Full text of Review (Federal Reserve Bank of Dallas) : May 1985
The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Economic Review
Federal Reserve Bank of Dallas
May 1985
1
Money in a Deregulated
Financial System
Gerald P. ODriscoll, Jr.
Despite fears to the contrary, deposit deregulation
will not result in either loss of monetary control or
the disappearance of money itself. Recent centralbanking literature has raised fears that deregulation
will lead to loss of monetary control. Similarly,
arguments from the theory of finance suggest that
deregulation will result in the disappearance of noninterest-bearing money and the loss of any distinction
between interest-bearing money and other financial
assets. Although valuable insights are provided by
these two bodies of literature, their stark predictions
are unwarranted.
13
Inflation and Permanent
Government Debt
W. Michael Cox
This article argues that post-World War II deficits
have been inflationary . A traditional assumption in
macroeconomics has been that government debt is
temporary. In that case, the issuance of debt in lieu
of current taxation does not stimulate aggregate
demand and, therefore, is not inflationary. This is
because debt issued today is matched by additional
savings to pay the future taxes necessary to retire the
debt. But in recent years Federal Government debt
has not been temporary . Government bonds have not
been matched by future taxes, and the bond-financed
deficits have contributed to the increase in the price
level. Statistical tests indicate that over the 1950-84
period, inflation in the United States was as closely
related to the growth in outstanding government debt
as to the growth in the monetary base .
This publication was digitized and made available by the Federal Reserve Bank of Dallas' Historical Library (FedHistory@dal.frb.org)
Money in a Deregulated
Financial System
Gerald P. O'Driscoll, Jr.
Senior Economist and Policy Advisor
Federal Reserve Bank of Dallas
A growing number of economists believe that
money would disappear in an unregulated financial
system . They argue that legal restrictions and the
regulatory system are responsible for the existence
of money as a distinct asset. In particular, they
predict that there would be no demand for noninterest-bearing financial assets like currency in an
unregulated financial system
In a less dramatic vein, a case has been made
that the long-run effects of deposit deregulation will
erode the central bank's ability to control monetary
growth. This would have serious implications to the
extent that monetary control is a necessary ingredient to a stable economy.
In 1980, Congress took the initiative in deregulating deposits by passing legislation providing
for the elimination of interest rate ceilings on
depository accounts. As a result, new types of
deposits paying market interest rates were introduced. Before this action, however, the Federal
The views expressed are those of the author and do not necessarily
reflect the positions of the Federal Reserve Bank of Dallas or the
Federal Reserve System. The author expresses his appreciation to
Lawrence H White of New York University for his helpful comments on earlier drafts of this article.
Economic Review I May 1985
Open Market Committee (FOMC) had announced a
change in operating procedure. Creater emphasis
was placed on control of reserves in order to control the monetary aggregates better. Deposit
deregulation might thus have made control of the
monetary aggregates impossible at the very time
that the FOMC had come to emphasize controlling
these aggregates.
The plan of this article is as follows. The first part
examines the central-banking literature on monetary
control in a deregulated environment. While it appears that the effects of deregulation thus far have
been less dramatic than many expected, the issues
raised in this literature are crucial for understanding
how banks would operate in a deregulated environment. The remainder of the article examines the
new literature on banking that has derived from the
theory of finance. In the latter, banks are not
viewed as creators of money but as pure financial
intermediaries.
Since both sets of literature focus on the effects
of paying competitive rates of interest on money,
this article examines the issue in some detail.
Analysis suggests that loss of monetary control is
unlikely. This conclusion is buttressed by the
historical experience of largely unregulated banking
systems.
The article also concludes that money would not
disappear in an unregulated financial system. Support for this position derives from analyzing the
essential properties of a medium of exchange and,
especially, from considering money's role as the
only perfectly liquid good. Yielding an explicit
return eliminates neither money's distinctive properties nor the demand for its special services.
Deregulation and monetary control
Deposit deregulation not only radically changes the
regulatory structure for banks but also significantly
affects the conduct of monetary policy.1 If deposits
were fully deregulated, the central bank could no
longer rely on disintermediation to control monetary
growth or to influence real economic activity. If
transaction accounts actually paid market rates of
interest, monetary policy could not change the opportunity cost of holding transaction balances. (See
Box A.)
The partial deregulation of deposits that has
already occurred does not correspond exactly to
this hypothetical case. First, deregulation is not
complete. Second, banks are still required to hold
reserves against all transaction accounts, which
drives a wedge between the market rates earned on
assets held by banks and the rates they can pay on
transaction accounts. 2 Nonetheless, the hypothetical
case certainly points to the direction in which the
monetary system is evolving. (See Box B.)
Moreover, the payment of near-market rates of interest on Super NOW accounts gives them a savings
feature heretofore lacking in transaction accounts; it
blurs the distinction between transaction and savings accounts. Thus, use of M1 as a "temporary
abode of purchasing power" is more attractive.
Observers anticipated that the introduction of Super
NOWs would lead to a once-and-for-all increase in
the demand for M1. 3
1. For purposes of this article, "bank" is used synonymously with
"depository institution." A fuller treatment would consider
other depository institutions (for example, savings and loan
associations) more explicitly. Further, this article focuses on
deregulation of deposits. Analysis of banks' expanded asset
powers would require a separate paper.
2. For an analysis, see John P Judd, " Deregulated Deposit Rates
and Monetary Policy, " Economic Review, Federal Reserve
Bank of San Francisco, Fall 1983, 27-44, especially 28-29.
2
Complicating the January 1983 advent of Super
NOW accounts, however, was the fact that in the
previous month the money market deposit account
(MMDA) had been introduced. MMDAs had a transaction feature: up to six third-party transfers could
be made per month (three by check). In theory,
then, MMDAs, which are a component of M2 but
not M1, could have "robbed" M1 of transaction
balances .
I n other words, the effects of all the deposit
deregulation on the demand for M1 were ambiguous. If money demand were less predictable,
however, then setting the proper targets for the
growth of M1 would be more difficult. These and
other considerations influenced the FOMC in its
decision in late 1982 to de-emphasize M1 temporarily in formulating monetary policy.'
Despite these concerns, the evidence accumulated thus far suggests that the worst fears have
been far from realized . Most shifting of accounts
apparently occurred within traditionally defined
aggregates. The evidence suggests that MMDAs attracted funds mainly from other components of M2
while Super NOWs attracted funds chiefly from
other components of M1. The comparatively small
amount of shifting that took place between
components of M1 and M2 appears to be nearly
offsetting.5
The fact that the partial deregulation of deposits
has had comparatively minor observable impact on
the monetary aggregates by no means impl ies that
concern over its potential effects was either
misguided or exaggerated . Individuals may still be
adjusting to the partial deregulation that has
already taken place. And, as indicated above, the
process of deregulating accounts at depository institutions is not over. (See Chart 1.)
Deposit deregulation is but one effect of an ongoing process of financial innovation. This process
3 For a fuller discussion of the potential effects. see Judd,
"Deregulated Deposit Rates and Monetary Policy "
4 Some analysts have argued that depository deregulation might
lead to permanent instability in the demand for money See
Judd, " Deregulated Deposit Rates and Monetary Policy,"
30-35. It should be noted that the FOMe has since returned to
targeti ng M1 .
5. See Frederick T Furlong, "New Deposit Instruments," Federal
Reserve Bulletin 69 (May 1983): 319-26; cf Judd, " Deregulated
Deposit Rates and Monetary Policy," 32-34
Federal Reserve Bank of Dallas
Box A
The Regulated Banking System, 1934-1980
Most of the distinctive features of the U S. banking
system were firmly established by the Banking Act of
1933 Among other things, the act specified lines of
commerce in which commercial banks could not
engage (for example, underwriting securities). Of
direct relevance here was the enactment of a system
of regulated interest rates on bank accounts. Payment
of interest on checking accounts was prohibited, as
was the payment of interest on interbank deposits.!
The Board of Governors of the Federal Reserve System
was empowered to establish interest rate ceilings on
savings accounts.
The Banking Act of 1933 mandated a sharp distinction between commercial banks, on the one hand, and
other depository and nondepository financial institutions, on the other hand Commercial banks were
heavily regulated as to the activities in which they
could engage. In return, however, they were given a
monopoly on the creation of checking accounts Two
consequences of this regulatory system are
noteworthy.
First, by sharply distinguishing between transaction
and nontransaction accounts, the Banking Act of 1933
indirectly established an identifiable empirical
counterpart to the theoretical concept of "transaction
money" in economics By effectively implementing
the intent of the act, the Federal Reserve System
prevented other types of deposits from acquiring a
transaction or medium-of-exchange feature .' The Fed
thereby delineated the category of deposit that should
be added to currency to determine the supply of transaction money The Banking Act of 1933 settled the
issue of which deposits are transaction accounts for
over 40 years.
Regulation Q facilitated monetary control in a second significant way. By selling a sufficient quantity of
assets from its portfolio, the Federal Reserve System
could push market interest rates further and further
above the maximum allowable interest rate payable
on bank accounts. This would raise the opportunity
cost of holding money.
Let i denote a short-term interest rate and d denote
the maximum allowable rate payable on the relevant
deposit category (zero for demand deposits) The opportunity cost of holding funds in a given deposit
category would be measured by (i-d). As i increases
relative to d, this cost rises. Individuals will tend to
Economic Review I May 1985
economize on their accounts at depository
institutions.
The process in which agents withdraw funds from
financial intermediaries and purchase financial assets
directly is described as disintermediation. Individuals
are not willing, however, to hold the same portfolios as
do financial intermediaries. Not being able to pool
risks, as do these intermediaries, individuals tend to
buy highly liquid, risk-free assets like U.S. Treasury
bills. Consequently, funds for mortgages and small
businesses dry up, slowing economic activity. This sectoral contraction shows up as slower growth in
nominal gross national product (GNP). (Had the
analysis begun with market rates above Regulation Q
levels, an expansionary monetary policy would have
produced the opposite results)
The conduct of monetary policy has thus been influenced by the regulatory structure. The policy question is how monetary control can be effected in a
partially or fully deregulated environment.
1. Payment of interest on bankers' balances helped integrate local
banks into a national capital market by enabling large correspon·
dent banks in money centers to bid for the funds of independent
local banks around the country The system was an effective
though imperfect substitute for a system of national branch bank·
ing See Eugene Nelson White, The Regulation and Reform of the
American Banking System, 1900-1929 [Princeton: Princeton Univer·
sity Press. 19B3), especially 65-74 Critics like Congressman Carter
Glass objected to the system's efficiency, for they did not want
funds moved from local markets to money centers See John H
Wood, "Familiar Developments in Bank Loan Markets," Economic
Review, Federal Reserve Bank of Dallas, November 19B3, 2-3
2 Financial institutions
In some New England states were eventually
able to slip through one legal loophole In 1972, state-chartered
thrifts in Massachusetts and New Hampshire began offering ac·
counts on which "negotiable orders of withdrawal" could be writ·
ten [hence the acronym "NOWs") These were essentially savings
accounts on which checks could be drawn. ApproXimately one year
later, Congress authorized NOW accounts for all depository institu'
tions in these states See Katharine Gibson, "The Early History and
Initial Impact of NOW Accounts," New England Economic Review,
Federal Reserve Bank of Boston, January/February 1975, 17-26 For
the spread to other New England states, see Ralph C Kimball. "Re·
cent Developments in the NOW Account Experiment in New
England," New England Economic Review, November/December
1976. 3-19; and Donald Basch. "The Diffusion of NOW Accounts
in Massachusetts," New England Economic Review,
November/December 1976.20-30 When introduced, the national
NOW accounts were modeled after these regional accounts
3
Chart 1
Super NOW Accounts as a Share of M1
PERCENT OF M1 MONEY STOCK
10
r---------------------------------------,
8
6
4
2
o
~----------~
__________
1983
1984
~L_
__________
~
1985
SOURCE : Board of Governors, Federal Reserve System
has the potential for causing even more far-reaching
changes in U.S. monetary institutions. The next section analyzes the direction in which recent innovations are taking the financial system .
