ȱ
ȱȱDZȱǰȱ
ǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱǯȱ¢ȱ
ȱȱȱȱ
ȱŘŜǰȱŘŖŖŞȱ
ȱȱȱ
ŝȬśŝŖŖȱ
ǯǯȱ
řŚŜŞŘȱ
ȱȱȱ
Pr
epared for Members and Committees of Congress
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
¢ȱ
This report provides a broad overview of Treasury debt management and examines changes in
debt sales implemented by the Clinton and Bush Administrations. It covers the following seven
topics: the concept of federal debt management, the concept of the term structure of interest rates,
theories of the term structure of interest rates, policy objectives, debt management (1945-1992),
Clinton Administration policies, and Bush Administration policies.
The Treasury sells securities in capital markets to finance budget deficits and to refinance its
maturing debt. The magnitude of Treasury sales of its securities can vary substantially from year
to year. Congress does not directly affect debt management decisions. However, Congress
oversees the operations of the Federal Reserve System, which cooperates with the Treasury in
pursuing debt management objectives. Treasury officials have testified before congressional
committees about federal debt management. The Government Accountability Office has written
reports concerning federal debt management as requested by congressional committees.
Federal debt management, narrowly defined, concerns Treasury’s decisions about sales of
Treasury bills, notes and bonds, which affect the term structure of the privately held interest-
bearing federal debt. The term structure of interest rates relates the average yield to maturity of a
particular type of debt instrument (such as U.S. Treasury securities) to its time (such as years) to
maturity. Financial economists have different theories concerning the causes of the term structure
of interest rates and the changes in the term structure over the business cycle. The four primary
theories are the expectations theory, the risk averse theory, the segmented market theory, and the
preferred habitat theory.
The Treasury must float debt maturities (securities), which are salable if the federal government is
to continue to operate. Subject to this mandatory objective, the Treasury has discretion in
pursuing three secondary objectives: economic stabilization, minimization of interest costs, and
neutrality (make debt sales as predictable as possible). The pursuit of one of these secondary
objectives may adversely affect another objective. Furthermore, these different objectives depend
on different theories of the term structure of interest rates. Beginning in FY1975, the Treasury
decided to pursue a policy of neutral debt management while attempting to increase the average
length to maturity of the marketable Treasury debt. According to this policy of neutral debt
management, the Treasury spread its borrowing needs widely across the maturity spectrum and
sold debt issues on a regular and predictable schedule.
The Clinton Administration shortened the debt maturity in order to reduce interest costs. This
policy was controversial and was reversed as budget surpluses occurred. The Bush
Administration also shortened the debt maturity in order to reduce interest costs, which has also
been controversial.
This report will not be updated.
ȱȱȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
ȱ
Concept of Federal Debt Management............................................................................................ 1
Concept of the Term Structure of Interest Rates.............................................................................. 2
Theories of the Term Structure of Interest Rates............................................................................. 3
Expectations Theories ............................................................................................................... 4
Pure Expectations Theory .................................................................................................. 4
Error-Learning Theory........................................................................................................ 6
Consistency with Data ........................................................................................................ 6
Risk Averse Theory ................................................................................................................... 7
Types of Risk ...................................................................................................................... 7
Consistency with Data ........................................................................................................ 8
Segmented Market Theory ........................................................................................................ 9
Effectiveness of Debt Management .................................................................................... 9
Consistency with Data ...................................................................................................... 10
Preferred Habitat Theory......................................................................................................... 10
Consistency with Data .......................................................................................................11
Prevailing View of Financial Economists ..........................................................................11
Policy Objectives........................................................................................................................... 12
Borrow Funds (Necessary Objective) ..................................................................................... 12
Alternative Secondary Objectives........................................................................................... 12
Economic Stabilization ..................................................................................................... 12
Minimum Interest Cost ..................................................................................................... 13
Neutrality .......................................................................................................................... 13
Debt Management (1945-1992) .................................................................................................... 13
Clinton Administration Policies .................................................................................................... 17
Bush Administration Policies ........................................................................................................ 19
ȱ
Figure 1. Hypothetical Term Structures of Interest Rates ............................................................... 5
Figure 2. Average Length of Privately Held Marketable Debt ...................................................... 14
ȱ
Table A-1. Example of Present Values .......................................................................................... 23
Table A-2. Maturity Distribution of Marketable Interest-Bearing Public Debt Securities
Held by Private Investors ........................................................................................................... 23
¡ȱ
Appendix. Examples of Interest Rate Risk.................................................................................... 21
ȱȱȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
ȱ
Author Contact Information .......................................................................................................... 24
ȱȱȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
his report provides a broad overview of debt management by the Department of the
Treasury (Treasury) and examines changes in debt sales implemented by the Clinton and
T Bush Administrations. It covers the following seven topics: the concept of federal debt
management, the concept of the term structure of interest rates, theories of the term structure of
interest rates, policy objectives, debt management (1945-1992), Clinton Administration policies,
and Bush Administration policies.
The Treasury sells securities on capital markets to finance budget deficits and to refinance its
maturing debt. The magnitude of Treasury sales of its securities can vary substantially from year
to year. Congress does not directly affect debt management decisions. However, Congress
oversees the operations of the Federal Reserve System, which cooperates with the Treasury in
pursuing debt management objectives. Treasury officials have testified before congressional
committees about federal debt management. The Government Accountability Office (GAO) has
written reports concerning federal debt management as requested by congressional committees.1
ȱȱȱȱȱ
In a broad sense, federal debt management encompasses policies that affect the maturity
composition of the privately held, interest-bearing federal debt and determine the characteristics
of federal debt instruments.2 The U.S. Treasury decides the maturities of new marketable debt.
This marketable debt consists of bills which mature in one year or less, notes which mature in 2
to 10 years, and bonds which mature in more than 10 years.3 The Treasury also sells savings
bonds, but these are not marketable. Furthermore, the Treasury conducts advance refundings; that
is, it offers holders of a particular outstanding debt issue the option of exchanging that debt issue
for a new debt issue with a longer maturity.
The Federal Reserve System (Fed) also can affect the maturity composition of federal debt
through open market operations. The Fed purchases Treasury securities in order to expand the
money supply and, conversely, sells Treasury securities in order to contract the money supply.
Obviously, the maturities of the securities that the Fed decides to buy (or sell) affects the maturity
composition of the Federal debt held by the rest of the public.4 Through open market operations,
the Fed has the capacity to partially offset Treasury debt management decisions. But the Fed
1 As examples, in September 1999, GAO published Debt Management in a Period of Budget Surplus in response to a
request from the chairman of the Senate Budget Committee and the chairman of the House Ways and Means
Committee; and in February 2001, GAO published Debt Management Actions and Future Challenges at the request of
the chairman of the House Ways and Means Committee, the chairman of the Social Security Subcommittee of the
House Ways and Means Committee, and the chairman of the Senate Budget Committee.
2 Privately held, interest-bearing federal debt accounts for a part of federal debt. Much of the federal debt is held in
federal trust funds—in other words, the federal government has issued IOUs to itself. For an overview of the federal
debt, see CRS Report RL31590, The Federal Government Debt: Its Size and Economic Significance, by Brian W.
Cashell.
3 Treasury bills are sold on a discount basis; that is, an investor earns the difference between the purchase price and the
redemption value. Treasury notes and bonds pay semi-annual coupons (interest). The initial sales price of a Treasury
note or bond may be above or below the redemption value, hence, in addition to interest payments, an investor may
have an acquisition discount or premium.
4 The Treasury excludes Fed holdings of Treasury securities in its calculation of the average length to maturity of the
privately held marketable Treasury debt. Hence, the Fed purchase (or sale) of a particular maturity (such as bills) on
financial markets would change the composition of the non-Fed privately held marketable Treasury debt and thus
change the average length to maturity.
ȱȱȱ
ŗȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
coordinates its open market operations with Treasury debt management policies. Thus, the
Treasury usually plays the dominant role in determining the maturity structure of federal debt
held by the investing public.
