. Errors were
generated by fitting a sixth-order autoregressive model to the
growth rate of M1.


APRIL 1984


Chart 1

Estimates of M onetary Uncertainty Under Actual
and Simulated Ml G row th

increased if forecasting errors begin to fall over an
increasingly w ider range. After all, the probability of
making an incorrect econom ic decision increases with
the probability o f making a large forecasting error. This
measure o f money growth variability, represented by
the solid red line in chart 1, shows that forecast errors
for M l growth have been considerably more variable
since 1979.

Pre-Test Estimation and
Lag Length Selection
The unknowns to be determined in this equation
prior to estimation are the lag lengths for money
growth, relative energy prices and m oney growth
variability (i.e., the n, p and q shown in table 1). These
values were chosen following procedures discussed
recently by Batten and Thornton.1- Pre-testing indi­

12Batten and Thornton (1983a, b) summarize an approach to the
selection of lag length and polynomial degree based on the work o'


cated the use o f contemporaneous and two lagged
quarterly values of the growth rate o f M l, contem pora­
neous and six lags for the relative price o f energy, and
contemporaneous and five lags for the measure of
money growth variability.13
The choice o f five lags for the measure of money
growth variability reflects the lagged responses of
money demand, velocity and GNP suggested bv theory
and depicted in figure 1. That is to say, increased
variability in m oney growth is expected to affect GNP
only after some lag; econom ic agents require sufficient
time both to discover the w ider band o f errors on
money growth forecasts and to adjust their behavior
accordingly. To test whether increased uncertainty

Geweke and Meese (1981); Mallows (1973); Schwartz (1978);
Akaike (1969); and Pagano and Hartley (1981).
,3The Pagano-Hartley t-ratios, final prediction errors and Mallows’
test statistic all suggested these lag lengths. These lag lengths were
fitted and chosen using ordinary distributed lag models without
polynomial smoothing.


APRIL 1984

Table 2
Implications of Alternative Pairs of Test Results for Estimated
Coefficients on Monetary Variability
Practical Implication

Individual Coefficients

Sum of Coefficients

1) Neither the level nor the growth rate
of GNP is affected

each equals zero

equals zero

2) The level of GNP is temporarily or
permanently lower but its long-run
growth rate is unaffected

some initial coefficients
are significantly negative

equals zero

3) Both the level and long-run growth
rate of GNP are permanently lower

some initial coefficients
are significantly negative

is significantly negative

has an effect on the level o f income, the relevant null
hypothesis is g() = g, = ... = g5 = 0, as shown in table 2.
Failure to reject this hypothesis w ould imply that
m oney growth variability had no effect on GNP.
If one or more individual coefficients indicate a sta­
tistically significant negative relationship between GNP
growth and money variability, the second issue o f in­
terest is whether this effect on the level and the growth
rate is transitoiy or permanent. In other words, it is
important to know whether greater money growth
variability causes a tem poraiy or permanent reduction
in the level and growth rate of GNP. This result can be
determined by testing the null hypothesis that X
k= 0
gk = 0. If this sum is not significantly different from
zero but some individual coefficients are significantly
negative, the results w ould im ply a transitoiy decline
in the growth rate o f GNP and either a temporary or
permanent reduction in its level. If this hypothesis also
is rejected, however, it can be determined that both the
level and growth rate o f GNP are permanently lower.
Implications o f possible test results are summarized in
table 2.

The results o f estimating the augmented GNP equa­
tion over the II/1962-IV/1983 sample period are given in
the first column o f table 1. The results reject each o f the
null hypotheses discussed above: some initial indi­
vidual coefficients for money growth variability are sig­
nificantly negative and their sum is significantly nega­

tive. Within the context o f the specified equation, these
results indicate that greater short-term variations in
the rate o f money growth tend to increase uncertainty
and money demand; as a result, permanent reductions
in both the level and the growth rate o f nominal in­
come are produced.
It also is important to note that the sum o f the

coefficients on monev growth ( X
b,l is not signifii= 0
cantly different from one after the addition o f a direct
measure o f money growth variability. This shows that
the one-to-one long-run correlation betw een the
growth rates o f money and nominal GNP remains, even
after the effect o f variable money growth is directly
taken into account.14
The significance tests on the other variables in­
cluded in the regression indicate that the strike vari­
able has negative effects on incom e growth. Also,
changes in the relative price o f energy have exhibited
som e significant positive long-run effects on GNP
growth. This latter result is not surprising; the impacts
o f short-run changes in relative energy prices are
usually measured as changes in inflation. Thus, the
relative energy price effect shows up in nominal GNP
(via the price change); and this explains the positive
sum coefficient for relative energy prices in this model.