Money and financial innovation
The effect of recent innovations is to move banks
and nonbanks alike in the same direction , Their
liabilities are (1) increasingly paying market interest
rates and (2) increasingly not subject to reserve
requirements. In the previous section, discussion
focused on the first characteristic. In this section,
attention will be paid to the second.
Though the two characteristics of deregulated
financial liabilities are logically distinct, they are
economically related . To pay competitive yields on
their liabilities, firms will minimize their holdings of
non-interest-bearing assets. Save for a brief interval
in 1980-81, nonbank financial firms were under no
requirement to hold non-interest-bearing reserves. As
a result, money market mutual funds (MMMFs), for
instance, have remained almost fully invested in
money market instruments. 6
In response to the development of MMMFs, banks
offered "sweep" accounts at the retail level. Market
forces thus moved banks and nonbanks alike to offer "zero-balance'" transaction accounts. A de4
posit balance is created to be instantly extinguished ,
Bank money (deposits) as we know it disappears in
the process, as do reserves and the central bank's
ability to conduct conventional monetary policy.
Banks developed zero-balance accounts to evade
the reserve requirements on transaction accounts. In
the absence of these requirements, the incentive to
adopt this subterfuge would be eliminated, While
bank deposits would be restored to their traditional
place, the tendency for banks to minimize their
reserve holdings would not diminish. Moreover, in
the process of circumventing reserve requirements
and deposit regulations, bankers have undoubtedly
acquired useful knowledge about minimizing the
level of reserves held against given deposit
balances.
Many of the forces moving the financial system
toward a world of nonreservable deposits are also
tending to minimize the use of currency. For in-
6 One author has found that money market mutual funds
generally hold a positive but small percentage of their assets
(less than 1 percent) in demand deposits at commercial banks;
apparently these deposits are used to honor redemptions
(Lawrence H. White, " Competitive Payments Systems and the
Unit of Account," American Economic Review 74 [September
1984): 707 n. 16).
Federal Reserve Bank of Dallas
Box B
A Chronology of Deposit Deregulation
Beginning in 1980, Congress took a more active role in
setting the pace of deposit deregulation. Depository
institution had been buffe ed by high and volatile interest rates and turbulent financial conditions for
more than a decad~ To a large extent. the financial
turbulence resulted from the volatility of inflation
rates. In addition, the breakdown of the old international monetary system of fixed exchange rates and its
replacement by a system of flexible exchange rates
put added burden on an already strained financial
system . Finally, the October 1979 decision of the
Federal Open Market Committee to place greater emphasis on controlling bank reserves (and less emphasis
on interest rales) injected an additional element of
uncertainty into financia l markets
Banks and other depo itory institutions were unable
to adjust freely the interest rates paid on deposits as
market rate changed. Member banks of the Federal
Reserve System were governed by Regulation Q, With
Similar regulations restricting other federally in ured
depository in tltutions. Consequently, as market interest rates rose above the respective maximum
allowable rates payable on deposits, deposi ory Institutions experienced outflows.
Before the introduction of money market mutual
funds (MMMFsl, Individuals had only limited ability to
invest in short-term. money market instruments. These
instruments not onl had high minimum denominations ($10,000 to $25.000 and higher) but also reqUired
a ophistication lacked by most small investor
Therefore, a natural limit on disintermediation existed
for a time.
The 1972 introduction of MMMFs changed all this
by radically altering the opportunities facing even the
small investor. The MMMF is a highly liquid a set pay109 competitive interest rates. It has a transaction
feature: investors can, in effect. place a sell order on
their shares b writing a check. Further, the MMMF
industry adopted an accounting procedure that
stabilizes the value of a share at $1 , making It a
perfeclly liquid investment for the individual.
In 1977, Merrill L nch bundled an MMMF with a
brokerage account and a line of credit supported b
the equit in the account The cash management account «(MAl gave its holder in tant overdraft
privileges.' With the growth of MMMFs and CMA , 40
years of regulations on bank liabilities began to
unravel
In response to the development of MMMFs, banks
offered " sweep" accoun at thEHetall level. Patterned
after established wholesale money managemen prac·
tices, such accounts automatically Invest all funds
over and above a predetermined minimum in money
market instruments.' Every night. the excess funds are
Economic Review I May 1985
"swept" into a portfolio of these financial instruments
to earn interest. Since the funds are no longer in demand deposit accounts, interest can be paid on the
funds and no reserves need be held against them .
The first major regulatory easing came with the
1978 introduction of the money market certificate.
This 6-month small time deposit paid an interest rate
equal to the 1SD-da Treasury bill discoun rate
established for the week in which the money market
certificate was Issued. While the minimum ba lance
was hefty-$10,OOO-banks circumvented this
minimum by lending depo itors some of the deposit.
The loan was ecured by the deposlL This subterfuge
was eventually institu lonalized b a regulation controlling the amount and terms of the loan. In the main.
however. the regulatory agencies were unwilling or
unable to deregulate deposits generally
Against this background, Congress passed the
Depository Institutions Deregulation and Monetary
Control Act in March 1980. In large part. this act
was deSigned to accelerate deposit deregulation
The Depository Institutions Deregula ion Committee
(DIDC)
as established to implement the orderly
phaseout of intere t rat ceilings. In this Splflt. the act
provided that all depository institutions could offer
NOW (negotiable order of withdrawal) accounts to individuals and nonprofi organizations.
Un atisfied with D IDC's progress in deregulating
deposits, Congress pa sed the Garn- St Germain
Depo itory Institutions Act of 1982. Among other
things. the act authorized all depository institutions to
offer a money market deposit account (MMDA)
" directly equivalent to and competitive with money
mark t mutual fund ." The MMDA was 0 ered effective December 1982. In the meantime, DIDC provided
for the introduction of Sup r NOWs-transactlon accounts for individuals paying market interest rates .
While the process of deresulat ng the nontran action components of the money tock IS n arl y complete. deregulated transaction accounts 'till make up
only a small percentage of total M1 (Chart 1). Clearly.
there is a sub tantial potential for further depos t
der gu lation. j
1 An overdraft is a loan made by th e bank to cover a check f o r which
there are insufficient ready funds in the account on wh ich it is
drawn.
2 These investments typically take the form of repur chase
agreements, in which the bank sells a share in a pool of securiti es to
the depositor with a promise to repur c hase In 24 hours . Technically,
the agreement is renewed each day, with the amount varying as
deposits and withdraw a ls are made in the account.
3. Ct . Thomas D. Simpson, Staff, Board of Governors of the Federal
Reserve System, "Impl ication s for Monetary Polley of Changes in
the FinanCial System " (Paper prepared for Brookings Panel on
Economic Activity. Wash ington, D.C, 5-6 April 1984), 11-12.
5
stance, the development of "debit cards" enables individuals to transfer cash without using currency. It
is perhaps fanciful to envision a world in which currency and coin are not used . But it is quite useful to
inquire into the institutional features of that world.
As we get closer and closer to it, presumably our
financial institutions will begin to look more and
more like those of the cashless, zero-balance financial system. 7
Economists and policymakers are searching for a
theory of a deregulated banking system. It turns
out, however, that this is one of the rare instances
of institutional change in which theory preceded
fact. Though a once-obscure literature on a fully
deregulated banking system has existed nearly 15
years, only the innovations of the last few years
made clear how fundamental are the theoretical
questions raised in it. I n the next section, I analyze
some of these issues.
Money in theory and practice
Money and prices. The classic reference in the
new literature on money is a 1970 article by Fischer
Black. 8 I n this seminal piece, Black imagines the
following:
a world in which commercial banks and other
financial institutions are free to offer checking accounts (and savings accounts) on any terms they
might want to set, and in which there are no
reserve requirements. Banks could pay interest on
demand deposits, and might not choose to
distinguish between demand deposits and time
deposits. Since there would be no reserve requirements, there would be no reason for Federal
Reserve open market operations.'
Black predicts that
In such a world, it would not be possible to give
any reasonable definition of the quantity of
money. The payments mechanism in such a world
would be very efficient, but money in the usual
sense would not exist.'o
At first blush, a world in which banks hold no
reserves and the public uses no currency would
seem to imply an infinite supply of money and an
infinite price level. Or, more generally, the quantity
of money and the price level would not be determinate. In Black's hypothetical world, however,
reserves and currency disappear because money as
a distinct asset no longer exists. The difficulty, as
we shall see, is not that the price level tends to
6
infinity but that money prices no longer exist.
In Black's model, there is complete depository
freedom. Depository institutions, which he terms
"banks," handle payments that are effected by
check or electronic transfer. Banks make loans, and
their income derives from the spread between the
rates paid on deposits and those charged on loans.
A loan is simply a negative bank balance-an
overdraft."
Black develops his analysis in the context of a
simple model of financial evolution, which begins
with a system of commodity money and ends in a
moneyless world . He hypothesizes that early in this
evolutionary process both real goods and financial
liabilities functioning as the means of payment are
priced in terms of a standard, abstract unit of
account.
At this point, it must be noted that Black's terminology is misleading. His use of "banks" is mentioned above. While we are accustomed to thinking
of banks as creators of money, there is no money in
Black's model. Yet he calls nonmoney financial
7 Grocery stores and gasoline stations are experimenting with
machines that electronically transfer funds for a purchase
from the vendee's to the vendor' s account when a debit card is
inserted One can easily imagine usage spreading until
cigarette and video-game machines would accept these cards .
A new generation of pay phones already accepts credit cards.
In 1986, French banks will begin introducing "smart cards,"
which can work in all these ways and more See " French Banks
Unveil Program to Circulate ' Smart' Credit Cards, " Wall Street
Journal, 6 March 1985, Southwest edition, sec 2. The use of
currency and coins will undoubtedly not disappear but might
become minimal before the end of this century.
8 . Fischer Black, " Banking and Interest Rates in a World Without
Money: The Effects of Uncontrolled Banking, " Journal of Bank
Research 1 (Autumn 1970): 8-20
9 . "Banking and Interest Rates," 9 Despite the supposition of
unregulated deposit creation, Black by no means assumes
laissez-faire in banking For instance, he specifies that " every
bank will be required to have capital equal to a certain fraction of its loans" (p 12).
10 "Banking and Interest Rates," 9 In addition, Black maintains
that "neither the quantity theory of money nor the liquidity
preference theory of money would be applicable" (p . 9).
" Traditional monetary theories will be inapplicable; in fact, it
will not be possible to define the quantity of money in meaningful terms" (p. 10)
11 . "Banking and Interest Rates," 10- 11 . Black's description of a
system of positive and negative balances is a virtual foretelling of the cash management account
Federal Reserve Bank of Dallas
assets created by banks "money." Finally, he refers
to the abstract unit of account as "dollars." So in a
world without banks as distinctive institutions,
without money and without currency, "banks"
create "money" denominated in "dollars." In addition, Black's evolutionary story begins with a commodity money, which really is money proper, but it
evolves into a financial asset that is no longer
money but is still called by the same name.