Finally, the broad concept of debt management includes Treasury’s ability to add or change the
characteristics of its debt issues. Examples of these characteristics include the length of maturity,
call features, conversion features, variable rate features, and indexation for changes in the price
level. The Treasury has issued securities with new maturities, discontinued some maturities, and
included or deleted call features from different security issues.5 The Treasury sells savings bonds
with fixed rates and price-indexed rates, but these bonds are not marketable; that is, they cannot
be resold on financial markets.6 In January 1997, the Treasury began issuing marketable securities
indexed to changes in the price level.
Unless specified otherwise, in this report debt management is narrowly defined as Treasury
decisions about debt sales that affect the maturity composition of privately held interest bearing
federal debt. This report focuses on marketable Treasury debt. Unless specified otherwise,
monetary policy is assumed to be held constant.
The Treasury’s changes in the maturity composition of the federal debt are important because of
their possible effects on the term structure of interest rates and the interest costs on the national
debt. A change in the term structure of interest rates on marketable Treasury securities (all else
held constant) will cause a parallel change in the term structure of private marketable debt
because Treasury securities and private securities of the same maturity are highly substitutable.
ȱȱȱȱȱȱ ȱȱ
The term structure of interest rates relates the average yield of a particular type of debt instrument
(such as U.S. Treasury securities) to its time (such as years) to maturity. The average yield to
maturity of a security is calculated based on annual interest payments and the difference between
the current market value of the security and its maturity value (or face value). Since it takes into
account the gain or loss that will occur at maturity, the average yield to maturity differs from the
coupon yield and the current yield. The coupon yield or nominal yield is the yield stated on the
face of the security. The coupon yield is not affected by any difference between the cost of a
security and its maturity value (or face value). The coupon yield is useful in indicating the dollar
amount of interest paid each year. The current yield of a security is simply the dollar amount of
interest divided by the current market price of the security.7
The distinction between the concepts of coupon yield, current yield, and average yield to maturity
can be illustrated by a simple example. Assume an individual purchases for $950 a bond with a
10-year maturity and a redemption value (or face value) of $1,000. The bond pays interest of $60
a year. The coupon yield or nominal yield per year would equal $60/$1,000 or 6%. The current
yield per year would equal $60/$950 or 6.32%.8 The average yield to maturity would be 6.70%.
5 A call feature gives the Treasury the option of redeeming a particular security issue before its date of maturity.
6 Current information about savings bonds is available at http://www.treasurydirect.gov, visited September 12, 2008.
7 Lester V. Chandler and Stephen M. Goldfeld, The Economics of Money and Banking, Seventh Edition (New York:
Harper & Row, 1977), pp. 58-59.
8 Ibid., p. 59.
ȱȱȱ
Řȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
The average yield to maturity is the average return on a security from the present date to the date
of final maturity, which is calculated based on all future interest payments and any capital gain or
loss at final maturity. In this example, the average yield to maturity of 6.70% was calculated by
using $950 for P (present value) and solving for i in the general formula for calculating the
present value of an interest-income security, which is explained in the second example in the
Appendix.
The term structure is the relationship of average yields to maturity and time to maturity of debt
instruments. Each maturity class of securities, such as three years to maturity, five years to
maturity, etc., has an average yield to maturity, and the term structure shows how they relate to
one another. For example, a Treasury note with three years to maturity could have an average
yield to maturity of 4.0%, a Treasury note with five years to maturity could have an average yield
to maturity of 5.0%, and a Treasury note with 10 years to maturity could have an average yield to
maturity of 6.0%. The term structure of interest rates is often depicted on a chart by measuring
the average yield to maturity on the vertical (or Y) axis and time of maturity on the horizontal (or
X) axis. This graph of the term structure is sometimes referred to as “the yield curve.” The term
structure is drawn for debt instruments outstanding at a given point in time.9
ȱȱȱȱȱȱ ȱȱ
Financial economists have different theories concerning the causes of the term structure of
interest rates and the changes in the term structure over the business cycle.10 The four primary
theories are the expectations theory, the risk aversion theory, the segmented market theory, and
the preferred habitat theory.
Currently, the prevailing theory among economists is the preferred habitat theory, which is a
combination of the first three theories. According to the preferred habitat theory, the primary
determinant of the term structure of interest rates is expectations about interest rates.
The validity and policy value of any theory can be evaluated by determining how well a theory
explains relevant data and predicts outcomes. The implications of each theory of the term
structure of the interest rate can be compared with three real-world phenomena. First, the yield
curve is usually upward sloping indicating that the yield rises as the length of maturity increase.
For example, during the period 1950-2003, in 88% of the months, the yield curve was upward
sloping.11 Second, usually short-term, intermediate-term, and long-term yields rise or decline
together.12 Third, over business cycles, usually short-term and long-term interest rates move in a
pro-cyclical fashion—rising during expansions and declining during recessions. Furthermore,
short-term rates fluctuate more than long-term rates over business cycles.13
9 On a daily basis, the Treasury calculates rates for the yield curve for Treasury securities. These data are available at
http://www.treasury.gov/offices/domestic-finance/debt-management/interest-rate/, visited September 12, 2008.
10 These theories assume that monetary policy is constant. However, if the Fed instituted a tighter (easier) monetary
policy then the term structure curve would shift up (down).
11 Lloyd B. Thomas, Money, Banking, and Financial Markets (Mason, Ohio: South-Western, 2006), p. 136.
12 Ibid., p. 137.
13 Ibid., pp. 137-138.
ȱȱȱ
řȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
¡ȱȱ
The expectations theory was the initial theory explaining the term structure of interest rates. Most
financial economists believe that the primary determinant of the term structure of interest rates is
interest rate expectations. There are two basic expectations theories for the term structure: the
pure expectations theory and the error-learning theory.
ȱ¡ȱ¢ȱŗŚȱ
A model based on the pure expectations theory has the following four assumptions:
1. All securities are riskless with respect to the payment of both interest and principal.
Moreover, the theory abstracts from tax considerations, call features, differing coupons, and
other such “imperfections.”
2. Investors hold with complete confidence a set of uniform expectations concerning all
future short-term rates of interest, and their forecasts are accurate.
3. There are no transactions’ costs, and consequently, switches among securities of different
maturities are unimpeded.
4. The behavior of all market participants is motivated by no objective other than profit
maximization, thus ensuring full use of every opportunity for profitable arbitrage among
different maturities.15
These assumptions imply that all maturities are perfect substitutes for each other. Each investor
seeks to maximize the yield on a given security over a given holding period. The dollar yield on a
security for any given holding period equals the coupon interest received plus the capital gain (or
minus the capital loss). The average yield to maturity equals the coupon interest received plus the
capital gain (or minus the capital loss) divided by the purchase price of the security. Because
securities of different maturities are perfect substitutes, for a given holding period, an investor
may purchase a security with a maturity equal to the length of the holding period, a series of
shorter-term securities with a combined maturity equal to the holding period, or a security with a
longer maturity than the holding period, which can be sold at the end of the holding period.
Furthermore, an investor may purchase any combination of different maturities.
In equilibrium, due to arbitrage, any combination of maturities should have the same average
yield to maturity for any holding period. For example, an investor would be indifferent between
purchasing a series of 90-day Treasury bills (and reinvesting the principal and interest) and
purchasing a 30-year Treasury bond (and reinvesting the interest). As a simplified case, assume
the 30-year bond is sold at par.16 Hence, the average yield to maturity on the 30-year bond would
14 Professor Irving Fisher developed the relationship between short-term and long-term rates of interest under
conditions of perfect certainty which was used by Professor J. R. Hicks in formulating the initial pure expectations
theory in 1939. The variation of the pure expectation model discussed in the section is based on the following source:
Friedrich A. Lutz, “The Structure of Interest Rates,” Quarterly Journal of Economics, vol. 55, 1940-1941, pp. 36-63.