14These results hold for a variety of variability measures, including a
moving standard deviation of money growth, squared money growth
rates and a multi-state Kalman filter estimate of the variance of
errors associated with one-quarter-ahead forecasts of money
growth. Unlike the criticisms of Allen with regard to uncertainty
results for money demand, these results for a GNP equation appear
to be robust with respect to the measurement of money growth
variability. See Allen (1982).


APRIL 1984


As a check o f the m odel’s robustness, the equation in
table 1 was re-estimated over a shorter II/1962-III/1979
sam ple period. This p eriod was chosen for tw o
reasons. First, the Federal Reserve changed its operat­
ing procedures in October 1979. Second, as shown in
chart 1, there was a sharp increase in money growth
variability after IV/1979. The results o f re-estimating the
GNP equation over the shorter sample period with new
values for VARM are given in the second column of
table 1.1S
The results for the shorter estimation period still
indicate that variable money growth temporarily low ­
ers the growth rate o f GNP. The long-run impacts on
the level and growth rate o f GNP, however, are no
longer significantly different from zero. Apparently, the
considerably lower variability o f money growth that
existed prior to 1980 did not produce anv long-run
impact on the growth o f GNP. Or, viewed differently,
even though variable money growth has a significantly
negative effect on GNP in both periods, permanent
reductions in its level and growth rate are found only
after 1980, w hen the variability o f m oney growth
The effects of money growth and relative energy
prices also follow lag patterns similar to those for the
longer sample period. However, the long-run effect of
relative energy prices is no longer significantly positive.
T h e o n ly o th e r a p p a re n t c h a n g e fro m th e full p e rio d
estim a tio n to this re s tric te d o n e is a d e c lin e in th e
e s tim a te d g r o w th rate o f v e lo c ity (the m o d e l’s co n sta n t
term ) to 3.0 fro m 5.5. H o w ever, sin ce th e g r o w th rate o f

velocity in this model is reallv a() +


gk, the implied

k= 0

velocity growth for the full-sample model is actually
4.14, which is not significantly different from 3.0.lfa In all
other respects, the results for both models are qualita­
tively similar and w ould seem to indicate that the
addition o f a money variability measure is robust with
respect to choice o f sample period.

' 5To reflect the less volatile pattern of money growth that prevailed
prior to 1980, the autoregressive model of money growth used to
generate values for the money variability measure was reestimated. An AR(1) model was found to whiten the residuals for a
model of money growth estimated over the pre-1980 sample.
16The F-statistic for H0: ac +

k= 0

critical value for F, 60 = 4.00.


gk = 3.0 is 1.83, less than the

The estimates reported in table 1 support the hy­
pothesized negative relationship betw een variable
money growth and GNP discussed elsewhere.17 H ow­
ever, the statistical measure o f m oney growth variabil­
ity is not expressed in units that have a clear econom ic
meaning. Therefore, the results in table 1 may be dif­
ficult to interpret directly, especially for policy pur­
poses. It may be useful to illustrate more intuitively
w hy some economists are concerned about the poten­
tial negative effects o f m oney volatility. This is done
below by using the equation in table 1 to repeat an
experiment recently suggested by Friedman.18
Friedman asked what the path o f GNP w ould have
been in recent years if the money stock had grown at
the following rates over these intervals: 7.1 percent
from III/1979 to III/1980; 6.1 percent from III/1980 to
III/1981; and 5.1 percent from III/1981 to III/1982.1UThe
6.1 percent three-vear average growth rate described
above is equal to its actual average over the same
period. The plots of both actual M l growth and Fried­
man’s smoothed m oney path are shown in the upper
panel o f chart 2.
While maintaining the same average growth rates of
money over four quarters, the Friedman scenario sig­
nificantly reduces the large quarter-to-quarter varia­
tions in M l growth that actually occurred over this
period. This result is shown clearly by the sharp de­
cline in money growth variability that is generated bv
these data; this new measure o f monetary uncertainty
is represented by the dashed line in chart 1. Over the
III/1979—III/1982 period, the more stable path o f M l
growth w ould have produced — in terms o f Fried­
man's analysis — a longer but less severe recession in
1980 and, beginning around mid-1981, an expansion
typical o f the postwar period (lasting about three
years). The projected path o f GNP under stable M l
growth is contrasted in the low er panel of chart 2 with
the projected path o f GNP under actual money growth.
The solid black line in the low er panel of chart 2 is the
path of GNP produced by a simulation o f the model
reported in the second column o f table 1 based on the

17For example, Friedman and Schwartz (1963b); Friedman; The
Search for Stable Money.