Despite the verbal confusion, Black's analysis
contains important insights. In what follows, I focus
on the analytical issues shorn of some of the
idiosyncrasies of the original article. I concentrate
on Black's article because it remains the fundamental contribution in the literature.
Goods trade for financial assets, both measured
by an abstract unit of account. A mutual fund of
common stock is the paradigm financial asset serving as a means of payment. Since the fund's value
fluctuates daily, the "money" price of goods
(measured in the unit of account) must constantly
be recomputed. I n other words, we are in a barter
world. Borrowing and lending of financial assets are
now introduced. Banks emerge as administrators
and guarantors (for a fee) of loans. I n the final
stage, banks also provide a means of payment. At
this point, banks resemble the financial intermediaries with which we are familiar. Black concludes
that
In none of these .. worlds was there any role for a
central bank. And the only effect that the financial sector had on the real sector was that as we
go to successively more efficient means of payment, we reduce the cost of making payments and
release real resources for other uses. In none of
these worlds was there any mechanism that would
cause uncontrolled inflation in the absence of a
central bank 12
Black's denial of a role for central banks follows
from his position that central banks exist to enforce
unnecessary regulations, regulations absent by
assumption from his model. His second conclusion
is a kind of neutrality proposition, one that beomes
the focus of Eugene Fama's development of Black's
framework.'3 Fama concludes that
A competitive banking sector is largely a passive
participant in the determination of a general
equilibrium, with no special control over prices or
real activity, which in turn means that there is
nothing in the economics of this sector that makes
it a special candidate for government control."
Economic Review I May 1985
In other words, banks merely finance real activities,
which are assumed to be invariant to the financing
process.
Toward which world are we evolving? Is it
Black's, in which money as we know it disappears?
Or is it one in which the banking system is completely unconstrained in its creation of conventional
deposits? If the latter, then the creation of additional money is virtually cost less and the classical
theory of money would apply.
Cost of production was the Classical explanation
of the long-run value of money (as well as of
everything else) under the rules of the gold standard, and Classicists did not consider the substitution of discretionary management of the money
supply. Classical thinkers saw no way, in the long
run, to maintain the value of a costless currency .
No one yet has proved them wrong on this point,
and we appear presently to be-however reluctantly- in the process of proving them right."
If the classical theory is correct, then some limitation on the ability of banks to create liabilities is
necessary. The limitation may be natural, such as a
resource cost under a commodity standard, or it
may be artificial, as that imposed by central banks
under a fiduciary or fiat standard. Without such a
limit, however, there is no anchor for nominal
values in an economic system . This issue is the
focus, directly or indirectly, of most of the rest of
this article.
Base money. If it were true that a largely
unregulated banking system would be self-limiting
in its production of liabilities, then the classical
case for regulating banks would be undermined. Further, if we are evolving toward a system of zerobalance accounts, then central banks will no longer
have anything to control. They would be both
superfluous and ineffectual.
In analyzing Black's contention about the stability of unregulated banks, I ask three questions of his
12. "Banking and Interest Rates," 15.
13
Eugene F Fama, " Banking in the Theory of Finance," Journal
of Monetary Economics 6 (january 1980): 39-57.
14 " Banking in the Theory of Finance. " 47
15
Will E Mason, "Winners and Losers: Some Paradoxes in
Monetary History Resolved and Some Lessons Unlearned,"
History of Political Economy 9 (Winter 1977): 478
7
model.
1. How are "dollar" prices determined?
2. In what are loans made?
3. How do banks settle among themselves?
Amazingly, Black's article provides no answer to
any of these questions. Answering the first question
is obviously important because Black incorporates a
unit of account in his model. For example, the value
of a deposit is its "dollar" price divided by the
"dollar" price of a basket of goods. "Dollars" are,
however, an entirely abstract unit of account in this
system; "dollars" do not exchange against goods.
Consequently, there is no market in which the
"dollar" price of anything is established. Black's
dollar prices are pure accounting prices; thus,
changes in them have no economic or market function. They neither clear excess demands nor result in
resource reallocation. 16
As to the second question, it is never clear what
assets are being borrowed. The reader is only told
that "the borrower writes a personal note and gives
it to the lender in exchange for certain assets."17
Black addresses but does not answer the third
question when he observes, "There will be an active
market in inter-bank funds."18 To settle debits and
credits of banks, clearinghouses must be able to
transfer an asset accepted as a final means of payment. For instance, is the transfer of assets like that
of gold or silver on a specie standard? That is, does
the transfer force an expansion or contraction of a
bank's operations? If not, what limitations are
there? 19
16. " Empirical observation alone could not detect the. say.
doubling of the accounting prices of all goods. In the market
place we can observe only the manifestations of money. and
hence real, prices" (Don Patinkin; Money, Interest, and Prices:
An Integration of Monetary and Value Theory. 2d ed. [New
York: Harper & Row. 1965). 16). Cf. White. "Competitive
Payments Systems." 700.
17
Black, "Banking and Interest Rates," 15
18 " Banking and I nterest Rates," 11 . Black never explains what
constitutes " inter-bank funds"
19. To answer these questions, one must specify the operative
monetary regime. On the importance of regimes in monetary
analysis, see, in the September 1984 Economic Review of the
Federal Reserve Bank of Dallas, Gerald P. O'Driscoll, Jr.,
" Expectations and Monetary Regimes, " 1-11 ; John H Wood,
"The Search for a Monetary Policy Rule in an Uncertain
World," 13-23; and W . Michael Cox, "What Is the Rule for
Financing Public Debt?" 25-31
8
In principle, individuals could transact with a
pure system of debits and credits operated by
banks. Banks cannot do so, however, because the
liability of one bank does not constitute payment to
another bank but the promise of payment. The final
means of payment cannot be the liability of another
bank but must come from outside the banking
system itself .
In the early stages of banking development,
specie alone settled what a bank owed on net to
other institutions. Moreover, the settlement process
involved pairwise settlements between institutions.
Clearinghouses arose to facilitate settlement and
process interbank transfers, thereby economizing on
transaction costs. Each member needed to settle
with only one institution-the clearinghouse-and
the settlement was for only the net amount of
adverse clearings with all other members. Further,
clearinghouses issued specie certificates to
economize on specie transfers . Finally, clearinghouses themselves eventually issued certificates
convertible into assets; these certificates at times
circulated alongside legal-tender currency.20
At all times, what constituted final payment for
commercial banks was neither the liability of
another bank nor an asset whose quantity was determined by another bank. In normal times, it was base
money: specie plus currency and, with the advent of
central banks, deposits at the central bank . In
modern parlance, it was and is base money.
I am making two points . First, the existence of
reserves or base money is inherent in the very concept of competitive banking. And second, the quantitative limit on reserves restricts the ability of
banks to create liabilities. This limit is not what
Black and Fama suppose, but it reinforces their
result that competitive banking can be a stable
system. 21
As a corollary, there must be a producer or supplier of reserves. The supplier need not be either a
20. On private clearinghouses, see Gary Gorton, " Private
Clearinghouses and the Origins of Central Banking,"
Business Review, Federal Reserve Bank of Philadelphia.
January/February 1984. 3-12; d White, "Competitive
Payments Systems," 705-6 .
21 . Black, in fact, argues that an autonomous limitation on bank
reserves would be inefficient ("Banking and Interest Rates,"
17-18); d . Fama, "Banking in the Theory of Finance," 39-40,
47-48.
Federal Reserve Bank of Dallas
central bank or even a financial institution. I n a
world of commodity money, reserves are a produced goodY
The remaining issue is to establish a role for currency . Here the argument is more pragmatic and
historical. First, with all the revolutions in the
payments mechanism, there is no documented
tendency for currency to be displaced. New
depository accounts (including nonbank accounts)
seem to be substitutes for existing types of accounts, not for currency . Debit cards still have a
very small share of the payments market. Moreover,
debit cards thus far seem to be better substitutes
for credit cards and checks than for currency. And
electronic funds transfers, which typically involve
very large amounts, are almost surely substitutes for
transfers by check, not currency .
It is a commonplace that the use of currency and
coin is more economical for small transactions . In
the economist's parlance, their use economizes on
transaction costs. Moreover, every other means of
transacting incurs costs not incurred when using
cash. For instance, a check may be drawn fraudulently, or there may for other reasons be insufficient
funds to pay the check upon presentment. The institution on which the check is drawn may even be
insolvent. Further, any computer transfer involves
potential loss through fraud or system failure. Cash
alone avoids these costs . The mere advance of
technology does not fundamentally alter the
benefit-cost calculus for using cash.
The demand for precautionary holdings of currency has been even more invariant to institutional
and technological change. Dale Osborne has
reported on estimates of the size of these stocks:
The public holds about $150 billion in currency,
more than $600 for each man, woman, and child
in the United States .... less than a third of this
staggering sum appears to be in use as exchange
media ....[Mjost of it is apparently hoarded against
civil disturbances, natural disasters, and nasty
divorces ..Since 1890, estimated hoards of all
denominations have remained an almost constant
3 percent of yearly GNPY
22 . On this point, see Gorton, "Private Clearinghouses and the
Origins of Central Banking "
23. Dale K Osborne, "What Is Money Today?" Economic Review,
Federal Reserve Bank of Dallas, January 1985, 13. I n the
original article the 1890 figure was misprinted as "1980" The
Economic Review I May 1985
Unless these fundamental determinants of the
precautionary demand for currency change, it is
highly unlikely that its role in our monetary system
will change. Over the same period, with all the
technological and institutional changes taking place,
currency in circulation has declined only one-third
(as a percentage of GNP).
Finally, largely unregulated banking systems have
existed in recent times .24 I n these "free banking"
systems, currency was produced privately, banks
could pay interest on demand deposits, and financial innovation occurred. 25 Moreover, the use of
non-interest-bearing currency flourished. The tentative conclusion, then, must be that circulating
money and a system of bank reserves will be part of
a deregulated financial system .
The quantity of money. Black is adamant that the
quantity of money in a competitive banking system
could not be meaningfully defined. 26 Being able to
measure empirically the quantity of money is not
the same as being capable of defining money; and
the definition of money is not the same as the concept of money.
The concept of money is ambiguous if not vague
stability of currency hoards over this period suggests that
neither tax avoidance nor illegal activities can account for
currency usage
24 There is a growing historiography of "free banking" systems
The two most well-known free banking systems were the Scottish, in the 18th and 19th centuries, and the American, which
began in the late Jacksonian period and ended with the Banking Act of 1863 On the Scottish system, see Lawrence H.