15 Burton Gordon Malkiel, The Term Structure of Interest Rates (Princeton: Princeton University Press, 1966), p. 18.
16 The par value is the redemption value if the security is held to maturity.
ȱȱȱ
Śȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
be approximately equal to the arithmetic average of the yields on a series of 90-day Treasury bills
over the 30-year period.17
Under the pure expectations theory, different expectations by investors of short-term interest rates
would result in term structure curves with different shapes. In equilibrium, for a given sequence
of expected Treasury bill rates, there would be a unique term structure of interest rates. Since the
long-term yield is an average of current and expected short-term yields, the long-term yield can
never fluctuate as widely as the short-term yield.18 In Figure 1, four hypothetical term structure
curves are depicted that correspond to different expectations by investors of short-term interest
rates. If investors expect future short-term interest rates to be constant, then the term structure
curve would be flat as shown by curve AA. If investors expect future short-term interest rates to
rise, then the term structure would be upwards sloping as shown by curve BB. If investors expect
future Treasury bill rates to decline over time, then the term structure curve would be downward
sloping as indicated by curve CC.
Figure 1. Hypothetical Term Structures of Interest Rates
Source: Adapted by CRS from Masera, p. 7.
Investors could expect that Treasury bill rates would vary over time. For example, investors could
expect Treasury bill rates to rise and then decline. These expectations could result in a humped
term structure curve such as curve DD.
17 These average yields to maturity are approximately equal because their calculations do not consider compounding.
18 R. S. Masera, The Term Structure of Interest Rates (Oxford: the Clarendon Press, Oxford, 1972), p. 14.
ȱȱȱ
śȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
The applicability of the pure expectations theory was criticized because of its underlying
assumptions. Some critics argued “that investors are simply not capable of predicting interest
rates for periods far into the future.”19 Early empirical studies tested the correspondence between
expected short-term rates and the later, actually observed short-term rates. These tests were
mostly negative in their results, which means that the theory is not supported by the evidence.20
But Professor David Meiselman correctly argued “that expectations need not be realized in order
for them to determine the rate structure in the manner asserted by the theory....”21 In other words,
investors purchase securities based on their expectations of yields to maturity rather than actual
yields to maturity.
Ȭȱ¢ȱ
Professor Meiselman formulated an error-learning theory which allowed unexpected changes in
interest rates to cause revisions of future interest rate expectations. For example, if actual short-
term rates turn out to be higher than expected short-term rates, then “the market will
systematically revise upwards expectations concerning short-term rates in the future.”22 Since
long-term rates are averages of present and future short-term rates, the entire term structure would
rise.23
Professor Meiselman maintained that his empirical tests supported his error-learning theory. But
critics disputed the accuracy of his data and his econometric procedures. Empirical tests by other
economists using more recent data failed to support Meiselman’s theory.24
¢ȱ ȱȱ
The pure expectations theory indicates that the average yield spread over a period of years should
be zero, which is inconsistent with the data showing average yields rising as the term to maturity
increases.25 The pure expectations theory assumes that various maturities are close substitutes for
one another; consequently, this theory correctly predicts that the yield of various maturities will
typically move in the same direction. Furthermore, according to the pure expectations theory,
short-term rates should always fluctuate more than long-term rates. But Treasury bond rates
sometimes have changed more than Treasury bill rates. Finally, the pure expectations theory
correctly predicts that interest rates on different maturities move in a pro-cyclical fashion over the
business cycle.26
19 Malkiel, p. 23.
20 Ibid., p. 28-29.
21 Malkiel, p. 35.
22 Ibid., p. 30.
23 Ibid., p. 31.
24 Ibid., p. 34-35.
25 Thomas, p. 137.
26 Ibid., p. 139.
ȱȱȱ
Ŝȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
ȱȱ¢Řŝȱ
Some critics argue that the expectations theories are incorrect because they first assume that all
securities are riskless. These critics maintain that securities contain risk and that most investors
are risk averse. The prevailing view of financial economists is that with time, risk increases but at
a decreasing rate.28 Hence, investors must receive a risk premium for purchasing securities with
longer maturities, but the amount of this risk premium increases more slowly than does the length
of the time to maturity. Thus, under the risk averse theory, the yield consists of the interest rate
under the pure expectation theory plus a risk premium.
¢ȱȱȱ
There are four major types of risk, but these types of risk tend to be interrelated. All of these types
of risk tend to increase with the length of the forecasting period. These types of risk are default
risk, interest rate risk, inflation risk, and political risk.
ȱȱ
Default risk is the probability that all promised payments of interest and repayments of principal
will not be fully paid (or paid at all). U.S. Treasury securities are often said to be free of default
risk, but this may not be the perception of every investor, so this risk could play some role.29
ȱȱȱ
An unexpected rise in interest rates causes the market value of fixed income securities (including
Treasury securities) to decline. For a given unexpected rise in the interest rate, the decline in the
market value of the security increases as the length of time to maturity rises. Conversely, an
unexpected decline in interest rates causes a rise in the market value of fixed income securities
and this rise increases with the length of time to maturity. Thus, a buyer of fixed income securities
assumes the risk that his interest rate forecast may be too low and the magnitude of this risk
increases as the time to maturity increases.30
ȱȱ
A change in the expectation of future inflation tends to drive up interest rates and thus lower the
market value of fixed income securities. Investors seek a “real” rate of return on their principal.
They deduct the expected change in the cost-of-living from the nominal rate of return to
determine the “real” rate of return. An increase in the expectation of future inflation causes
investors to demand a higher nominal interest rate, particularly on long-term securities. The
27 Most financial economists only discuss interest-rate risk and refer to this theory as the liquidity preference theory.
This author has broadened the theory to cover four types of risk.
28 For most securities, all four types of risk, described in the following pages, would increase at a decreasing rate. But
U.S. Treasury securities usually are assumed free of default risk.
29 “ The Junkification of American T-Bonds,” The Economist, vol. 311, no. 7,604; May 27, 1989, pp. 77-78.
30 For examples of interest rate risk, see Appendix.
ȱȱȱ
ŝȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
longer the period of forecasting, all other things being equal, the more difficult it is to accurately
forecast the rate of inflation.
ȱȱ
New government rules, regulations, laws, and court decisions or changes in existing ones may
adversely affect the market value of securities, including Treasury securities. For example, higher
marginal tax rates on interest income may encourage investors to shift funds into tax-free
municipals or growth stocks, which are expected to earn more lightly taxed capital gains.
Although actual default by the U.S. Treasury is almost unthinkable, there may have been a fear
that the federal government would monetize more of the debt and thus cause unexpected inflation
which would, in turn, lower the “real” value of the national debt and interest payments.31 Some
financial experts argue that approval of current proposals to provide extensive federal credit
assistance to financial institutions may result in the monetizing of more federal debt and thus
cause unexpected inflation.
Professors Alessandro Missale and Olivier Jean Blanchard found that during the period 1960
through 1989, three OECD nations had debt-to-GDP ratios which rose more than 100%. These
three nations were Ireland, Italy, and Belgium. As their debt-to-GDP ratios rose further and
further above some threshold (debt of 40% to 50% of GDP), these countries reduced the average
maturity of their publicly held debt. The authors provide a tentative explanation that as the debt-
to-GDP ratio rose investors became more fearful that their governments had a greater incentive to
inflate away the real value of their national debts. Investors were increasingly reluctant to
purchase government securities of longer maturities. Hence, in order to sell debt, these
governments had to sell shorter maturities, and consequently, the average maturity of outstanding
debt declined.32
Political risk is more pronounced for private debt instruments. For example, more stringent
pollution controls required by the Environmental Protection Agency could adversely affect the
value of a particular corporation’s bonds. If the corporation’s output causes significant pollution
then more stringent pollution controls would raise the production costs of the corporation, lower
profits, and thus increase the probability that the corporation would default on its bonds.
¢ȱ ȱȱ
Most financial economists believe that most investors are risk averse. Usually the term structure
of interest rates is upwards sloping, which is consistent with the risk averse theory. But, if risk
aversion were the sole explanatory factor for the term structure and risk is always monotonically
related to time then the term structure curve would always be upwards sloping, which it is not.
Hence, most financial economists believe that risk aversion explains only part of the shape of the
term structure curve. The risk averse theory is consistent with yields of various maturities
31 For a current viewpoint on this issue, see Gerald P. O’Driscoll Jr., “Washington Is Quietly Repudiating Its Debts,”
(Opinion), The Wall Street Journal, vol. 252, no. 45, August 22, 2008, p. A15.
32 Alessandro Missale and Olivier Jean Blanchard, “The Debt Burden and Debt Maturity,” The American Economic
Review, vol. 84, no. 1, March 1994, pp. 309-319.
