Academic journal article Journal of Money, Credit & Banking

Do Excess Holding-Period Returns Depend on the Composition of Outstanding Federal Debt?

Academic journal article Journal of Money, Credit & Banking

Do Excess Holding-Period Returns Depend on the Composition of Outstanding Federal Debt?

Article excerpt

This note revisits an old question, the impact of maturity composition of outstanding federal debt on financial markets. The subject is worth revisiting for two reasons. First, in 1993 the U.S. Treasury announced intentions to finance more of its debt with short-term instruments. This policy change derives partly from the Treasury's desire to reduce financing costs since the spread between short and long rates was at that time large by historical standards. Officials and financial pundits are quoted as saying that a benefit of the policy will be a reduction in long-term interest rates.(1) Second, the average maturity of outstanding federal debt has undergone two large swings since the Accord, declining from 5.2 years in April of 1953 to 2.7 years in January of 1976 and then rising to 5.7 years in February of 1991. The large swings in average maturity give us hope that if debt policy does affect the structure of asset returns, the relationship will be discernible.

Many past studies have examined the relationship between the average maturity of outstanding federal debt, or some other measure of debt structure, and the spread between short- and long-term interest rates. Although the portfolio balance theory, initially suggested by Tobin (1958, 1963) and expounded most recently by Friedman (1992), predicts that debt structure will affect the spread between short- and long-term interest rates if assets of different maturity lengths are imperfect substitutes in portfolios, empirical studies have generally been unable to detect any relationship between the short-long spread and debt structure.(2),(3) Various explanations have been offered for the apparent lack of relationship. Most obviously, except for a constant risk premium, investors could view securities of various maturities as perfect substitutes. Integrated world capital markets could also explain the results. With interest rates determined in world markets, changes in domestic asset supplies might not be of sufficient size relative to world markets to alter the structure of interest rates. Finally, it has been argued (Modigliani and Sutch 1967, Roley 1982) that changes in the maturity structure of federal debt might, over time, induce offsetting changes in the composition of private sector debt. However, lack of a long-run relationship does not rule out a short-run relationship as (if) private sector adjustments take place.

Part of the inability to detect a relationship between federal debt structure and the structure of asset returns may be the method of testing. Although some studies have estimated multiequation structural models of asset returns [Roley (1982) among others], Melino (1988) cites the severe data requirements and the restrictions that must be imposed on such models to make this approach tractable. He further notes that the results may be sensitive to the imposed structure. This study eschews a structural approach in favor of a more parsimonious reduced-form approach. If debt policy has any impact on financial markets, the policy ought to be detectable in a short-run response function.

Our measures of financial market response are the excess holding-period returns on Treasury notes and bonds of various maturity lengths (defined as the difference between the return over a one-month period on a bond or note of given maturity length and the return on a one-month Treasury bill). While past studies have examined the spread between short- and long-term interest rates, holding-period returns are the more appropriate measure of the immediate financial market response to changes in federal debt policy because they incorporate the capital gains or losses that accrue to holding longer term assets.(4) Federal debt policy is measured by the average maturity of outstanding federal debt (AVM).(5)

Our reduced-form model is based on the results found in Bollerslev, Engle, and Wooldridge (1988), Cutler, Poterba, and Summers (1990, 1991), Hall and Miles (1992), and Mankiw (1986). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.