White, Free Banking in Britain: Theory, Experience, and Debate,
1800-1845 (Cambridge: Cambridge University Press, 1984) For
some recent work on the American system, see Arthur J
Rolnick and Warren E. Weber, "Free Banking, Wildcat Banking, and Shinplasters, " Federal Reserve Bank of Minneapolis
Quarterly Review, Fall 1982, 10-19; the modern classic work is
Hugh Rockoff, " The Free Banking Era: A Reexamination, "
Journal of Money, Credit, and Banking 6 (May 1974): 141-67
25
In the United States, for instance, both the thrift and life insurance industries arose in the 19th century to compete with
commercial banks And, of course, deposits supplanted currency as the primary means of payment For the modern era,
see John H Wood, "Familiar Developments in Bank Loan
Markets," Economic Review, Federal Reserve Bank of Dallas,
November 1983,1-13 Wood maintains, "The present American
financial system was essentially in place by 1880" (p 2)
26 Sometimes, however, it appears that money can be defined b.u t
its quantity is not determinate (" Banking and Interest Rates,"
15,19)
9
in this literature. It is not entirely surprising, then,
that no meaningful definition of money would be
apparent. And if money cannot be defined, it cannot be measured. Causation, however, does not
necessarily run the other way: the mere fact that we
cannot quantify our definition does not imply that
we do not have one. Further, identifying the realworld counterpart of our concept of money is a
process logically independent of developing the
concept itself.27
It is not clear that measuring the quantity of
money in the Black-Fama competitive banking
system would be impossible, nor would it seem impossible to give a "meaningful" definition of money.
For instance, there is no insurmountable obstacle to
measuring a commodity money. Even in this case,
however, Black complains that "a commodity used
as means of payment also has other uses, and it
may not be clear when it is to be counted as part of
the money supply, and when it is to be counted as
involved in one of its other uses." 28
First, it must be noted that if this argument is accepted, there never has been a meaningful definition of money. Second, the problem here is surely
not one of definition but of measurement. The
classical economists, for instance, distinguish between specie in circulation and gold and silver
devoted to nonmonetary uses. If measurement were
sometimes difficult, this problem did not negate the
important distinction that defined or delimited
money and nonmoney.29 Third, as the second point
suggests, money always has close substitutes. So,
too, do most other goods. Oranges and grapefruit
are close substitutes at the margin. This does not,
however, preclude distinguishing them .
At times, Black seems to suggest that money
cannot be meaningfully defined because it is endogenously produced by private suppliers.30 Most
money supplies have been at least partly endogenous; certainly commodity money supplies
were . 31 It is true that an endogenous money supply
is subject to continuous variation . But at each moment it has a determinate size.
Not only, then, is the disappearance of outside
money unlikely, but a deregulated banking system
introduces no new analytical problems of defining
money.
Interest on money
Black, Fama, and others have seemingly taken for
granted that a deregulated banking system would or
should evolve into a world without money. Fama
associates financial sophistication and freedom in
banking with a society
so advanced that terms like money, medium of exchange, means of payment, and temporary abode
of purchasing power have long ago fallen from its
vocabulary, and all written accounts of the ancient "monetary age" were long ago recycled as
part of an ecology movement. J2
The central-banking literature explicitly links the
monetary control issue to the payment of interest
on money. If interest is paid on money, will money
blend in with other assets?
The answer must surely be no. Again, monetary
history and economic theory both support this
conclusion. Limitations on the payment of interest
are of comparatively modern origin . Moreover,
developed countries similar to the United States (for
example, Britain and Canada) never implemented
such a highly restricted system and removed the
deposit regulations they had sooner. There has been
neither a noticeable lack of monetary control
(vis-a-vis the United States) nor signs of the imminent disappearance of money as we have known it.
Even more telling, of course, is the fact that the
Federal Reserve System operated between 1914 and
1934 without the relevant controls on deposits.
Whether or not the Fed always consciously exercised control over the money supply, it never lacked
31
27 On this point, see Will E. Mason, "The Empirical Definition of
Money: A Critique," Economic Inquiry 14 (December 1976):
525-36.
26. "Banking and Interest Rates, " 15.
29. Mason, "Winners and losers," 461.
30. Cf. White, "Competitive Payments Systems," 709-11
10
" We could not take exogeneity as a defining characteristic of
money. That our money is exogenous ... would be a theorem,
true today but not true at all times. Money was not exogenous
when it was gold , [T]he acceptance of exogeneity as either a
defining characteristic or a theorem true of all future moneys
would mean that a system of privately produced money is a
contradiction in terms However hard the essentials of such a
system may be to grasp, it does not appear to be selfcontradictory" (Osborne, "What Is Money Today?" 13-14)
32 "Banking in the Theory of Finance, " 55 .
Federal Reserve Bank of Dallas
the control.
The theoretical argument goes to the concept of
money itself. Certainly there are properties of
money illuminated by analyzing it as just another
asset in a theory of finance. The approach almost
inevitably overlooks, however, the insights gleaned
from treating money as a unique good with special
properties These properties explain why, even in
competitive banking systems, money yields less interest than other highly liquid financial assets.
The more traditional theory of money emphasizes
that money itself is the product of an evolutionary
(unregulated) process. l l Money has certain properties that are the product of neither regu lation nor
historical accident but reflect inherent characteristics valued in the marketplace. Prominent among
these characteristics is money's liquidity.
Money is not simply highly liquid, as are many
other assets, but is perfectly liquid. It trades in
every market and need never be sold at a discount.
In other words, money is the good circulating
routinely as the medium of exchange and the entire stock of which can be spent ("sold")
simultaneously.14
Experience of recent years indicates that even the
most highly liquid, short-dated nonmonetary asset is
subject to some price risk. It is for this reason that
individuals are willing to incur substantial opportunity costs in holding money balances. Throughout
monetary history, in highly regulated and substan-
33. The classic presentation of this theory is in Carl Menger, "On
the Origin of Money," trans . Caroline A Foley, Economic
Journal 2 (June 1892): 239-55 . For a development of Menger' s
analysis, see Gerald P O'Driscol1. Ir , and Mario J. Rizzo, The
Economics of Time and Ignorance (Oxford and New York:
Basil Blackwell, 1985), 191-98; cf White, "Competitive
Payments Systems," 703-6 Robert A Jones presents a
modern development of part of Menger's thesis in "The
Origin and Development of Media of Exchange," Journal of
Political Economy 84 (August 1976, pt 1): 757-75; for a recent
analysis of liquidity, see Robert A Jones and Joseph M
Ostroy, "Flexibility and Uncertainty," Review of Economic
Studies 51 (January 1984): 13-32 Also relevant is SHerbert
Frankel, Two Philosophies of Money: The Conflict of Trust
and Authority (New York: St Martin's Press, 1977)
34 Osborne identifies these two features as essential properties
of a medium of exchange: "both the simultaneity and routinecirculation tests are essential in all systems; the former to
identify a stock instead of velocity or a mixture of stock and
velocity, and the latter to ensure that the stock is money"
("What Is Money Today? " 2)
Economic Review I May 1985
tially unregulated monetary systems, people have
demanded absolutely liquid assets. This demand has
been evidenced in the continued large holdings of
non-interest-bearing currency .
Money yields less than a market interest rate
because it also yields nonpecuniary services. Indeed,
it is money's nonpecuniary services that distinguish
it from nonmoney financial assets, which yield only
a pecuniary returnY
Put another way, if money were to yield market
interest rates, then its total return would be higher
than for any other asset. For money would yield
both a pecuniary return equal to that on highly
secure, short-dated assets and a nonpecuniary return
of "liquidity" services.
The last point sheds light on the recent controversy surrounding the "legal restrictions theory of
money."J& This theory focuses on the apparent
paradox that at one and the same time, individuals
hold both non-interest-bearing government currency
and interest-bearing government securities. Advocates of the theory identify legal restrictions on
the use of interest-bearing government securities for
media of exchange as the source of the demand for
non-interest-bearing currency . The following stark
prediction reveals the affinity between the legal
restrictions theory and the Black-Fama approach:
Laissez-faire means the absence of legal restrictions that tend, among other things, to enhance
the demand for a government's currency. Thus, the
imposition of laissez-faire would almost certainly
reduce the demand for government currency. It
could even reduce it to zero. A zero demand for a
government's currency should be interpreted as
the abandonment of one monetary unit in favor of
another-for example, the abandonment of the
dollar in favor of one ounce of gold. Thus, my
prediction of the effects of imposing laissez-faire
35 See Benjamin Klein, " The Competitive Supply of Money,"
Journal of Money, Credit, and Banking 6 (November 1974):
425.
36. See Neil Wallace, "A Legal Restrictions Theory of the Demand for 'Money' and the Role of Monetary Policy," Federal
Reserve Bank of Minneapolis Quarterly Review, Winter 1983,
1-7; John Bryant and Neil Wallace, " The Inefficiency of
Interest-bearing National Debt," Journal of Political Economy
87 (April 1979): 365-81; and John Bryant, "Analyzing Deficit
Finance in a Regime of Unbacked Government Paper,"
Economic Review, Federal Reserve Bank of Dallas, January
1985, 17-27
11
takes the form of an either/or statement: either
nominal interest rates go to zero or existing
government currency becomes worthless. 37
The argument presented here denies the thesis of
the legal restrictions theory by questioning its
assumption of "a paradoxical pattern of returns
among assets."18 Non-interest-bearing currency has
an implicit yield of convenience and liquidity services; its implicit rate of return will, in equilibrium,
equal the explicit pecuniary rate of return of highly
liquid, interest-bearing government assets.
Some monetary assets may yield explicit interest
No asset yielding interest at the competitive rate,
however, could routinely circulate as the medium of
exchange; were it to do so, it would in fact earn a
supracompetitive return (as emphasized earlier).
More to the point, assets yielding explicit interest
(a pecuniary return) at the competitive rate could
never provide absolute liquidity. These assets would
be subject to a price risk not heretofore associated
with routinely circulating media of exchange. 19
To put the matter in the same stark terms as does
Professor Wallace, my prediction is that under
laissez-faire, there would be no tendency for either
the interest rate to go to zero or non-interest-bearing
currency to become worthless .
It is possible to trivialize the difference between
money and nonmoney financial assets by suggesting
that the use of money merely saves some transaction costs . These "transaction costs" are, however,
the prohibitive costs of barter that are avoided in a
monetary economy. Neither the legal restrictions
theory nor theorists in the Black-Fama tradition
have addressed how the costs can be avoided in
their system.'o
Conclusion
This article began by considering the problem of
central bank control of the money supply in a partially deregulated environment. It then analyzed the
stability of a monetary system in a world of substantially unregulated financial institutions.
One theoretical issue arose in the two sets of
literature examined in this article: Will the payment
of interest on money lead to either loss of monetary
control or the disappearance of money itself? The
answer in both cases is no. The answer is justified
by theoretical, institutional, and historical analysis.
Deposit deregulation will not result in an unstable
banking system. It will evolve in ways that we
find unsettling and unfamiliar and that present
challenges to the monetary authorities. For instance,
central banks can no longer rely on disintermediation and credit crunches to influence monetary
growth. We are not, however, in totally uncharted
waters." Neither bank reserves nor currency is likely
to disappear in the near future . Moreover, there are
inherent limits on the ability of private banks to
create liabilities .
Deregulation may even highlight the central
bank's role in controlling the growth of base money
under the present monetary regime, which embodies
monetary targets. It is in this way that a nominal
anchor is provided in a system of pure fiduciary
currency.
Laissez-Faire Approach to Monetary Stability," Journal of
Money, Credit, and Banking 15 [August 1983): 307) They do
not distinguish their presumably sophisticated barter system
from " crude" barter, so we must presume it faces all the difficulties of any barter system By way of solution, they merely
predict, " The profit motive will surely lead competing private
firms to offer convenient methods of payment" (p. 307). In
this literature, these payments systems are typically equitybased (like MMMFs) White has pointed out that it is improbable that shares of an equity fund would ever become a
generally accepted medium of exchange ("Competitive
Payments Systems," 710).