ȱȱȱ
Şȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
typically moving in the same direction.33 The risk averse theory also is consistent with the yield
curve exhibiting a regular cyclical pattern.34
ȱȱ¢ȱ
Some critics of the expectations theory argue that it is incorrect because of its assumptions that all
debt instruments are perfect substitutes and that there are no transactions’ costs. These critics
maintain that the loan market is segmented into compartments according to maturities. The
leading proponent of this segmented market theory is Professor J. M. Culbertson, who in 1957
totally rejected the expectations theory. He contended that the supply and demand in markets for
specific maturities determine their respective interest rates. According to Culbertson:
There are a variety of impediments to mobility of funds in debt markets: legal restrictions on
some types of borrowing and on debt holdings of institutional investors, desire of investors
for portfolio diversification, customary investment standards applied to financial institutions,
lags in establishment or revision of financial institutions, specialization of investors on
technical grounds, impediments to geographical movement of funds in cases requiring
judgment or administrative activity, etc.35
He stated that “lenders as much as borrowers generally insist upon a maturity that is related to the
purpose of borrowing.”36 A business prefers to borrow long-term in order to construct a plant.37
Life insurance companies invest funds from the sales of annuities in long-term securities in order
to minimize risk. Pension funds also prefer to invest in long-term securities in order to match debt
maturities with projected payouts. Hence, life insurance companies and pension funds have
limited needs for liquidity. But commercial banks must maintain a more liquid position because
of possible sudden large withdrawals by depositors. Government regulations concerning asset
holding reflect these differences in liquidity requirements.38
Furthermore, according to the pure expectations theory, short-term rates should always fluctuate
more than long-term rates. But Treasury bond rates sometimes have changed more than Treasury
bill rates. For example, during the period of January 1, 1994, through May 15, 1994, the 30-day
Treasury bill rate often changed on a daily or weekly basis by less than the 30-year bond rate.39
ȱȱȱȱ
Professor Culbertson correctly argued that the expectations theory and the segmented market
theory have different implications for the feasibility of using debt management as a policy
instrument. If the expectations theory is valid then a change in the composition of Treasury debt
sales would be neutralized by portfolio changes of investors and the term structure of interest
33 Thomas, p. 137.
34 Thomas, p. 138.
35 J. M. Culbertson, “The Term Structure of Interest Rates,” The Quarterly Journal of Economics, vol. 71, no. 4,
November 1957, p. 503.
36 Ibid., p. 494.
37 Ibid.
38 Malkiel, pp. 26-27.
39 Board of Governors of the Federal Reserve System, Federal Reserve Statistical Releases—Selected Interest Rates,
January 1, 1994-May 15, 1994.
ȱȱȱ
şȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
rates would be unchanged. This occurs in the expectations theory because debt instruments of all
different maturities are perfect substitutes. In contrast, the segmented market theory assumes that
debt instruments of different maturities are not perfect substitutes. Hence, the segmented market
theory suggests that changes in Treasury debt sales can affect the supply of debt in different
markets, and consequently, the term structure of interest rates.40 Professor Culbertson writes that
If used actively in a coordinated manner, monetary and debt management policies can play
an essential role in dealing with both inflationary and deflationary problems, by enforcing an
appropriate behavior on interest rates, through their impact upon conditions in debt markets,
and through their influence upon the liquidity position of the economy.41
Professor Culbertson did not present a specific testable model, but instead he presented concepts
based on extensive anecdotal evidence. Furthermore, data limitations made any possible
econometric testing difficult. Consequently, there has not been quality econometric testing, which
has either supported or rejected the segmented market theory.42
¢ȱ ȱȱ
The segmented market theory is based on institutional preferences for different maturities by
lenders and borrowers and is consistent with the prevalence of upward sloping yield curves.43 The
segmented market theory is inconsistent with data that show yields of various maturities typically
move in the same direction. Because financial markets for different maturities are distinct, a rise
in the interest rate for one maturity should not affect the interest rate on other maturities. Finally,
the segmented market theory offers a plausible explanation using institutional factors for the
cyclical pattern of the term structure of interest rate. For example, during a recession, loan
demand declines and banks put their excess funds into Treasury bills, which raises Treasury bill
prices and lowers their yield.44
ȱ
ȱ¢ȱ
Professors Franco Modigliani and Richard Sutch (M-S) argued that the expectations theory, the
risk averse theory, and the segmented market theory have merits and shortcomings. Consequently,
M-S advocated the blending of these three theories into a preferred habitat theory.45 They
described their preferred habitat theory as
basically an adaptation of the expectational theory of the structure of interest rates under
certainty to a world in which (1) future rates are in fact uncertain; (2) transactors, both final
wealth holders and final borrowers, have definite preferences as to the length of time they
want to keep their funds invested or for which they require financing (that is, they have a
preferred maturity habitat); and (3) types of transactors generally exhibit risk aversion, and
40 Culbertson, pp. 487-488.
41 Ibid., p. 517.
42 Masera, pp. 30-31.
43 Thomas, p. 137.
44 Thomas, p. 139.
45 Their model is an adaption of a model constructed by F. De Leeuw. For a presentation of this model, see F. De
Leeuw, A Model of Financial Behavior, Chapter 13 in J. Duesenberry, G. Fromm, L. Klein, and E. Kuh (eds.), The
Brookings Quarterly Econometric Model of the United States (Chicago: Rand McNally, 1965), pp. 494-503.
ȱȱȱ
ŗŖȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
hence, other things equal, would prefer to match maturities in their portfolios to their habitat
so as to be sure of the return or cost. In addition to final transactors, the model also
recognizes the existence of arbitragers, or intermediaries, prepared simultaneously to borrow
and lend in different maturities when the differences in expected returns provide sufficient
inducement to compensate for the risk involved in the operation.46
According to M-S, investors have some preferred portfolio of maturities, but they can be induced
to alter their portfolios if their returns from changing are sufficiently high.
Modigliani and Sutch conducted extensive empirical analyses using their model. They examined
the relationship between the yield on Aaa rated corporate bonds and the rate on commercial paper.
They concluded that
The expectation model can account remarkably well for the relation between short- and long-
term rates in the United States. Furthermore, the prevailing expectations of long-term rates
involve a blending of extrapolation of very recent changes and regression toward a long-term
normal level.
There is no evidence that the maturity structure of the federal debt, or changes in this
structure, exert a significant, lasting or transient influence on the relation between the two
rates.47
Modigliani and Sutch’s model and empirical results were harshly criticized by some economists.
Professor Neil Wallace argued that the M-S analysis has serious data problems and econometric
shortcomings.48 Professor Reuben A. Kessel believes that M-S should have employed more
sophisticated econometric techniques. He also argues that the M-S model is more of a market
segmentation model than an expectations model.49
¢ȱ ȱȱ
The preferred habitat theory is consistent with data indicating that usually the term structure of
interest rates is upward sloping, yields of various maturities usually move in the same direction,
and the yield curve exhibits a regular cyclical pattern.50
ȱ ȱȱȱȱ
Probably the prevailing view of financial economists is that some variation of the preferred
habitat model best explains the term structure of interest rates. But the relative importance of
46 Franco Modigliani and Richard Sutch, “Debt Management and the Term Structure of Interest Rates: An Empirical
Analysis of Recent Experience,” The Journal of Political Economy, vol. 75, no. 4, part 2 supplement, August 1967, p.
569.
47 Modigliani, Franco and Richard Sutch, “Innovations in Interest Rate Policy,” The American Economic Review, vol.
56, no. 2, May 1966, p. 196.
48 Neil Wallace, Comment, The Journal of Political Economy, vol. 75, no. 4, part 2 supplement, August 1967, pp. 590-
592.
49 Reuben A. Kessel, Comment, The Journal of Political Economy, vol. 75, no. 4, part 2, supplement, August 1967, pp.
592-593.
50 Thomas, pp. 137-139.
ȱȱȱ
ŗŗȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
expectations, risk aversion, and segmented markets is disputed among financial economists,
which makes the formulation of debt management policy controversial.
¢ȱȱ
Prior to the Administration of President Bill Clinton, the consensus of financial economists was
that federal debt management had one necessary objective and three alternative secondary
objectives.
