IV Issues in Domestic Public Debt Management
- Manmohan Kumar, and Pablo Guidotti
- Published Date:
- March 1991
The above analysis has emphasized the key role of fiscal adjustment in dealing with the crisis facing the G-15 countries. This section examines a number of specific debt-management strategies which could complement fiscal adjustment. These strategies should be viewed in the context of developments noted in Section II, where it was shown that the growth of domestic public debt in highly indebted developing countries was associated with a sharp increase in inflation. Since 1982, price stabilization programs have been undertaken in a number of these countries. In some cases, programs have been successful in temporarily reducing inflation but, soon thereafter, new and sometimes virulent outbursts of inflation have occurred. In part, the lack of success has resulted from difficulties faced by the countries’ authorities in making stabilization efforts credible. This lack of credibility has meant that real interest rates often remained extremely high in the early stages of the adjustment programs, casting further doubt on their sustainability. The uncertainty associated with repeated failures to stabilize inflation rates has further increased the likelihood of runs on domestic assets. It is not growth in domestic debt per se that is responsible for this vicious circle, but rather the underlying fiscal situation that is reflected, in part, in the evolution of domestic debt. This section examines how the characteristics of domestic debt contribute to the developments mentioned above and how domestic debt management—in particular, debt indexation and changes in the maturity structure of debt—can contribute to achieving a better outcome.
I. Nominal Versus Indexed Debt
In most countries, domestic public debt is issued in nominal terms. Put simply, this means that government liabilities are denominated in domestic currency and the amortization schedule is not explicitly indexed to reflect changes in the purchasing power of money. Because its real value is affected by changes in the price level, the presence of nominal debt has important implications for the effects of inflation on the fiscal situation. This subsection examines the connection between nominal debt and inflation and reaches three main conclusions: (1) inflation affects the real value of nominal debt only if it is unanticipated; therefore, persistent inflation should not be expected, in principle, to reduce the debt burden; (2) the existence of nominal debt may exacerbate inflation, because the public recognizes that policymakers may have an incentive to tolerate inflation and consequently demands a corresponding inflation premium; and (3) debt indexation may contribute to reducing inflationary expectations and, thus, enhance the credibility of stabilization efforts by reducing the danger of high real interest rates.
a. Nominal Public Debt and Inflation
Since, superficially at least, the real value of nominal government debt could be reduced by an increase in the price level, it is worth examining the extent to which inflation might be regarded as “effective” in reducing the debt burden. Undertaking an analysis of this issue requires one to recognize that public expectations about inflation play a crucial role in determining this “effectiveness.” Furthermore, it should be emphasized that even if it appears that the debt burden might be reduced through inflation, inflation has other significant adverse allocative and distributional consequences which have to be taken into account. For instance, because of the uncertainties it engenders, inflation may have considerable adverse effects on investment and growth. Also, because of fiscal lags, government revenues may be significantly reduced in real terms by inflation.
Inflation reduces the real debt burden only if it takes the public by surprise, for if inflation is anticipated, bondholders will require a nominal return to compensate them for the expected capital loss associated with the fall in the real value of nominal debt. In fact, to the extent that inflation is anticipated, the nominal interest rate will rise to compensate bondholders point for point so that, in the end, the real return actually paid on the public debt will be the same as it would have been if no inflation had occurred. In such a case, inflation has no direct effect on the debt burden. Thus, inflation has an effect only if it results in a negative ex post real rate of interest or, at least, in a real rate of return lower than that available in international financial markets.31
If policymakers expect that inflation will not be anticipated by the public, they may be tempted to put aside their misgivings concerning its adverse effects in an effort to lower the real value of public debt. Given the serious costs of inflation, such a policy is very much a second-best approach, but it may nevertheless be considered attractive because it enables policymakers, in the short run, to avoid raising taxes or reducing expenditures.