37. Wallace. "A Legal Restrictions Theory, " 4
38
Wallace, " A Legal Restrictions Theory," 1
39 Much the same point can be made about the equity-based
medium of exchange in the analysis of Black and Fama (see
note 40). Some categorize their theories as instances of the
legal restrictions theory Wallace refers to Fama (and Robert
E. Hall) as providing "other discussions of the legal restrictions theory" ("A Legal Restrictions Theory," 1 n, 2)
40 The barter issue is sometimes acknowledged but never addressed in the literature For instance. Robert L Greenfield
and Leland B Yeager admit that although their proposed
payments system, which builds on the work of Black, Fama,
and Robert Hall, " is barter . , it is not crude barter" (" A
12
41
Though Wood focuses on bank assets , particularly loans, this
is the main point of "Familiar Developments in Bank Loan
Markets "
Federal Reserve Bank of Dallas
Inflation and Permanent
Government Debt
W. Michael Cox
Senior Economist
Federal Reserve Bank of Dallas
Interest-bearing public bonds, which correspond to
no government capital formation and which no taxpayer rationally expects to have to help retire ... ,
have effects similar to those of the noninterest
IOUs we call M [money] .
- Paul A. Samuelson, Economics
At the start of the Second World War, the market
value of privately held gross federal debt in the
United States totaled $56 billion. During the war the
average annual budget deficit jumped sharply to
$42 billion, and outstanding government debt
climbed to $220 billion. It is reasonable to presume
that individuals viewed debt issued during the war
as temporary, intended for future retirement.
Following the war, however, significant budget
surpluses did not arise to finance net debt retirement. To the contrary, budget deficits continued
and have grown in magnitude. Privately held gross
federal debt now stands at more than $1,300 billion .
It is not surprising, then, that attention in recent
years has shifted toward matters involving public
debt. Key among these is the question of govern-
The views expressed are those of the author and do not necessarily
reflect the positions of the Federal Reserve Bank of Dallas or the
Federal Reserve System.
Economic Review I May 1985
ment's long-term debt intentions. Is government
debt temporary and implicitly backed by future
taxes? Or has the government adopted a policy of
permanent government debt, with no intention of
future debt retirement? The question is crucial
because a regime of permanent government debt is
fundamentally different from a regime of temporary
government debt.
This article focuses on the inflationary implications of permanent government debt. Specific attention is paid to two propositions: the Ricardian
Equivalence Theorem-which claims that taxatio.n
and debt finance are equivalent-and the Quantity
Theory of Money-which claims that the price level
is related to the stock of money only (and not
government debt). It is argued that both propositions are valid in a regime of temporary government
debt but each is invalid with permanent government
debt. In a regime of permanent government debt,
taxation and debt finance are not equivalent, and a
switch from current taxation to debt finance has
inflationary effects much the same as those of a
switch to currency finance.
Tests for the inflationary effects of government
debt in the United States show that over the
1950-84 period the price level was determined as
13
much by the stock of privately held government
debt as by base money. The results of Granger-Sims
causality tests relating the price level to both
money and debt, as opposed to money only or debt
only, indicate that debt is as important an explanatory variable as base money over this period .
Considering the frequency of budget deficits in recent years (23 of the past 24 years) and their comparatively huge magnitude, these results indicate
that government debt is viewed as permanent, with
the same inflationary implications as money.
The Ricardian Equivalence Theorem
The proposition that government bonds rolled over
forever are like money, with potentially the same inflationary implications as money, is not new. 1 Strictly speaking, the argument is as old as the debate on
the Ricardian Equivalence Theorem, which states
that debt issuance on the part of government is
equivalent to current taxation .2 According to this
theorem, the choice of debt finance as opposed to
current taxation does not affect anything - not even
the price level- because the issuance of government debt implies equivalent (in the present
discounted-value sense) future tax I iabil ities. J In
short, the theorem views government bonds as
backed by taxes.
The classic exposition of this hypothesis is as
follows:
Besides valuing government expenditures as income, households may regard deficit financing as
equivalent to taxation . The issue of a bond by the
For a discussion of this proposition see, for example, Paul A
Samuelson, Economics, 8th ed (New York: McGraw-Hili Book
Company, 1970), 326-27; Carl F Christ, " Patinkin on Money,
Interest, and Prices," Journal of Political Economy 65 (August
1957): 347-54; Don Patinkin, Money, Interest, and Prices: An
Integration of Monetary and Value Theory , 2d ed (New York :
Harper & Row, 1965); Marco Antonio Campos Martins, " A
Nominal Theory of the Nominal Rate of Interest and the Price
Level," Journal of Political Economy 88 (February 1980):
174-85; John Bryant and Neil Wallace, " The Inefficiency of
Interest-bearing National Debt," Journal of Political Economy
87 (April 1979): 365 - 81 ; and Preston J. Miller, " Deficit Policies,
Deficit Fallacies," Federal Reserve Bank of Minneapolis
Quarterly Review, Summer 1980,2-4
2 The writing of David Ricardo from which the Equivalence
Theorem has been drawn may be found in The Works and Correspondence of David Ricardo, ed, Piero Sraffa (Cambridge:
Cambridge University Press, 1951), 4:185-87.
3. There is an effect on private savings, as outlined later
14
government to finance expenditures leads to
future interest payments and possible ultimate
repayment of principal. That is, it implies future
taxes that would not be necessary if the expenditures were financed by current taxation . If a
typical household were to save the entire amount
that was made available to it by a switch from
current taxation to deficit financing, the interest
on the saving would meet the future tax charges to
pay interest on the government bonds; the amount
saved would be available to meet possible future
taxes imposed to repay the principal of the government bonds.'
The economic implications of the Equivalence
Theorem are clearly important and are well
recognized . If debt finance and current taxation are
equivalent, then "government debt will be absorbed
without any real effects on the economy."; Switching from current taxation to debt issuance has no
aggregate demand, interest rate, or even price effects when the two financing means are equivalent.
In recent years the assumptions underlying the
Equivalence Theorem have been attacked by many
parties and defended by many others , Supporters
of the Equivalence Theorem contend that (1) individuals are rational in the sense of fully discounting their future tax liabilities and (2) government
debt is temporary and fully backed by taxes. 6 If,
4
Martin J Bailey, National Income and the Price Level: A Study
in Macroeconomic Theory, 2d ed (New York: McGraw-Hili
Book Company, 1971), 156,
5 See Roger C Kormendi, " Government Debt, Government
Spending, and Private Sector Behavior, " American Economic
Review 73 (December 1983): 994-1010
6 Two other points of contention regarding the Equivalence
Theorem are government capital formation and intergenerational bequest motives With regard to the first one, the backing for public debt need not be taxes but could be profitable
investment goods purchased with the proceeds from bond
sales Since there is little evidence, however, that the government is ~ngaging in massive and highly profitable investment,
this matter is not seriously considered here With regard to the
second point, Robert J, Barro demonstrates that " finite lives
will not be relevant to the capitalization of future tax
liabilities so long as current generations are connected to
future generations by a chain of operative intergenerational
transfers" (" Are Government Bonds Net Wealth? " Journal of
Political Economy 82 [November/December 1974]: 1095), He
assumes, however, that "the government may not reissue the
bonds when they come due and that the principal is paid off
at the beginning of the next period by an additional lump-sum
tax levy" (p, 1102), That is, Barra explicitly focuses on temporary government debt and does not consider permanent debt
Feder;a\ Reserve B;ank of O;al\;as
however, individuals do not accurately take account
of their future tax liabilities or if the government
follows a policy of permanent debt finance, then
the theorem is invalid and government debt is not
neutral.
In the past, it was largely assumed that government debt is in fact backed by taxes, and attention
was focused on the question of whether individuals
rationally perceive that backing. It was argued that
if individuals are rational, they will recognize that
government bonds are equivalent to future taxes.
If the household has a definite consumption plan
for the future, and if it knows the future tax effects of the shift from current taxation to deficit
financing, then it will save all the disposable income it gets from the switch from current taxes to
a bond issue. 7
Furthermore,
If future tax liabilities implicit in deficit financing
are accurately foreseen, ... [then] the behavior of the
community will be exactly the same as if the
budget were continuously balanced. 8
The term "rational" has commonly been used to
describe individuals who are "not confused or misinformed" but, on the basis of complete information,
consistently discount their future tax liabilities. 9
Note that government bonds are not net nominal
wealth in this case; they are neutral.
The assumption that government bonds are
perceived as net wealth by the private sector is
crucial in demonstrating real effects of shifts in
the stock of public debt In particular, the standard effects of "expansionary" fiscal policy on aggregate demand hinge on this assumption . Government bonds will be perceived as net wealth only if
their value exceeds the capitalized value of the
implied stream of future tax liabilities 10
It is important to point out that there are two
7
Bailey, National Income and the Price Level, 156
8
Bailey, National Income and the Price Level. 158 The government budget equation Bailey is implicitly using here is that of
the Treasury alone, not the consolidated budget equation of
the Treasury and the Federal Reserve See note 28 for a discussion of the consolidated budget equation, and see the section
on temporary and permanent government debt for a discussion
of the budget equation distinction between permanent government debt and permanent deficit finance
9
See Bailey, National Income and the Price Level. 157
10 Barro, "Are Government Bonds Net Wealth?" 1095
Economic Review I May 1985
very different problems regarding the capitalization
of future tax liabilities. One is the issue of rationality, which questions whether individuals fully
perceive their future tax liabilities. The other matter
is government's debt policy, which is concerned
with what in fact those tax liabilities are . If individuals are rational and if government debt is temporary, then "there is no persuasive theoretical case
for treating government debt...as a net component
of perceived household wealth."" If, however, individuals are "confused or misinformed" or if
government debt is permanent, then government
debt will not be perceived as backed by taxes, and
government bonds will be net wealth since their
value exceeds the capitalized value of future tax
liabilities.
Temporary and permanent government debt
Previous studies have accepted the notion that
government debt is backed by taxes and have questioned whether individuals are rational enough to
perceive that backing . This section does the opposite by accepting individual rationality and questioning government's long-term debt intentions.
Focus is on temporary and permanent government
debt. '2 It is argued that the Ricardian Equivalence
Theorem is valid in a regime of temporary government debt, provided individuals are rational, but is
not valid in a regime of permanent government debt
11
See Barro, "Are Government Bonds Net Wealth?" 1116
12
In the earlier literature and in this article, the matter of temporary versus permanent is considered only for government
debt (and not also for money) Recently, however, Bryant and
Wallace, "The Inefficiency of Interest-bearing Natonal Debt";
Bruce D. Smith, "Money and Inflation in Colonial Massachusetts," Federal Reserve Bank of Minneapolis Quarterly
Review, Winter 1984,1-14; and others have emphasized the
importance of permanence as applied to both types of
government paper These models focus on government
pol icies regarding the fiscal deficit or surplus, rather than
specifically on government policies regarding its debt These
two types of pol icies are clearly not the same since budget
deficits can be financed with the issuance of either debt or
currency, and a regime of temporary deficit finance requires
that the budget be balanced in the long run whereas temporary government debt requires only that debt issued during
deficit be retired during surplus. Notice in the Patinkin model
that money is viewed as always a component of private
nominal wealth This view may be reconciled in light of the
recent work by treating all money as issued permanently;
then, only the permanence of government debt is in question
15
Box A
A Model of Temporary and Permanent Government Debt
Market
Condition for equilibrium
= N(w),
Labor services . .