ȱȱǻ¢ȱǼȱ
The Treasury must float debt maturities, which are salable if the federal government is to continue
to operate. Currently, the U.S. Treasury is experiencing no difficulty in selling any maturity, but a
change in debt marketing practices conceivably might create problems for the salability of debt—
at least in the short run. Furthermore, a large increase in the volume of debt to finance might
complicate the issuance of new debt. Subject to this necessary objective, the Treasury has
discretion in pursuing three secondary objectives.
ȱ¢ȱȱ
The pursuit of one of these secondary objectives may adversely affect another secondary
objective. Furthermore, these different objectives depend on different theories of the term
structure of interest rates.
ȱ£ȱ
Supporters of the economic stabilization objective maintain that the term structure of interest
rates is explained primarily by the segmented market theory. They argue that debt management
should be used to lessen macroeconomic fluctuations.
For instance, during a recession, the Treasury should lower the average maturity of the national
debt by issuing fewer long-term securities and more short-term securities. This reduced supply of
long-term securities would raise their price, and consequently, the long-term rate of interest
would fall. On the other hand, the increased supply of short-term securities would lower their
price and increase the short-term rate of interest. Since the long-term rate of interest has a more
pronounced effect on investment than does the short-term rate of interest, total investment would
increase, which would cause a multiple expansion of gross domestic product.
Conversely, according to this view, during a boom, the Treasury should raise the average maturity
of the debt by issuing more long-term securities and fewer short-term securities. This policy
would raise the long-term rate of interest, lower investment, and reduce the rate of inflation.
Some critics of countercyclical debt management argue that it would reduce the stability of
financial markets by causing uncertainty and direct competition with the private sector for long-
term funds during boom periods. Furthermore, opponents charge that the countercyclical policy
would raise interest costs on the national debt because the Treasury would sell more long-term
securities when interest rates are high and fewer long-term securities when interest rates are low.
ȱȱȱ
ŗŘȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
Other critics who support the expectations theory argue that Treasury debt management cannot
alter the term structure of interest rates, and consequently, cannot be used as a countercyclical
tool.
ȱ ȱȱ
Some financial economists recommend that the Treasury should attempt to minimize interest
costs on the national debt, which would require that the Treasury have some ability to forecast
interest rates. Under this policy, the Treasury should issue more long-term securities and fewer
short-term securities when interest rates are perceived to be low, often during recessions, in order
to lock in low interest costs. Conversely, when interest rates are perceived to be high, often during
booms, the Treasury should sell more short-term securities and fewer long-term securities, and
thus, reduce the level of future obligations to pay high interest rates on securities.
Critics argue that it is difficult for the Treasury to forecast interest rates. Furthermore, to the
degree that the financial markets are segmented, this policy of minimizing interest costs usually
would be procyclical since interest rates are typically “low” during recessions but “high” during
booms.
¢ȱ
Proponents of a neutrality policy for debt management argue that the Treasury should borrow
funds across the entire maturity structure with regularly scheduled sales. They assert that the
Treasury should attempt to make debt sales as predictable as possible in order to minimize
uncertainty in financial markets and consequently reduce interest costs. Hence, the Treasury
should not alter the composition of debt sales because of changes in the business cycle or changes
in interest rate forecasts.
A brief presentation of the history of debt management from 1945 through 1992 will provide a
better understanding of the pursuit of debt management objectives and a background to examine
debt management policies implemented by the Clinton Administration and the Bush
Administration. Previous sections of this report have largely discussed Treasury debt management
without considering policies of the Federal Reserve System, but this brief historical overview of
debt management demonstrates their interrelationship.
ȱȱǻŗşŚśȬŗşşŘǼȱ
From 1945 through March 1951, the Federal Reserve System continued its policy of pegging
interest rates which it had begun in March 1942.51 This policy initially imposed ceilings on
different interest rates ranging from 3/8 of one percent on 90-day Treasury bills to 2 ½% on 25-
year Treasury bonds.52 The Federal Reserve purchased without limitation all Treasury securities at
those prices, which ensured that their average yields to maturity did not rise above the pegs.
Under this pegging policy, the Fed abandoned control over the money supply. The structure of the
interest pegs was also the term structure of privately held marketable Treasury debt. As long as
51 Dudley G. Luckett, Money and Banking (New York, McGraw-Hill, 1976), p. 555.
52 Chandler and Goldfeld, pp. 555-556.
ȱȱȱ
ŗřȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
private investors were confident that the Fed would continue to peg rates, they preferred to hold
longer term securities because of their higher interest yields.53 Hence, from 1945 through March
1951, the average length to maturity of the privately held marketable Treasury debt was high, as
shown in Figure 2. These computations of the average length to maturity exclude Treasury
securities in Government accounts and in Federal Reserve banks.54
Figure 2. Average Length of Privately Held Marketable Debt
Source: Treasury Bulletin, March 2001, p. 29.
53 Ibid., p. 556.
54 U.S. Department of the Treasury, Treasury Bulletin, June 1994, p. 30.
ȱȱȱ
ŗŚȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
In 1950, a conflict between the Fed and the Treasury over pegging was discussed publicly. The
Fed wanted to discontinue pegging interest rates in order to be free to conduct an anti-inflationary
monetary policy. The Treasury wanted pegging of rates to continue in order to lessen interest
costs on the national debt and to maintain stable financial markets.55 On March 4, 1951, the
Treasury and the Fed jointly announced the accord, an agreement ending the pegging of interest
rates. The Fed gradually pursued an active monetary policy while assisting in maintaining orderly
markets for Treasury issues.56 As shown in Figure 2, after the accord, the average maturity of
privately held marketable Treasury debt declined.
After the accord, the Fed was concerned about the stability and strength of the Government
securities market; consequently a subcommittee of the Federal Open Market Committee was
assigned the duty of studying this market. This subcommittee found that the Government
securities market lacked “depth, breadth, and resiliency.”57 Hence, this subcommittee
recommended that the Fed conduct open-market operations by buying and selling only Treasury
bills. In 1953, the Fed adopted this bills only policy which it continued through 1961.58 As shown
in Figure 2, the average maturity of the privately held marketable Treasury debt fluctuated during
this period of bills only.
In the late 1950s, the Fed became concerned about the large persistent balance-of-payments
deficits. Because of the recession in 1960, the Fed also was concerned about the slow rate of
growth of the U.S. economy. In 1961, the Fed decided that the appropriate policy was to raise
short-term interest rates and simultaneously reduce long-term rates. The Fed believed that a rise
in the short-term interest rate would attract short-term capital to the United States and thus reduce
the U.S. balance-of-payments deficit. The Fed also believed that a reduction in long-term interest
rates would raise investment, and consequently, increase economic growth. From 1961 through
1965, the Federal Reserve with the assistance of the Treasury implemented this policy which was
called operation twist. The Fed simultaneously bought long-term Treasury securities and sold
Treasury bills. The Treasury altered the composition of its debt sales by raising the proportion of
debt sold which consisted of bills.59 Hence, the average length to maturity of the privately held
marketable Treasury debt rose from 1961 through 1965 as shown in Figure 2.
Modigliani and Sutch found that during operation twist that “the spread between rates on long-
term Government bonds and the bills rate declined from 150 base points down to 35 base
points.”60 But, they maintained that “...the spread typically tends to close in a period of recovery
and rising short-term rates....”61 They concluded that operation twist only slightly reduced the
spread between long-term and short-term rates.62 But other economists have argued that it is
difficult to assess the effects of operation twist because of its limited scale.63
55 Chandler and Goldfeld, pp. 560-561.
56 Ibid., p. 564.
57 Luckett, pp. 564-565.
58 Ibid., p. 565.
59 Luckett, pp. 566-568.
60 Modigliani and Sutch, p. 79.
61 Ibid.
62 Ibid., p. 116.
63 Chandler and Goldfeld, p. 581.
ȱȱȱ
ŗśȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
Federal Reserve swaps of shorts for longs were small, as were its net purchases of longer
maturities. Moreover, the Treasury offset these operations to some extent by issuing longer-
term securities.64
Nevertheless, the prevailing view of financial economists was that operation twist was a
disappointment. Hence, support for using debt management as a countercyclical tool diminished.