The extent to which high levels of nominal public debt may encourage policymakers to tolerate inflation, rather than undertaking fiscal adjustment, can be illustrated by the following example. Consider the primary fiscal improvement that would be required to match a given reduction in the real value of the stock of nominal debt achieved by a once-for-all and unanticipated increase in the price level. Consider, for instance, a once-for-all and unanticipated increase in the price level of 10 percent, with a stock of nominal debt of 25 percent of GDP. This increase generates a fall in the real value of domestic debt of 2.3 percent of GDP (i.e., 25 × (0.1 ÷ 1.1)). This is equivalent to the fall in real domestic debt which would be obtained by a primary fiscal improvement of 0.6 percent of GDP per year over a period of five years.32 Similarly, the same increase in the price level, when the debt-to-GDP ratio is 100 percent, has an effect equivalent to a fiscal improvement of 1.2 percent of GDP per year sustained for ten years.
This example shows why a large stock of nominal debt may provide a temptation to tolerate inflation.33 Precisely because of this temptation, a large stock of nominal government debt may be destabilizing. This may be particularly true when the government’s ability to enforce a credible stabilization program is questionable. When the public doubts that stabilization will be successful, a vicious circle may result from the interrelationship between nominal debt, interest rates, and inflation. Given a large stock of nominal public debt, and to the extent that the public anticipates future inflation, nominal interest rates will rise. But this increase in interest rates will increase the debt burden even further, making inflation appear even more “attractive” to the policymaker although, in the final analysis, it will be much more costly to the economy, and to society generally, than alternative policies such as tax increases or fiscal tightening.
Furthermore, unless there is an active interest rate policy, the economy could end up not having a nominal anchor.34 This is so because depending on whether interest rates are freely determined by the market, an economy can be consistent with entirely different inflation rates. On the one hand, suppose that the public expects a high rate of inflation, leading to high nominal interest rates. If the government attempts to achieve low inflation, then the cost of debt servicing may be too high (because of the associated high real interest rate). Consequently, policymakers may be forced to take the high-inflation route, thereby validating expectations. On the other hand, if inflationary expectations are low, the nominal interest rate will be correspondingly low, making a strategy of low inflation feasible, validating expectations in this case as well.
It is important to note that the possibility of an economy’s not having a nominal anchor suggests that real interest rates may vary substantially over very short periods during stabilization programs and that these fluctuations may be triggered largely by changes in public expectations about inflation. Therefore, in designing adjustment programs for a country with a large nominal public debt, it appears that special attention should be devoted to the effects of alternative real interest rates on the fiscal position. In particular, an adjustment program should be designed so that the effects of possible large fluctuations in real interest rates can be accommodated within the program, thereby increasing public confidence that the policy will not be reversed.
b. Implications for Interest Rate Policy and Sterilization
The interaction between non-indexed debt and inflationary expectations has important implications for interest rate policy in the context of stabilization. This is especially so when a country faces insolvency or near-insolvency and has a large stock of nominal debt. In this context, a policy of keeping real interest rates high to reduce inflation, while possibly effective in the short run, may, in the end, turn out to cause a stronger inflationary outburst when the domestic debt burden becomes intolerable because of the high real interest rates. Depending on the starting inflation rate, the initial increase in interest rates may generate a substantial increase in the stock of nominal public debt as the higher interest is capitalized. For reasons discussed earlier, the growth of debt may induce the public to expect higher inflation as the deteriorating fiscal situation begins to undermine the credibility of the stabilization effort. As inflationary expectations worsen, nominal interest rates rise pari passu, and, if inflation remains unchanged, real interest rates increase, weakening the government’s credibility even more. Thus, a policy of high real interest rates, which was initially implemented with the objective of reducing inflation, may end up forcing abandonment of the stabilization program.35