Q(w)
Commodities
F(y, r, ~ + ~B) = y,
Bonds
B(Y, R'pM + kB)
= kB
p
p'
Money.
L(Y, R'pM + kB)
= p'
M
p
Fy
> 0,
Fr
< 0,
Fv
> O.
Definitions
w = real wage rate .
Y
real income.
r = real interest rate.
M
= stock of cash balances (base money).
p = price level.
k = fraction of outstanding government debt that is viewed by the
private sector as not backed by taxes .
B
M
+
kB
market value of privately held government debt
nominal wealth of the private sector
M
v == -
p
kB
+ -p
= real wealth of the private sector.
R = nominal interest rate.
NOTE : This m odel may be found in Don Patinkin, Mone y, Interest, and Prices , 2d ed (New York :
Harper & Row, 1965), 289-90
16
Federal Reserve Bank of Dallas
regardless of the rationality issue.'3 The implications
of both temporary and permanent government debt
for determination of net wealth, aggregate demand,
the interest rate, and the price level are outlined."
Temporary government debt is defined as a
regime in which government debt sold during a
budget deficit is later retired (principal plus interest)
during surplus.' 5 Under this regime the issuance of
U .S. Treasury paper to finance a current budget
deficit requires a future budget surplus so that the
principal may be retired and the interest paid.
Treasury debt outstanding under this regime is
backed government paper. It is backed by taxes.
13. In this section and throughout the article, it is assumed that
the economy in question is closed In an open economy,
domestic budget deficits may be financed through external
borrowing, without required reductions in internal private
spending. This case is clearly reasonable when government
debt is issued temporarily, but a regime of permanent government debt requires that the foreign economy as a whole run
permanent current account surpluses
14. Gerald P. O ' Driscoll, Jr, advocates that " Ricardo in fact
denied that taxation and public debt are equivalent. The
'Ricardian Equivalence Theorem' is, consequently, a
misnomer, largely because Ricardo was not a Ricardian on
this issue Rather, Ricardo enunciated a nonequivalence
theorem" ("The Ricardian Nonequivalence Theorem, " Journal
of Political Economy 85 [February 1977]: 207) As O ' Driscoll
points out (p 209), Ricardo in fact said, " This argument of
charging posterity with the interest of our debt, or of relieving
them from a portion of such interest, is often used by otherwise well informed people, but we confess we see no weight
in it" (The Works and Correspondence of David Ricardo
4:187)
15
Much of the early literature considered only models of temporary deficit finance or temporary government debt. Recently, however, a number of models have come forth to consider
either permanent deficit finance or permanent government
debt Among those considering permanent deficit finance are
John Bryant, "Analyzing Deficit Financ e in a Regime of Unbacked Government Paper, " Economic Review, Federal
Reserve Bank of Dallas, January 1985,17-27; and John Bryant
and Neil Wallace, " A Price Discrimination Analysis of
Monetary Policy, " Review of Economic Studies 51 (April
1984): 279-88 See also Preston J Miller, " Higher Deficit
Policies Lead to Higher Inflation," Federal Reserve 8ank of
Minneapolis Quarterly Review, Winter 1983,8-19 Among
those considering permanent government debt are Samuelson,
Economics , 326-27; Patinkin, Money, Interest. and Prices ;
Bryant and Wallace, " A Price Discrimination Analysis of
Monetary Policy"; and Neil Wallace, " A Legal Restrictions
Theory of the Demand for 'Money' and the Role of Monetary
Policy," Federal Reserve 8ank of Minneapolis Quarterly
Review , Winter 1983,1-7
Economic Review I May 1985
Permanent government debt is defined as a
regime in which government debt sold during a
budget deficit is not retired later, nor is the interest
financed with tax revenues from a budget surplus.
Principal and interest are entirely deficit-financedeither rolled over or financed by printing currency .' 6
Government debt issued under this regime is not
backed by taxes .
They [government bonds]. like currency, are pieces
of paper backed by nothing - not by tangible
assets, not by future taxes ... [T]hey are valued fiat
paper that adds to the nominal wealth of the
private sectoL 17
16. It is also important to demonstrate that a regime of perma·
nent government debt is feasible in the long run In recent
years, a number of studies have specifically considered the
feasibility matter, and several models that have been
developed show how government can run continuous budget
deficits or how government debt can exist permanently.
Rather than reviewing these models here or attempting to
determine all possible scenarios under which permanent
government debt is achievable, the feasibility of permanent
government debt will be demonstrated by outlining one simple possible scenario_
Specifically, suppose that the stocks of money and bonds
each grow at a rate of 10 percent per year, forever. In this
case the stock of money relative t6 bonds remains constant,
and the model implies that the real interest rate remains constant and private nominal wealth grows at a rate of 10 percent per year Assuming, for simplicity, that real income is
not growing and that real government spending and taxes are
constant, equilibrium is achieved with a 10-percent rate of inflation in wages and prices, a constant value of real cash
balances, constant real government debt, and constant real
private wealth. With real government debt and the real interest rate constant, real interest payments on the debt are
also constant
This solution can easily be verified by noting that if Yo' ro'
Po' and Wo are equilibrium values of real income, the real interest rate, the price level, and the wage rate for (M = Mo) and
(8=8 0) in the model, then Yo' ro,1 .10Po' and 1.10Wo are
equilibrium values for (M = 1.10M o) and (8 = 11080 ), Hence,
even though nominal government debt is growing unbounded,
real principal plus interest is not, and the economy's ability to
finance the debt is not in question
It is important to point out that the primary factor responsible for preventing the unbounded growth of the real value
of the debt or real interest payments is inflation . And the factor responsible for the inflation is the issuance of unbacked
government paper (of both the Federal Reserve and the
Treasury) In short. through its inflationary effect. permanent
government debt erodes its own real value and thereby acts
to prevent unbounded growth of the real debt of the
government
17
Miller, " Deficit Policies, Deficit Fallacies, " 2_
17
The issue of whether government bonds are
backed by taxes is clearly more general than the
setting of any specific model. Nevertheless the
simple model contained in Box A, set forth' early
in the literature by Don Patinkin, is helpful for investigating the implications of temporary and
permanent government debt. 1 8 It is important to
keep in mind that the primary benefit of this early
model is expository. The weaknesses of the model
for illustrating the dynamic properties of ongoing
government deficit policies are well recognized.
Nevertheless, it contains the principal ideas underlying a regime of temporary government debt as well
as one of permanent government debt.
The arguments on the left-hand side represent
market demand functions, and those on the righthand side represent market supply. The demand for
labor is assumed to be inversely related to the real
wage rate, whereas labor supply is positively related
to the real wage rate. Aggregate commodity demand is positively related to real income and
private real wealth but inversely related to the real
interest rate. The demand for bonds is assumed to
be directly related to real income, the nominal interest rate, and real wealth and to be proportional
to the price level. 19 Finally, aggregate money demand is assumed to be directly related to real income and real wealth, inversely related to the
nominal interest rate, and proportional to the price
level.
Notice, in particular, that nominal wealth of the
private sector consists of cash balances plus a fraction, k, of the stock of privately held government
bonds. By definition, k is the fraction of outstanding
government debt that is perceived as not backed by
taxes. As defined by Patinkin,
This definition of k focuses on the individual's
perception of the backed ness of government
debt-that is, on the rationality issue. We may, instead, accept the notion that individuals are rational
and focus on the backedness question. That is,
following Samuelson, define k as the fraction of
Interest-bearing public bonds .. which no taxpayer
rationally expects to have to help retire ."
The implications of both temporary and permanent
government debt may then be determined easily by
using the model.
Clearly, the Ricardian Equivalence Theorem
claims that k equals zero. Substituting (k =0) into
the model verifies that the issuance of government
debt in lieu of current taxation does not affect aggregate demand, the interest rate, or even the price
level in this case. Hence, the Ricardian Equivalence
Theorem is valid in a regime of temporary government debt, provided individuals are rational.
A regime of permanent government debt corresponds to a value of 1 for k. The economic
implications of permanent government debt may
easily be determined by substituting (k=1) into the
model. In this case, private nominal wealth consists
of cash balances plus private holdings of outstanding government debt. An increase in the stock of
privately held government bonds adds to nominal
wealth, stimulates aggregate demand (until a new
equilibrium is established), and drives up the price
level. The model clearly illustrates that permanent
government debt is not neutral; in particular, it is
inflationaryn
As is well understood, government can cause inflation by printing more money It can also cause inflation by printing more bonds Additions to the
stocks of money or bonds, by increasing the total
amount of nominal wealth, increase private
demands for goods and services . The increased
demands, in turn, push up the prices of goods.23
k is a constant (greater than zero and less than
one) reflecting the degree to which individuals do
not discount the future tax liabilities connected
with government bonds
20
21
Economics, 326
22
The result that permanent government debt is inflationary is
not unique to the setting of this particular model. See Bryant,
"Analyzing Deficit Finance in a Regime of Unbacked Government Paper," for example, for an analysis of the inflationary
implications of unbacked government paper in an
overlapping-generations model Note also that money is not
neutral with permanent government debt. In particular, an increase in the stock of money has interest rate effects when
part of government debt is permanent.
23
Miller, "Deficit Policies, Deficit Fallacies," 2 .
18 See Patinkin, Money, Interest, and Prices, 289 - 94, for a
presentation and discussion of this model.
19 The bond demand function is an "excess" demand function'
that is, it is the demand for bonds by the private sector in e'xcess of the private supply of bonds The variable B represents
the market value of outstanding government bonds, rather
than the number of bonds outstanding
20, Money, Interest, and Prices, 289
18
Federal Reserve Bank of Dallas
The Quantity Theory of Money
This section considers the Quantity Theory of
Money under both temporary government debt and
permanent government debt. It is argued that the
Quantity Theory is valid in a regime of temporary
government debt, provided individuals are rational,
but is not valid with permanent government debt
regardless of the rational ity issue. 2'
According to the crude Quantity Theory, "prices
must always be proportional to the amount of
money-so that doubling M must exactly double
P ."2S I t is easy to demonstrate that this proposition
holds under a regime of temporary government
debt, provided individuals are rational. This is the
case because Treasury debt is perceived as fully
backed government paper and is not net nominal
wealth, and a switch to debt finance does not affect
any thing- not even the price level. Substituting
(k = 0) into the model verifies that net wealth consists only of money balances and that the price
level is not affected by an increase in the volume of
Treasury paper. Furthermore, doubling the stock of
money exactly doubles the price level. Hence, the
Quantity Theory holds.
With permanent government debt, however, this
is clearly not the case. Substituting (k = 1) into the
model demonstrates that the price level is proportional to the stock of money plus permanent debt
but is not proportional to the stock of money alone.
Doubling the stock of money less than doubles the
price level because the price level is also related to
the stock of outstanding government debt. Hence,
the Quantity Theory of Money is not valid in a
regime of permanent government debt.