Beginning in 1965, high long-term interest rates prevented the Treasury from issuing bonds
because Congress had forbidden the Treasury from issuing bonds with a rate of return more than
4.25%.65 Consequently, the average yield to maturity of the privately held marketable Treasury
debt began declining sharply, as shown in Figure 2.
The Treasury was concerned about this rapid decrease in average maturity. In 1967, the Treasury
requested and received congressional permission to raise the maximum maturity on notes from
five years to seven years.66 There was no congressional ceiling on the rate of return on Treasury
notes. In March 1971, Congress enacted a limited exception to the 4.25% ceiling on Treasury
bonds.67 In 1973, the Treasury obtained congressional permission for an increase in the maximum
maturity on notes from seven years to ten years.68 Despite these changes, the average length to
maturity of Treasury debt continued to decline as shown in Figure 2.
From 1966 through 1974, the Treasury issued debt “on a sort of ad hoc basis.”69 The Treasury did
not announce in advance its plans for sales. Thus, potential buyers were faced with uncertainty
about which Treasury securities would be issued.
Beginning in FY1975, the Treasury was faced with the difficulty of financing persistently high
budgetary deficits. The Treasury was concerned about the stability of financial markets. The
Treasury decided to pursue a policy of neutral debt management while attempting to lengthen the
average length to maturity of the privately held marketable Treasury debt. According to this
policy of neutral debt management, the Treasury spread the Treasury’s borrowing needs widely
across the maturity spectrum. The Treasury had a regular and predictable schedule of debt
issues.70 The Treasury did not vary the composition of its debt issues based on interest rate
forecasts. This policy minimized investor uncertainty. In 1975, the Treasury started a program of
regularization of debt issuance with the scheduled cycle of sales of 2-year, 4-year, and 5-year
notes.71 The Congress expanded the quantity of Treasury bonds that could be sold without being
subject to the 4.25% interest rate ceiling. On November 10, 1988, the interest rate ceiling of
Treasury bonds was repealed.72 From December 1975 until February 1992, the average length to
64 Ibid.
65 Marcia Stigum, The Money Market (Homewood, Ill., Dow Jones-Irwin, 1990), p. 319.
66 Ibid., p. 320.
67 U.S. Department of the Treasury, Treasury Bulletin, June 1994, p. 30.
68 Stigum, p. 321.
69 Ibid., pp. 323-324.
70 For an analysis of this policy change, see Kenneth D. Garbade, “The Emergence of Regular and Predictable as a
Treasury Debt Management Strategy,” Economic Policy Review—Federal Reserve Bank of New York, March 2007, 19
p.
71 Stigum, p. 323.
72 U.S. Department of the Treasury, Treasury Bulletin, June 1994, p. 30.
ȱȱȱ
ŗŜȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
maturity of the privately held marketable federal debt gradually increased as shown in Table A-2
and Figure 2.
ȱȱȱ
During the Clinton Administration, the Treasury pursued three main debt management goals:
sound cash management—ensuring that Treasury cash balances are sufficient at all times, lowest
cost financing for the taxpayers, and the promotion of efficient capital markets.73 To achieve these
goals, the Treasury was guided by five interrelated principles: (1) maintaining the “risk-free”
status of Treasury securities, (2) maintaining consistency and predictability in issuing debt, (3)
ensuring market liquidity, (4) financing across the yield curve, and (5) employing unitary
financing (i.e., aggregating all of the U.S. Government’s financing needs and borrowing as one
nation).74
In its 1993 budget, the Clinton Administration announced a change in federal debt management.75
The Treasury would issue less long-term debt and more short-term debt. The Clinton
Administration maintained that shortening the debt maturity would reduce interest costs because
short-term securities usually have a lower interest rate than long-term securities.76 In early 1993,
the Treasury reduced the quarterly auction of 30-year bonds from $10 billion to $9.25 billion.77
On May 5, 1993, the Treasury announced a reduction in the sale of 30-year bonds from quarterly
sales of $9.25 billion or $37 billion annually to semiannual sales of $11 billion or $22 billion
annually.78 The Treasury also announced the discontinuation of the issuance of 7-year notes,
which had been amounting to about $40 billion per year.79 The Treasury replaced these 7-year
notes with additional sales of bills and 2-year and 3-year notes.80
It is not clear whether the Administration considered possible effects on the term structure of
interest rates. Ms. Deborah Danker, the Assistant Secretary of Domestic Finance, stated that the
Treasury was “not attempting to manipulate the yield curve or drive down long-term rates.”81 Mr.
Frank Newman, Undersecretary, said that by borrowing less at longer maturities, the Treasury
was not trying to reduce long-term market rates.82 But, if that happened, he reportedly said, it
“would not break my heart.”83 There is some anecdotal evidence that, at least in the short-run, the
73 Testimony of Gary Gensler, Assistant Secretary of Financial Markets, U.S. Department of the Treasury, House
Committee on Ways and Means, Hearing on Managing the Public Debt in an Era of Surpluses, June 24, 1998, p. 1.
74 Ibid., pp. 1-2.
75 During the Clinton Administration, changes in the average length of marketable interest-bearing debt securities held
by private investors are shown in Figure 2 and Table A-2.
76 D. Keith Sill, “Managing the Public Debt,” Business Review, Federal Reserve Bank of Philadelphia, July/August
1994, p. 3.
77 U.S. Congressional Budget Office, Federal Debt and Interest Costs (Washington, GPO, May 1993), p. 67.
78 Ibid., p. 70.
79 Ibid.
80 Ibid.
81 Stephen A. Davies, “Long Bond Sales Will Be Slashed By Nearly Half In Treasury Plan,” The Bond Buyer, vol. 304,
no. 29,146; May 6, 1993, p. 4.
82 David Wessel, “Savings from Treasury Borrowings Won’t Meet Projections, Official Says,” The Wall Street Journal,
vol. 221, no. 99, May 21, 1993, p. A2.
83 Ibid.
ȱȱȱ
ŗŝȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
debt management changes caused rates of interest on 30-year Treasury bonds to decline.84 In the
1993 budget, the Office of Management and Budget forecasted savings of $16 billion over fiscal
years 1994 through 1998 because of this new policy of issuing shorter debt.85 Actual savings were
below estimates, but a specific amount has not been estimated.
These changes in the composition of Treasury debt issues were criticized by many financial
analysts. They argued that OMB forecasted interest rate levels that could be too low, and
consequently, savings in interest costs could be overstated. After the Treasury announced the
changes in the composition of its debt sales, two major American corporations, Walt Disney and
Coca-Cola, sold bond issues with maturities of 100 years.86 Furthermore, at least a half dozen
corporations issued 50-year bonds, and many other corporations increased the average length of
their long-term debt.87 Unlike OMB and CBO, these corporations apparently judged that long-
term rates were low and decided to lock-in existing long-term rates. These corporations likely had
a different forecast of interest rates than OMB and CBO.
In May 1996, the Treasury began selling debt with a longer average maturity.88 This change in the
composition of debt issuance has contributed to the stabilization of the average length of the
marketable interest-bearing debt.
On January 29, 1997, the U.S. Treasury began issuing inflation-protection (price-indexed)
securities which were 10-year notes.89 The Treasury issued 5-year inflation-protection notes in
1997 and 30-year inflation-protection bonds in 1998. By the end of 1998, the Treasury established
a regular schedule of offerings of 5-year, 10-year, and 30-year inflation-protection securities.
The decline in the size of the deficit for fiscal years 1995, 1996, and 1997 caused a reduction in
the rate of growth of the national debt, and consequently, a decrease in the rate of growth of the
volume of debt issuance. This decrease combined with the May 1996 decision to sell debt with a
longer average maturity and the issuance of inflation-protection securities stabilized the average
length of marketable interest-bearing public debt. The debt management changes of the Clinton
Administration implemented in February 1993 would have eventually shortened the average
length of maturity of the Treasury from five years and eleven months in February 1993 to
approximately four and one-half years.90
The budget savings forecast by OMB from February 1993 through May 1, 1996 were
substantially overstated because interest rates over most of this period exceeded those forecast by
84 Lauricella and Thomas T. Vogel, Jr., “Many Treasury Prices Rise on More Speculation About Possible Changes in
U.S. Borrowing Mix,” The Wall Street Journal, April 18, 1993, p. C17.
