There is no question that Peter Spiro is correct in pointing out that there was a mistake in my article on fiscal deficits and interest rates (Tanzi (1985)). At the time I received his comment I was already aware of the mistake because my colleague, Lans Bovenberg, had discovered it in connection with some additional work on interest rates that we were doing jointly. I had planned to acknowledge the mistake at the time that the paper reporting the new work would be issued. The mistake is a common one, but nevertheless an embarrassing one. When the data reflecting the deficit were copied, they simply were left with the negative signs that they would have if one were referring to fiscal balances, rather than fiscal deficits. I regret the error.

Since I wrote that paper (some time in early 1985), several articles have pointed out that changes in the policy rule and operating procedures governing the behavior of the U.S. Federal Reserve System make the kind of statistical analysis carried out in my paper, which took short-term interest rates as the dependent variable, somewhat suspect (see Dwyer (1985) and Kaufman and Lombra (1986)). If the authors of these articles are right, one should resist the temptation simply to change the sign of the deficit variable in the estimated equations of my paper. If the paper did not in fact prove that fiscal deficits caused an increase in short-term interest rates, it should not be interpreted as necessarily having proven that fiscal deficits lowered interest rates.¹ Despite the error made, I remain convinced that, if a relationship exists between interest rates and fiscal deficits, it must be a positive one. The main problem is to establish that relationship statistically. As Spiro writes, “… the empirical evidence for a positive effect of the deficit on the level of interest rates remains… elusive …”

Perhaps I could use the vehicle of this reply to report briefly on the additional work that I have been carrying out with Bovenberg on fiscal deficits and interest rates.² Part of this work is closely related to some of the points made by Spiro. I will limit my reporting to those points.

Just as Spiro did, we lagged the deficit variable by a couple of years. Furthermore, we used semiannual rather than annual data. The use of the lagged variable can prevent some of the endogenous feedback from the interest rates to the deficit.³ The use of the semiannual data may help to identify better the dynamics of the relationship. These changes eliminated the negative sign on the deficit variable and, for the 1960–84 period, provided a positive sign. That is, according to these results, fiscal deficits would increase short-term interest rates. These equations are not as good, however, as those originally reported in my 1985 article, and of course the questions raised by Dwyer (1985), Kaufman and Lombra (1986), and others about the validity of this approach apply to these later equations as well. The results of our work are not at variance with those reported by Spiro in his equations (2) and (3) and discussed in his comment. His observations on the role of international capital flows and on the monetary variable are certainly pertinent.

Spiro also discusses an article by Hoelscher (1986) that, in Spiro’s words, “does not claim to show anything more than that a higher deficit raises bond yields relative to short-term interest rates.” Spiro concludes, however, that “… findings discussed here also place Hoelscher’s results in a new light. Deficits do raise long-term interest rates relative to short-term rates, but at the same time they lower the level of short-term interest rates. Therefore it is difficult to infer from regressions of that type by how much deficits have raised the absolute level of the long-term interest rate.” Actually, it was in the determination of long-term rates that we found the most interesting results in our current work on interest rates.

For a variety of reasons, interest rates on long-term bonds are much more stable than those on short-term bills. The specific framework used in my 1985 article to try to explain the behavior of short-term rates could not be used, without major modifications, to explain the behavior of long-term rates (see Tanzi (1985, pp. 570–71)). For long-term rates, short-term economic fluctuations (as reflected in changes in the annual gap variable) could not be expected to play much of a role. Furthermore, only expectations of inflation over long periods, and only fiscal deficits projected over several future years, would be relevant. As I wrote in 1985 (p. 571), “…the fiscal deficit of a particular year, unless it was seen to reflect a more permanent stance of fiscal policy, should not have a large influence on bonds that cover much longer periods.” This statement implies that long-term rates must be related to the rate of inflation, expected over several future years, as well as to the fiscal deficits, also expected over several future years.

In the work with Bovenberg, we used as the independent variables ten-year observed inflation forecasts and five-year averages of forward-looking, cyclically adjusted deficits measured as shares of trend GNP; the dependent variable was the rate of interest on ten-year U.S. Treasury securities and, for some other equations, the rate of interest on three-year U.S. Treasury securities. For the 1961–85 period we found surprisingly good results: the coefficients of the variables for expected inflation and for the fiscal deficit were positive and highly significant (at the 1 percent level), and the adjusted coefficients of determination (${\overline{R}}^2$) were very high. Thus the statistical results did seem to support the prevalent view that the fiscal deficit has a positive effect on long-term interest rates. Similar results were also obtained, with the required adjustments, in connection with three-year bonds.⁴ The importance of these results derives from the fact that, unlike Hoelscher’s findings, they do not refer to the behavior of long-term bonds in relation to that of short-term bills. Thus the results are immune to Spiro’s criticism of Hoelscher’s results. Our results indicate that fiscal deficits increase long-term interest rates in an absolute rather than a relative sense.

In conclusion, I wish to thank Peter Spiro for a useful comment,

In statistical matter (except in the resumes and restimenes) throughout this issue,

• Dots (…) indicate that data are not available;

• A dash (—) indicates that the figure is zero or less than half the final digit shown, or that the item does not exist;

• A single dot (.) indicates decimals;

• A comma (,) separates thousands and millions;

• “Billion” means a thousand million, and “trillion” means a thousand billion;

• A short dash (-) is used between years or months (for example, 1981–83 or January-October) to indicate a total of the years or months inclusive of the beginning and ending years or months;

• A stroke (/) is used between years (for example, 1981/82) to indicate a fiscal year or a crop year;

• Components of tables may not add to totals shown because of rounding,