,9The experiment stops at this point because money growth acceler­
ated sharply and varied substantially over subsequent quarters.


APRIL 1984

Chart 2

GNP Growth and Alternative M oney Growth Paths
Actual and Smoothed Money Series

Actual and Smoothed Paths of GNP Growth




Simulat on with actual money
* /\
/ \y





Simulation with smoi jthed m o n e y /







1 /
\ /
\ A

» /



\ Y/ 1i
1 /
1 /
M N1 1' /
\ V '' /#





\\ 111 y
\ J
A '
I — ■v v 7 \ '
1 \
/ \l '
\i »





// /





11 //



* Actual


Y '
A\ '»

Iku ' / /
\v /,/ ^
\ \' l f







/ > S A\.

















APRIL 1984


smoothed m oney growth figures listed above.20 The
results are quite similar to Friedman's conjecture;
moreover, they depict clearly what some economists
claim are the prospective benefits o f more stable
money growth. The simulated path o f GNP growth —
under reduced quarter-to-quarter variation in M l
growth — shows higher average growth and much
narrower variation than does actual GNP growth over
this period. For example, actual GNP growth ranged
between —2 and 20 percent; under more stable money
growth, however, the simulated rates of growth in GNP
vary between 7 and 12.5 percent. Moreover, while
simulated GNP growth using actual money growth
rates fell to zero in III/1982 and was 5 percent or below
in three o f the 12 quarters shown, the simulated path of
GNP growth under less variable money growth fell
below 7.5 percent on only one occasion. In summary,
the contrasting results shown in chart 2 suggest that
more stable money growth could promote a higher
average level o f GNP growth and reduce the range in
which GNP growth fluctuates.

A number o f recent studies have argued that variabil­
ity in the quarter-to-quarter growth rate of money has
increased m oney demand and, therefore, decreased
the growth rate o f GNP in the short run. This article
investigates the link between variable money growth
and GNP bv adding a measure o f m oney growth
variability to a specific m odel o f GNP.
Th e results suggest that increased quarter-toquarter variation in the growth rate o f M l has some
transitory negative effects both on the level and growth
rate of nominal GNP; moreover, in more recent year's,
when the variation in money growth has increased
nearly threefold, there is some evidence that the
effects on the level and growth rate o f GNP have been
permanent reductions. If the effect of money variability
on inflation is small, as is generally thought, these
results imply a permanently low er level and, perhaps,
smaller growth rate o f real GNP.

A simulation experiment based on these results
illustrates the potential benefits o f more stable money
growth. Within the context o f the m odel used, growth
in nominal GNP w ould have been higher, on average,
and more stable since 1979 if the quarter-to-quarter
growth in M l had been substantially less variable than
it actually has been since then.

Akaike, H.

“Statistical Predictor Identification,” Annals of the Insti­
tute of Statistical Mathematics, vol. 21 (1969), pp. 203-17.

Allen, Stuart D. “Klein’s Price Variability Terms in the U.S. Demand
for Money,” Journal of Money, Credit and Banking (November
1982), pp. 525-30.
Andersen, Leonall C., and Jerry L. Jordan. “Monetary and Fiscal
Actions: A Test of their Relative Importance in Economic Stabiliza­
tion,” this Review (November 1968), pp. 11-24.
Batten, Dallas S., and Daniel L. Thornton. “Polynomial Distributed
Lags and the Estimation of the St. Louis Equation,” this Review
(April 1983a), pp. 13-25.
----------- “Lag-Length Selection Criteria: Empirical Results from
the St. Louis Equation,” Federal Reserve Bank of St. Louis Work­
ing Paper No. 83-008 (1983b).
Berson, David W. "Money Growth Volatility, Uncertainty, and High
Interest Rates," Federal Reserve Bank of Kansas City Economic
Review (November 1983), pp. 23-38.
Blejer, M. I. "The Demand for Money and the Variability of the Rate
of Inflation: Some Empirical Results,” International Economic Re­
view (June 1979), pp. 545-49.
Bomhoff, Eduard J.
Holland, 1983).