85 Ibid.
86 Thomas D. Lauricella and Constance Mitchell, “Coca-Cola Joins Disney at the Very Long End With the Sale of $150
Million of 100-Year Bonds,” The Wall Street Journal, vol. 222, no. 16, July 23, 1993, p. C15.
87 Allen Sloan, “Unlike Uncle Sam, Disney Sees Nothing Goofy in Long-Term Bonds,” The Washington Post, July 27,
1993, p. D3.
88 U.S. Department of the Treasury, Treasury News, May 1, 1996, pp. 1-3.
89 CRS Report 97-134, Treasury Inflation-Protection Securities: Description, Goals, and Policy Issues, by James M.
Bickley, p. 1. (This report is out-of-print and available from the author.)
90 Jerome H. Powell, “More Vulnerable to Inflation,” The Wall Street Journal, vol. 221, no. 98, May 20, 1993, p. A16.
ȱȱȱ
ŗŞȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
OMB.91 OMB’s forecast of budget savings after May 1, 1996 were not relevant because of policy
changes in debt management.
Since short-term rates are usually lower than long-term rates, a lower average length of maturity
that occurred between February 1993 and January 1995 may have reduced interest costs in the
long run.92 But, as the average length of maturity of the debt declines, the Treasury must refinance
the debt more frequently, and consequently, yearly interest costs will be more volatile and
uncertain.
Beginning with FY1998, the federal government began running budget surpluses; that is, the flow
of revenue into the U.S. Treasury exceeded the outflow of expenditures. The Treasury lowered
the amount of outstanding publicly held debt by reducing new debt issuance as existing federal
debt issues matured. In addition, on March 9, 2000, the Treasury conducted its first buyback
operation of outstanding Treasury securities before maturity.93 For fiscal years 1998-2001, budget
surpluses caused a decline in the level of publicly-held debt. Because fewer short-term securities
were issued, the average length rose from five year and five months at the end of FY1997 to six
years and one month at the end of FY2001.94
ȱȱȱ
For the Bush Administration, the stated debt management objective is the “lowest cost of
financing over time.”95 The Treasury has four constraints in achieving this objective:
• Uncertainty: forecast errors, legislation, etc. all create uncertainty in deficit
forecasts, debt limit problems.
• Size: Treasury is too large to behave opportunistically.
• Fluctuations in non-marketable debt: savings bonds, state and local government
securities.
• Short-term balances: adequate cash balances to cover expenses.96
The Treasury maintains that the goal of lowest cost over time implies a diversified debt portfolio
that spread debt across maturities in order to reduce risk, diversity the investor base, improve cash
management and facilitate regular and predictable issuance.97 The Treasury argues that it should
91 For OMB’s interest rate forecasts, see U.S. Congressional Budget Office, An Analysis of the President’s February
Budgetary Proposals (Washington, GPO, March 1993), p. II-2.
92 The actual effects on long term rates has not been estimated.
93 For an analysis of paying down the federal debt, see CRS Report RS20302, Paying Down the Federal Debt: A
Discussion of Methods, by James M. Bickley. For an examination of the buybacks of Treasury debt during 2000-2001,
see Kenneth D. Garbade and Matthew Rutherford, Buybacks in Treasury Cash and Debt Management, Federal Reserve
Bank of New York Staff Report no. 304, October 2007, 65 p.
94 See Table A-2 and Figure 2.
95 Karthik Ramanathan, Overview of U.S. Treasury Debt Management, Office of Debt Management, U.S. Department
of the Treasury, June 2008, p. 3. Available at http://treas.gov/offices/domestic-finance/debt-management/, visited
August 26, 2008.
96 Ibid., p. 4.
97 Ibid., p. 6.
ȱȱȱ
ŗşȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
not interfere with price discovery, in other words, the price of a security should be determined
through basic supply and demand factors related to the market.98 The Treasury asserts that it is a
regular and predictable market participant and not a market timer.99
At the start of the Bush Administration in January 2001, the average length of market interest-
bearing public debt held by private investors was five years and nine months.100 During the two
terms of the Bush Administration, this average length of debt declined and was four years and six
months at the end of May 2008.101 Since short-term interest rates were below long-term rates, the
shorting of the debt by the Bush Administration resulted in an immediate savings in interest costs.
But, as the shorter is the average length of maturity of the debt, the more frequently the debt must
be refinanced, and consequently, yearly interest costs will be more volatile and uncertain. During
much of the Bush Administration, interest rates were relatively low in comparison to interest rates
over the past 35 years. Hence, the Treasury may have failed to “lock in” low interest rates by
limiting the sale on longer term securities.102
The complexities of federal debt management are indicated in the following statement by James
Clouse, then Acting Deputy Assistant Secretary for Federal Finance in 2006:
The pragmatic approach [to debt management] focuses largely on minimizing the cost
of debt financing over time while also giving some weight to other factors such as
variability in interest payment, the diversification of the investor base, operational risk,
and flexibility. To some extent, the pragmatic approach is amenable to standard analysis
of decision-making under uncertainty. For example, given the presence of a term
premium in longer-term Treasury yields, a strict cost minimization criterion would
probably favor—at the margin—shorter-term financing. On the other hand, yields on
longer-term securities are less variable over time than short-term yields and longer-term
securities can be employed to lessen rollover risks. An optimal debt portfolio might then
involve some appropriate balancing of these considerations.
While a useful exercise, there is a fundamental difference between this standard sort of
analysis and Treasury debt management in practice: the textbook exercise assumes that
Treasury is a price-taker in financial markets, but in reality the strategies and tactics
employed by Treasury debt managers can affect the market prices of Treasury securities.
It is this interplay of market expectations and debt management policy that makes
Treasury debt management more subtle and nuanced than a simple optimal cash
management exercise.103
98 Ibid., p. 7.
99 Ibid., p. 8.
100 U.S. Department of the Treasury, Treasury Bulletin, June 2001, p. 27.
101 U.S. Department of the Treasury, Treasury Bulletin, September 2008, p. 25.
102 See Table A-2 for data on the maturity distribution and average length of marketable interest-bearing public debt
securities held by private investors.
103 James Clouse, Acting Deputy Assistant Secretary for Federal Finance, U.S. Department of the Treasury, Statement
before the Information Management Network, April 19, 2006, p. 3. He is currently employed at the Federal Reserve
Board.
ȱȱȱ
ŘŖȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
¡ǯ ¡ȱȱ ȱȱŗŖŚȱ
The purposes of this appendix are to provide examples of interest rate risk and to demonstrate that
interest rate risk increases as the time to maturity increases. The present capital value of an
income producing asset is calculated by a process called capitalization, which involves
discounting the expected flow of money receipts. A dollar received in the future is worth less than
a dollar today because the dollar today can be immediately invested and earn interest
compounded over time. An investor generally would not buy a security unless the yield on the
purchase price is as high as that available on other comparable securities. These concepts are
demonstrated by the following examples of three different maturities: 1-year, 10-year, and
perpetuity.
First, assume a potential investor is offered a 1-year security which promises to pay $1,060 in one
year. The general formula for simple discount is shown by the following equation:
P = A/(1+i)
where P is the present value, A is the dollar amount receivable at the end of the interest period,
and i is the rate of interest for that period stated in hundredths. In this example, if the prevailing
market yield is 6 percent per year, that is, i = 0.06, then the investor would be willing to pay
$1,000, as shown in the following calculation:
P = $1,060/(1+0.06) = $1,060/(1.06) = $1,000
If the prevailing market yield had been 8 percent then the investor would have paid only $981.48,
which was calculated as follows:
P = A/(1+i) = $1,060/(1+0.08) = $981.48
If an investor purchases the 1-year security when the going market yield is 6 percent and the
going market yield rises immediately to 8 percent then the market value of the 1-year security
would decline from $1,000 to $981.48. This decline of $18.52 is 1.852 percent of the initial
purchase price of $1,000.