Although the discussion here is for the case where
all government debt is permanent, the results may
easily be generalized to the case where a portion of
the debt (k) is permanent and a portion (1 - k) is
temporary. The basic result is unchanged-the price
level is related to the stock of unbacked government paper, (M+kB)-and government bonds issued
24 Of course, irrationality could be of a very peculiar type
Namely. individuals could perceive future tax liabilities even
though they do not exist, in which case the Equivalence
Theorem would still hold. That case is not taken seriously
here
25
See Samuelson, Economics, 326, for this statement of the
Quantity Theory of Money
Economic Review I May 1985
permanently "have effects similar to those of the
non interest IOUs we call M [money]."26 Only in the
case where all government debt is temporary (k = 0)
does the Quantity Theory hold. As expressed by
Samuelson:
believers in a crude quantity theory ...should reformulate their theory to say:
Doubling M and permanent public debt [kB]
will, other things equal, double all Ps and leave all
relative Ps, physical quantities, and interest rates
unchanged in the new long-run equilibrium . 27
Policy implications
In this section the policy implications of permanent
government debt are addressed. It is argued that to
the extent government debt is permanent, monetary
policy is limited in its ability to control the price
level. When a portion of the debt is temporary,
however, monetary policy retains its ability to control the price level, but a larger open market exchange is required than that commonly recognized.
Monetary policy retains a limited ability to affect
the price level even when all government debt is
permanent, but that effect is secondary and most of
the power of monetary policy is lost.
To make the analysis as general as possible, both
temporary and permanent government debt are considered. The exposition may be simplified by assuming that a portion, (1 - k), of each unit of public
debt is temporary and a portion, k, is permanent.
That is, for each unit of public debt outstanding, the
fraction (1 - k) of the principal plus interest is
financed by future taxes, and the fraction k is
financed by debt issuance or money creation.
Consider now the effect of open market operations.28 In particular, suppose that the Federal
26. See Samuelson, Economics, 326
27. Economics, 326-27
28 See Patinkin, Money, Interest, and Prices, 291- 94, for a
discussion of the effect of open market operations with permanent government debt The budget equation of government used here is the consolidated budget equation of the
Treasury and the Federal Reserve . The Treasury finances
deficit spending of Congress with the sale of Treasury
securities, a portion of which the Federal Reserve purchases
by the creation of currency (base money) Hence, deficitfinanced government spending ultimately involves some combination of increased holdings of Treasury paper plus Federal
Reserve paper on the part of the private sector Interest
earned on Treasury securities by the Federal Reserve is also
assumed to be returned to the Treasury
19
Box B
Summary of the Implications of Temporary and Permanent Government Debt
Determined from Using the Patinkin Model
Government Debt Is Temporary
(k = 0)
Government Debt Is Permanent
(k
= 1)
A Portion of Government Debt
Is Temporary, and
a Portion Is Permanent
(0
< k < 1)
Ricardian Equivalence Theorem
Holds. Government debt is neutral.
In particular, government debt is
not inflationary. The interest
rate is independent of the stock of
outstanding government debt.
Ricardian Equivalence Theorem
Does Not Hold. Government debt
is not neutral. The issuance of
government debt is inflationary
and raises the interest rate.
Ricardian Equivalence Theorem
Does Not Hold. Government debt
is, in general , not neutral The
theorem holds only for the portion
of government debt that is temporary and not for that which is permanent. The issuance of permanent
government debt is inflationary and
raises the interest rate.
Quantity Equation of Money Holds.
More generally, the price level is
related to the stock of money only
and not to government debt.
Quantity Equation Does Not Hold.
The price level is related to both
the stock of money and government
debt but not to money only .
Quantity Equation Does Not Hold.
The price level is related to both
the stock of money and permanent
government debt but not to money
only.
Monetary Policy Works. Open
market exchanges of Treasury
securities for base money are
effective in controlling private
financial wealth and the
price level.
Monetary Policy Works But Is
Weakened by the Presence of
Government Debt. Open market
exchanges of Treasury securities
for base money are effective only
to the extent they involve a
capital gain or loss in private
wealth.
Monetary Policy Works But Is
Weakened by the Presence of the
Permanent Government Debt. To
achieve a given price effect, a
larger open market exchange of
base money for Treasury securities
is required than if the debt were
all temporary.
NOTE: Based on the assumptions that individuals are rational and the economy is closed
20
Federal Reserve Bank of Dallas
Reserve purchases $1 worth of Treasury securities.
The purchase has the effect, on the one hand, of increasing the volume of unbacked government paper
by $1 as base money is created to buy the Treasury
securities. On the other hand, k percent of the
Treasury securities purchased were permanent and
therefore unbacked by taxes. Hence, the net effect
of the open market purchase is to increase the
volume of privately held unbacked government
paper by $(1 - k), rather than the $1 commonly
believed. A similar result follows for an open
market sale of Treasury securities. Private nominal
wealth falls by only $(1 - k) when the Federal
Reserve exchanges $1 in Treasury securities for $1 in
money.
The presence of permanent government debt
clearly inhibits the power of monetary policy to
control the price level, but it does not render policy
powerless. Without permanent government debt (for
k = 0), a $1 reduction in private nominal wealth may
be directly achieved with the open market sale of
$1 in Treasury securities; with permanent government debt, a $1 reduction may be achieved with the
sale of $1/(1 - k) in Treasury securities. Hence,
monetary policy retains its ultimate ability to control nominal wealth and the price level, but larger
exchanges of money for debt are required .
The choice of $1 for the money-bond exchange is
motivated not only by convenience but also by the
secondary effects that such a small exchange allows
the analysis to disregard. In particular, the interest
rate implications of exchanging $1 in bonds for $1
in money in the above examples have been ignored .
The importance of the interest rate effect may be illustrated by considering the case where all government debt is permanent (k = 1). Note that in this
case, nominal wealth of the private sector is simply
base money, M, plus the market value of privately
held government debt, 8 - that is, (M + 8). Open
market operations move M and 8 in opposite directions but not by equal amounts, since a change in
the stock of interest-bearing paper (8) relative to
non-interest-bearing paper (M) in the economy affects the interest rate.
Does an open market sale of Treasury securities,
then, still constitute "tight monetary policy"? To
answer this question, we need only focus on the interest rate effects of reducing the stock of money
relative to bonds in the model. A reduction in M
relative to 8 raises interest rates, which lowers the
Economic Review I May 1985
price of all outstanding government debt. 29 The
resulting capital loss on outstanding government
debt causes an overall loss of private nominal
wealth. 30 This result may easily be verified in the
model by noting that only a reduction in overall
(M + 8) is consistent with new equilibrium following
the open market sale of Treasury securities. Hence,
open market sales of Treasury securities by the
Federal Reserve still constitute "tight monetary
policy," but their effect is clearly secondary and
weakened by the presence of permanent government debt.
The inflationary effects
of government debt, 1950-1984
This section provides empirical support for the
hypothesis that government debt is inflationary by
testing for the price effects of government debt in
the United States over the 1950-84 period. 31 The
procedure used was to apply Granger-Sims causality
tests to examine the relationship between the price
level and both money and debt and to compare the
results with those from tests involving money only
and debt only.32
29. It should be noted that the interest rate effect of the money·
bond exchange is not common to all models See, for example, Bryant and Wallace, "The Inefficiency of Interest-bearing
National Debt"
30 The effect on private nominal wealth of an open market sale
of Treasury securities may be demonstrated mathematically
by defining V as the par value of government bonds and Pa as
the price of a government bond (per $1 of par value), so that
B equals PaV An open market sale of L\V units of Treasury
securities at the unit price Pa involves a reduction in L\M units
of currency, so that PaL\V equals - L\M Private nominal
wealth equals (M+kB), which equals (M+kPaV); hence, for
(k=1), increments to private nominal wealth may be written
as (L\M + PaL\V + VL\Pa ) The effect on private nominal wealth
of the open market sale may then be written as simply VL\Pa,
which is the capital 1055 on outstanding government debt.
31 . Previous empirical studies have looked for aggregate demand
effects or interest rate effects of government debt in order to
assess the neutrality question Data on the market value of
Treasury debt reported recently by W Michael Cox and Eric
Hirschhorn, "The Market Value of US Government Debt;
Monthly, 1942-1980," Journal of Monetary Economics 11
(March 1983): 261-72, and by W Michael Cox, "The Behavior
of Treasury Securities; Monthly, 1942-1984," Journal of
Monetary Economics, forthcoming, make it possible, however,
to test directly for the inflationary effects of public debt.
32 The Granger-Sims procedure is used here because it permits
tests of significance of groups of variables See C W , J
21
The first step was to select particular regression
forms with which to test the causes of inflation. By
strict interpretation of the Quantity Theory of
Money, the price level at a point in time is proportional to the stock of money available at that time.
It is widely recognized, however, that money's
effect on prices is not purely contemporaneous.
Money may affect prices with a lag or even with a
lead. Broadly interpreted then, the Quantity Theory
claims that the price level is directly related to the
quantity of money (only). Economists supporting this
principle (whom I will call monetarists) regard inflation as purely a monetary phenomenon. This summarizes the specific hypothesis to be tested here
and will be referred to as the monetarist hypothesis,
Hm'
Hm' An increase in the outstanding volume of
Federal Reserve paper is inflationary, but
an increase in the volume of Treasury paper
is not.
The competing hypothesis, H mb , is that both money
and debt are inflationary. Specifically,
H mb . An increase in the outstanding stock of
either type of government paper (Federal
Reserve notes or Treasury notes) is
inflationary.
Table 1 summarizes the results of all the regression equations. Rather than present the entire set of
estimated coefficients and standard errors for each
regression (approximately 130 statistics), the
statistics reported will be limited to only those
necessary for comparing the explanatory power of
money and debt. Specifically, Table 1 reports the
sum of squared errors (SSE) and the adjusted R2 (i~2)
for each regression equation . By construction, the R2
statistics measure the percentage of variation in the
price variable (adjusted for degrees of freedom) accounted for by the hypothesized explanatory
Granger, "Investigating Causal Relations by Econometric
Models and Cross-spectral Methods," Econometrica 37 (July
1969): 424-38, and Christopher A Sims, "Money, Income, and
Causality," American Economic Review 62 (September 1972):
540-52, for a detailed description of the procedure
variables . The higher the R2, therefore, the higher is
the explanatory power of the hypothesized set of
variables. 33 In contrast, SSE measures the unexplained variation in the price variable by squaring
the residual (actual less predicted) values and summing these over all observations. The lower the SSE,
therefore, the greater is the explanatory power of
the hypothesized set of variables.
Equation 1 reports the results of the regression involving money only, equation 2 involves debt only,
and equation 3 includes both money and debt. Including both money and debt makes it possible to
conduct a direct test of the monetarist hypothesis.
Clearly, equation 3 "nests" equation 1, in that it
contains all the explanatory variables present in
equation 1 plus the additional debt variables. Equation 1, then, is a restricted form of equation 3, the
specific restriction being that the estimated coefficients on the debt variables are not significantly different from zero. This restriction may easily be
tested by comparing the sum of squared errors for
equation 1 (SSE1) with that for equation 3 (SSE3).
The test is referred to as a joint F test, and the
specific statistic is
(SSE1 - SSE3)/"
SSE3/(n -k)
where r is the number of restrictions in the moneyonly regression, n is the number of observations,
and k is the number of explanatory variables in the
unrestricted (money and debt) regression. Note that
this statistic follows an F distribution and has a
critical value of 2.01 for significance at the 95percent level. Calculation of the statistic gives an
actual value of 2.37. Hence, Hm is rejected in favor
of Hmb 34
To gain a better appreciation for the weight of
this finding, it is instructive to subject the money
variables to the same rigorous test-that is, test
Hmb against the hypothesis that only debt is inflationary, H b. Specifically, consider the hypothesis
H b . An increase in the outstanding stock of
Treasury notes is inflationary, but an increase in the stock of Federal Reserve notes
is not.