Monetary Uncertainty (Amsterdam: North-

Carlson, Keith M. “Does the St. Louis Equation Now Believe in
Fiscal Policy?” this Review (February 1978), pp. 13-19.
Evans, Paul.

“Price-Level Instability and Output in the U.S.,” Eco­

nomic Inquiry (April 1983), pp. 172-87.

Fey, R. A., and N. C. Jain. "Identification and Testing of Optimal Lag
Structures and Causality in Economic Forecasting,” in O. D.
Anderson and M. R. Perryman, eds., Applied Time Series Analysis
(Amsterdam: North Holland, 1982), pp. 65-73.
Fieleke, Norman S. “Fluctuations in Economic Activity and the
Money Supply: An Overview,” New England Economic Review
(May/June 1982), pp. 5-14.
Friedman, Milton. “What Could Reasonably Have Been Expected
From Monetarism: The United States,” paper presented at the
Mont Pelerin Society Meeting, Vancouver, British Columbia, Au­
gust 29, 1983.
Friedman, Milton, and Anna J. Schwartz. A Monetary History of the
United States, 1867-1960 (Princeton University Press, 1963a).

20The dashed line was derived by using actual money growth rates
and the errors from an autoregressive model fit to actual money
data. These data provided the basis for projected GNP growth from
111/1979-111/1982 under actual monetary conditions. Actual M1 data
then were replaced with Friedman’s figures for the 111/1979-111/1982
interval. The autoregressive model for money growth then was
re-estimated over 1/1960-111/1982 to generate a new error series
and a new measure of money growth variability. The coefficient
estimates reported in the second column of table 1 were used to
re-simulate GNP growth over 1979-82 in an environment of more
stable money growth. These simulated results are shown by the
solid black line in chart 2.


________ “Money and Business Cycles,” supplement to The Re­
view of Economics and Statistics (February 1963b), pp. 32-64.
________ Monetary Trends in the United States and the United
Kingdom: Their Relation to Income, Prices and Interest Rates,
1867-1975 (University of Chicago Press, 1982).

Geweke, J., and R. Meese. “Estimating Regression Models of Finite
But Unknown Order," International Economic Review (February
1981), pp. 55-70.
Hafer, R. W. “The Role of Fiscal Policy in the St. Louis Equation,”
this Review (January 1982), pp. 17-22.

Klein, Benjamin. “The Demand for Quality-adjusted Cash Balances:
Price Uncertainty in the U.S. Demand for Money Function,” Journal
of Political Economy (August 1977), pp. 691-715.
Levi, Maurice D., and John H. Makin. "Fisher, Phillips, Friedman
and the Measured Impact of Inflation on Interest,” Journal of Fi­
nance (March 1979), pp. 35-52.
Mallows, C. L. “Some Comments on Cp," Technometrics (Novem­
ber 1973), pp. 661-75.
Mascaro, Angelo, and Allan H. Meltzer. “Long- and Short-Term
Interest Rates in a Risky World,” Journal of Monetary Economics
(November 1983), pp. 485-518.
Pagano, Marcello, and Michael J. Hartley. “On Fitting Distributed
Lag Models Subject to Polynomial Restrictions,” Journal of Econo­
metrics (June 1981), pp. 171-98.

APRIL 1984
Schwartz, G.

“Estimating the Dimension of a Model," Annals of

Statistics, vol. 6 (March 1978), pp. 461-64.

Slovin, Myron B., and Marie Elizabeth Sushka. "Money, Interest
Rates and Risk," Journal of Monetary Economics (September
1983), pp. 475-82.
Smirlock, Michael. “Inflation Uncertainty and the Demand for
Money,” Economic Inquiry (July 1982), pp. 355-63.
Tatom, John A. “Energy Prices and Short-Run Economic Perfor­
mance," this Review (January 1981), pp. 3-17.
“The Search for Stable Money.” James A. Dorn, ed., The Cato Journal
(Spring 1983).
Tobin, James.

“Liquidity Preference as Behavior Toward Risk," The

Review of Economic Studies (February 1958), pp. 65-86.