If the prevailing market yield had been 4 percent then the investor would have paid $1,019.23,
which was calculated as follows:
P = A/(1+i) = $1,060/(1+0.04) = $1,019.23
If an investor purchases the 1-year security when the going market yield is 6 percent and then the
going market yield immediately declines to 4 percent then the market value of the 1-year security
would rise from $1,000 to $1,019.23. This rise is $19.23 or 1.923 percent of the initial purchase
price.
Second, assume a potential investor is offered a 10-year security that promises to pay $60 at the
end of each year for ten years plus $1,000 at the end of the tenth year. In addition, assume that the
market rate of interest is 6 percent.
104 This appendix is based closely on the following source: Chandler and Goldfeld, pp. 59-64.
ȱȱȱ
Řŗȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
The general formula for calculating the present value of an interest-income security is the
following:
P = A1/(1+i) + A2/(1+i)2 + A3/(1+i)3 + . . . + An/(1+i)n + F/(1+i)n
where P is the present value of the security, the A’s are the dollar amounts of interest received at
the ends of each interest period, F is the amount of principal repayment, i is the rate of interest at
which the amounts are discounted, and n is the number of interest periods.
Hence, the present value of this 10-year security would equal $1,000, which was calculated as
follows:
P = $60/(1+0.06) + $60/(1+0.06)2 + $60/(1+0.06)3 + . . . + $60/(1+0.06)10
+ $1,000/(1+0.06)10 = $1,000
If an investor purchases the 10-year security when the interest rate is 6 percent and then the
interest rate immediately rises to 8 percent, then the market value of the security would fall from
$1,000 to $865.80, which was calculated by using 0.08 for i in the formula. This decline of
$134.20 is 13.42 percent of the initial purchase price of $1,000.
If an investor purchases the 10-year security when the interest rate is 6 percent and then the
interest rate immediately declines to 4 percent then the market value of the security would rise
from $1,000 to $1,162.22, which was calculated by using 0.04 for i in the formula. This increase
of $162.22 is 16.222 percent of the initial purchase price.
Third, assume that a potential investor is offered a security that pays $60 per year in perpetuity.
The formula for calculating the present value of a perpetuity is the following:
P = A/i
where P is the present value of the security, A is the dollar amount of interest received at the ends
of each interest period, and i is the rate of discount. Thus, in this example, the present value of the
security would be $1,000 which was calculated as follows:
P = A/i = $60/0.06 = $1,000
If an investor purchases this perpetuity when the interest rate is 6 percent and then the interest
rate rises to 8 percent then the market value of the security would fall from $1,000 to $750
($60/0.08). This decline of $250 is 25 percent of the initial purchase price of $1,000.
Conversely, if an investor purchases this perpetuity when the interest rate is 6 percent and then
the interest rate declines to 4 percent then the market value of the security would rise from $1,000
to $1,500 ($60/0.04). This rise of $500 in the value of the security would equal 50 percent of the
initial purchase price.
Table A-1 summarizes these three examples. These examples demonstrate that, for a given
change in the interest rate, the present value of an outstanding security would vary more as the
term to maturity increases, all other things being equal.
ȱȱȱ
ŘŘȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
Table A-1. Example of Present Values
Present Value If Discounted at
Description of Debt
6%
4%
8%
An obligation to pay $1,060 at the end of the year
$1,000.00
$1,019.23
941.48
An obligation to pay $60 annually for ten years and $1,000 at the
1,000.00 1,162.22 865.80
end of ten years
An obligation to pay $60 a year in perpetuity
1,000.00
1,500.00
750.00
Source: Chandler and Goldfeld, p. 63.
Table A-2. Maturity Distribution of Marketable Interest-Bearing Public Debt
Securities Held by Private Investors
(millions of dollars)
Maturity
Average
Class
Lengtha
Fiscal Amount Outstanding,
Year
Privately Held
Within 1
1 to 5 5 to 10 10 to 20 20 Years
Year
Years
Years
Years and Over Years Months
1969
156,008
69,311
50,182
18,078
6,097
12,337
4
2
1970
157,910
76,443
57,035
8,286
7,876
8,272
3
8
1971
161,863
74,803
58,557
14,503
6,357
7,645
3
6
1972
165,978
79,509
57,157
16,033
6,358
6,922
3
3
1973
167,869
84,041
54,139
16,385
8,741
4,564
3
1
1974
164,862
87,150
50,103
14,197
9,930
3,481
2
11
1975
210,382
115,677
65,852
15,385
8,857
4,611
2
8
1976
279,782
150,296
90,578
24,169
8,087
6,652
2
7
1977 326,674 161,329
113,319
33,067
8,428
10,531
2
11
1978 356,501 163,819
132,993
33,500
11,383
14,805
3
3
1979 380,530 181,883
127,574
32,279
18,489
20,304
3
7
1980 463,717 220,084
156,244
38,809
25,901
22,679
3
9
1981 549,863 256,187
182,237
48,743
32,569
30,127
4
0
1982 682,043 314,436
221,783
75,749
33,017
37,058
3
11
1983 862,631 379,579
294,955
99,174
40,826
48,097
4
1
1984 1,017,488 437,941
332,808
130,417
49,664
66,658
4
6
1985 1,185,675 472,661
402,766
159,383
62,853
88,012
4
11
1986 1,354,275 506,903
467,348
189,995
70,664
119,365
5
3
1987 1,445,366 483,582
526,746
209,160
72,862
153,016
5
9
1988 1,555,208 524,201
552,993
232,453
74,186
171,375
5
9
1989 1,654,660 546,751
578,333
247,428
80,616
201,532
6
0
ȱȱȱ
Řřȱ
ȱȱDZȱǰȱǰȱȱȱǻŗşŚśȬŘŖŖŞǼȱ
ȱ
Maturity
Average
Class
Lengtha
Fiscal Amount Outstanding,
Year
Privately Held
Within 1
1 to 5 5 to 10 10 to 20 20 Years
Year
Years
Years
Years and Over Years Months
1990 1,841,903 626,297
630,144
267,573
82,713
235,176
6
1
1991 2,113,799 713,778
761,243
280,574
84,900
273,304
6
0
1992 2,363,802 808,705
866,329
295,921
84,706
308,141
5
11
1993 2,562,336 858,135
978,714
306,663
94,345
324,479
5
10
1994 2,719,861 877,932
1,128,322
289,998
88,208
335,401
5
8
1995 2,870,781 1,002,875
1,157,492
290,111
87,297
333,006
5
4
1996 3,011,185 1,058,558
1,212,258
306,643
111,360
322,366
5
3
1997 2,998,846 1,017,913
1,206,993
321,622
154,205
298,113
5
5
1998 2,856,637 940,572
1,105,175
319,331
157,347
334,212
5
10
1999 2,728,011 915,145
962,644
378,163
149,703
322,356
6
0
2000 2,469,152 858,903
791,540
355,382
167,082
296,246
6
2
2001 2,328,302 900,178
650,522
329,247
174,653
273,702
6
1
2002 2,492,821 939,986
802,032
311,176
203,816
235,811
5
6
2003 2,804,092 1,057,049
955,239
351,552
243,755
196,497
5
1
2004 3,145,244 1,127,850
1,150,979
414,728
243,036
208,652
4
11
2005 3,334,411 1,100,783
1,279,646
499,386
281,229
173,367
4
10
2006 3,496,359 1,140,553
1,295,589
589,748
290,733
179,736
4
11
2007 3,634,666 1,176,510
1,309,871
677,905
291,963
178,417
4
10
2008
(May)
4,162,323 1,618,739
1,329,756
718,171
333,602
162,056
4
6
Source: Adapted by CRS from Council of Economic Advisors, Economic Report of the President, Washington,
GPO, 2008, p. 329; and U.S. Treasury, Treasury Bulletin, September 2008, p. 25.
Notes: Through FY1976, the fiscal year was on a July 1—June 30 basis; beginning with October 1976 (FY1977),
the fiscal year is on an October 1—September 30 basis. 2008.
a. Treasury inflation-protected securities—notes, first offered in 1997, and bonds, first offered in 1998—are
included in the average length calculation from 1997 forward.
ȱȱ ȱ
James M. Bickley
Specialist in Public Finance
jbickley@crs.loc.gov, 7-7794
ȱȱȱ
ŘŚȱ