33 Rl'S in this study are naturally low because a common trend
was extracted from each series With the trend left in (running
levels against levels). the i?l' s were all above 99, but the error
terms were highly autocorrelated
22
34 Specifically, the set of explanatory variables BI + 2 , BI + 1 , BI •
BI _ 1 , .. , BI _ 6 cannot be rejected as having insignificant
predictive content.
Federal Reserve Bank of Dallas
Table 1
RESULTS OF CAUSALITY TESTS RELATING INFLATION
TO THE GROWTH IN MONEY AND GOVERNMENT DEBT
Equation 1
Equation 2
Equation 3
Equation 4
Equation 5
...
..
..
...
Explanatory
variables'
SSE
R2
Money only
Debt only
Money and debt
Money plus debt
Money plus .59 debt
.001470
.001264
.001201
.001260
.001256
.345
.437
.415
.431
.432
Dependent variable
Inflation
(Pt ) = growth rate of the consumer price index.
Independent variables
Money
(Mt ) = growth rate of the St.
Louis Fed
monetary base.
Debt t ) = growth rate of the market value of
privately held gross federal debt.
Money plus debt (~1 t) = growth rate of M t + Bt .
Money plus .59 debt (W2 t) = growth rate of M t + .59B t ·
(B
Specific explanatory variables, by equation
1
Pt -1' Pt- 2, PI - 3, Mt+2' Mt+1' Mt, Mt- 1, ... , Mt - 6 ·
3 .. -
PI - 1, Pt - 2. Pt- 3, M1+ 2 , M1+ 1 ,M1, Mt - 1, ... , Mt - 6 ,
81+ 2,81+ 1 , 8t , 81- 1 , ... , 81- 6 ,
1. Including three lags on inflation.
NOTE: All series are seasonally adjusted residuals, calculated according to the
process described in the Appendix. Data were quarterly over the period
1950-84. Because of the lags involved in the estimation equations, all
regressions were estimated for the first quarter of 1952 to the second
quarter of 1984. See the Appendix for a complete description of the
regression procedure used in testing the hypotheses.
SSE is the sum of squared errors.
R2 is the coefficient of determination adiusted for degrees of freedom
Economic Review I May 1985
23
This hypothesis is clearly the antithesis of Hm; it
claims that debt is inflationary but money is not.
Also, there is clearly an analogous nesting of
hypotheses . In particular, equation 2 amounts to a
restricted form of equation 3, with the specific
restriction that the coefficients on the money
variables are zero. Calculation of the F statistic
gives an actual value of 0.56, which is substantially
less than the value of 2.01 required for significance
at the 95-percent level. Hence, Hb cannot be rejected. Strictly interpreted, the results of this test
imply that money does not add any significant
predictive content not already found in the debt
variable. The purpose of considering this hypothesis,
however, is not to cast doubt on the largely unquestioned role of money in affecting prices but, rather,
to indicate the degree of rigor the test possesses.
The results of the tests, therefore, should be more
broadly viewed as implying that government debt is
at least as good as money in explaining the behavior
of the price level.
Among the regressions reported in Table 1 are
two equations testing the specific hypothesis that
money plus debt is inflationary. For each of these
regressions, a nominal wealth variable is first
created by adding money and a measure of government debt. The first nominal wealth variable, W1, is
simply money plus all government debt. Hence, this
measure of nominal wealth corresponds to the value
of all privately held government paper. The second
measure, W2, is money plus 59 percent of government debt. Using this fraction of debt is justified
because over the 1950-84 period the principal on
government debt was essentially rolled over and, on
average, each $1 in interest was financed with only
41 cents in taxes. 35 Therefore, government debt may
essentially be viewed as an infinite-lived consol on
35
24
In "What Is the Rule for Financing Public Debt?" Economic
Review, Federal Reserve Bank of Dallas, September 1984,
25-31, W Michael Cox investigates the Treasury's rule for
financing public debt over the 1950-81 period . The basic finding of the study is that government debt may be viewed as an
infinite-lived consol (since the principal on public debt has
continuously been rolled over) on which 41 percent of the interest has been tax-financed and 59 percent has been deficitfinanced . This implies a value of 59 for k Cox also finds
evidence of a shift in the Treasury's interest-financing rule in
the early 1970s. That case is not taken up in the empirical
work here.
which 59 percent of the interest is deficit-financed.
I n terms of the above discussion, this constitutes a
regime under which 59 percent of outstanding debt
is permanent and 41 percent is temporary .
Equations 4 and 5 report the results of the regressions involving W1 and W2, respectively. The question of whether government debt is inflationary may
be examined again by comparing the adjusted R2· s
across equations 1, 2, 4, and 5. Note that in the
regression involving money only, R2 is approximately
.35 whereas for money plus government debt it is
.43. Although the improvement may not appear
substantial, debt alone actually outperforms debt
plus money. Again, the purpose of this comparison
is not to question the role of money in affecting
prices but to call attention to the rigor of the examination to which the debt variable is subjected.
In light of the results of these tests, it is difficult
to dismiss government debt as an important variable
in explaining inflation. The tests indicate that over
the 1950-84 period the price level was as much
determined by the volume of outstanding interestbearing government paper (debt) as by non-interestbearing paper (money). Apparently, Treasury notes
are viewed as unbacked government paper, with the
same inflationary implications as money .36
Conclusion
It is commonly believed that an open market sale of
Treasury securities by the Federal Reserve reflects
"tight" monetary policy and that the sale reduces
inflation. The argument is that by exchanging
Federal Reserve paper for Treasury paper, the
Federal Reserve reduces private wealth because
money is a component of private nominal wealth
but government bonds are not. The article here
shows that this result is true under a regime of temporary government debt but not true with permanent government debt. When the Treasury adopts a
regime of permanent government debt, Treasury
36
The results presented here are derived over the extended
period 1950-84 Clearly, in recent years the relationship between government debt and inflation has not behaved as
would be predicted by this study . Strictly speaking, the
arguments here concerning the inflationary implications of
permanent government debt are correct only for the case of a
closed economy The feeling is that this breakdown is temporary and due mainly to external considerations, such as the
unprecedented current account deficit.
Federal Reserve Bank of Dallas
securities are unbacked paper, with inflationary effects much the same as those of money.
Put differently, government spending requires
some type of "tax," and there are essentially only
two taxes a government may levy on its citizens.)?
One is a direct tax (income tax, head tax, sales tax,
and so on), and the other is an inflation tax. Con-
37
Again, this argument is strictly correct only for the case of a
closed economy See notes 13 and 36 for a discussion of the
open-economy considera tions.
tinuous budget deficits call for the issuance of unbacked government paper-money or bonds-and
involve an inflation tax regardless of whether the
paper is Federal Reserve paper or Treasury paper.
Tests relating the price level to money and
government debt in the United States indicate that
over the 1950-84 period the price level was as
strongly linked to the volume of Treasury notes as
to Federal Reserve notes. Apparently, public debt is
viewed by the private sector as unbacked paper,
with price implications much the same as those of
money.
Appendix
Testing for Inflationary Effects of the Government Debt
Preparation of the data
As a first step, individual data series were chosen to
represent the price, money, and government debt
variables. Selection of the data series was based on
two criteria _ First, it was generally desirable to test the
hypothesis over as long a period as possible Second,
to avoid spurious econometric problems, statistically
comparable raw data series were needed. These
criteria led to the selection of seasonally unadjusted
monthly observations for each variable over the sample period 1950-84
For the price level, the consumer price index was
used. The monetary base was chosen as the appropriate money variable because it is the only part of
money that is net wealth. Specifically, the St louis
Fed monetary base was used since data are available
back to 1950; for other measures of the base, data
begin with 1959. For the government debt variable, a
market value measure was required , Ideally, what is
needed is the market value of privately held permanent government debt, The author has recently
reported statistics on the market value of privately
Economic Review I May 1985
held gross federal debt, and that is the series used
here.' The question of which part of the debt is permanent and which part is temporary is taken up in the
text
After selection of the data, quarterly averages of the
monthly observations were calculated, and stationary
representations of the three variables were sought
Because each series exhibited a clear upward trend,
the data were transformed into growth rates . The
resulting growth rate variables still exhibited a mildly
upward drift; hence, each variable was regressed on
time (and a constant), and the residuals retrieved. Each
residual series showed a significant seasonal pattern,
so the data were adjusted using seasonal dummies .
The resulting seasonally adjusted residual growth rate
series were selected as the stationary representations
1 The market value series reported by W Michael Cox. " The
Behavior of Treasury Securities: Monthly. 1942-1984," Journal 01
Monetary Economics . forthcoming. are the monthly market values
of gross federal debt (marketable plus nonmarketable debt)
held by the private sector
25
of the price, money, and debt variables.
Specification of the regression equations
In selecting the specific forms of the regression equations to be used in testing the h'ypotheses, the current
value of the price variable (P) was regressed against
lead, current, and lag values of money (M) and debt
(B), as well as against lag values of the price level. By
including lag values of the price level, the regression
equation is essentially being allowed to attribute as
much as possible of the current behavior of the price
level to past values of the price level. Any remaining
power that money and debt have in explaining the
price level may then be properly attributed to these
variables. It is in this sense of incremental predictive
content that the procedure is referred to as a CrangerSims "causality" test!
Extensive leads and lags on money and debt and
lags on prices were first introduced and then individually eliminated until a significant loss in overall
explanatory power of the' regression (as measured by
adjusted R2) was encountered. This procedure resulted
in the following regression structure: lags of three
26
quarters on prices and leads of two quarters, a contemporaneous value, and lags of six quarters on both
money and debt. At this stage the residuals indicated
significant first-order autocorrelation, which, if untreated, would lead ordinary least squares to provide
inconsistent estimates of the regression coefficients.
Hatanaka's two-step "residual adjustment" procedure
was used to correct the problem .'
2 For a discussion of this procedure. see C w J Granger, "Investigating Causal Relations by Econometric Models and Cross-spectral
Methods," Econometrica 37 (J uly 1969): 424-38, and Christopher A
Sims, "Money, Income, and Causality," American Economic
Review 62 (September 1972): 540-52
3 See Michio Hatanaka, "An Efficient Two-Step Estimator for the
Dynamic Adjustment Model with Autoregressive Errors," Journal
of Econometrics 2 (September 1974): 199-220 The Hatanaka procedure is the same as the usual autocorrelation correction except
that an extra regressor (a lagged residual estimator obtained by the
method of instrumental variables) is included This procedure
yields estimators that are asymptotically equivalent to the maximum likelihood estimators and, hence, are consistent and asymptotically efficient in the presence of normality
Federal Reserve Bank of Dallas
FEDERAL RESERVE BANK OF DALLAS
STATION K, DALLAS, TEXAS 75222
ADDRESS CORRECTION REQUESTED
BULK RATE
U.S. POSTAGE
PAID
PERMIT NO. 151
The Economic Review is published by the Federal Reserve Bank of Dallas and
will be issued six times in 1985 (January, March, May, July, September, and
November). This publication is sent to the mailing list without charge, and additional copies of most issues are available from the Public Affairs Department,
Federal Reserve Bank of Dallas, Station K, Dallas, Texas 75222. Articles may be
reprinted on the condition that the source is credited
@FedFRASER