Similarly, the presence of a large stock of nominal debt may also have important implications for the role of sterilization in the context of a price-stabilization program under a fixed exchange rate regime. Policymakers must react to the capital inflow that typically occurs during the first stages of the program. On the one hand, if the new funds are not sterilized, the domestic money supply will take a sizable upward jump, which may be interpreted as a signal that the government’s anti-inflationary stance is weakening. On the other hand, if the capital inflow is sterilized by issuing nominal bonds, then, for reasons discussed above, the increase in domestic (non-indexed) debt may weaken the credibility of the government’s price-stabilization program and result in unduly high interest rates. This, in turn, may end up forcing the government to abandon the stabilization effort and to devalue the country’s currency as the capitalization of interest rapidly increases the debt burden.36
c. Debt Indexation and Credibility
Debt indexation occurs when the domestic-currency value of the principal in a debt contract is indexed so as to maintain its purchasing power. The choice of the appropriate index, in itself, poses serious technical problems. The two most prevalent methods of indexation are (1) linking it to the consumer price index, or the exchange rate, or a combination of the two; and (2) linking the interest rate on long-term bonds to some short-term interest rate. The latter method is conceptually different from the type of price indexation discussed in this section. Interest rate indexation will be discussed later on in the context of debt maturity.37
It has been suggested that a possible solution to the problem of losing the nominal anchor is to index the public debt. (See Calvo (1988).) The reason why debt indexation may be thought of as a useful tool for controlling inflation is precisely because it breaks the link, noted earlier, between inflation and government debt.38 This occurs because the principal is indexed to maintain its real value, and, consequently, changes in the price level have no effect on the debt burden. This implies that for indexed debt, the above-mentioned reasons for tolerating inflation have no validity. Thus, by eliminating the potential link between inflation and debt, indexation can contribute to lowering the public’s expectations of inflation. Furthermore, indexed debt could be attractive because it can be thought of as introducing an element of policy pre-commitment (namely, the indexation formula itself). Hence, debt indexation can lead to a reduction in nominal interest rates and thereby help the economy reach a better (low-inflation) equilibrium.
The policy of debt indexation should, however, be used with considerable caution owing to factors relating both to the efficacy of debt indexation and to the impact that such indexation may have on other variables in the economy. It should be emphasized that debt indexation alone is by no means enough to prevent inflationary outbursts and that the effects of debt indexation will depend crucially on the soundness of underlying macroeconomic policies. Here, one can distinguish two outcomes. On the one hand, there may be instances in which a stabilization program is, in principle, sustainable, assuming there is adequate fiscal adjustment, if the private sector believes in it and in the ensuing reduction of inflation. But the program may become unsustainable if the private sector does not believe in it—that is, if there is a credibility gap which results in very high ex ante real interest rates leading to the loss of nominal anchor discussed earlier. It is precisely in such instances that a device—such as debt indexation—which enhances policy credibility may provide a bridge to sustainability by reducing interest rates.
On the other hand, debt indexation may be ineffective, and possibly even counterproductive, if fiscal adjustment is inadequate. Inflation clearly may accelerate, notwithstanding debt indexation, because of high fiscal deficits and the associated monetary policy stance. An illustration of a situation in which the presence of debt indexation is actually counterproductive is provided by the case in which there is an increase in the public sector’s deficit—for instance, because of an unanticipated increase in expenditure or a terms of trade shock—and the government is temporarily unable to raise the required revenue through conventional means. In such a case, where monetary financing is the only alternative, inflation will accelerate despite the presence of debt indexation.39 In fact, it is likely that, in this instance, debt indexation will exacerbate inflation, since it reduces the base of the inflation tax.40
Furthermore, debt indexation is effective only as long as open default is not a viable option. If default were a possibility, indexation would be ineffective in reducing ex ante real interest rates, because the inflation premium would be replaced by a default premium. If the probability of default were to be sufficiently high in the eyes of the public, then the default premium would be correspondingly high. In such a situation, the government could face a serious shortage of funds or, possibly, a run (that is, a stampede to cash in government securities).41
Two additional concerns are often raised in connection with debt indexation. First, there is a common concern that any device—such as indexation—which promises to moderate the adverse consequences of inflation is likely to weaken the resolve of governments to address the problem of inflation.42 While this argument may have some validity, especially for other forms of indexation, it may be less applicable to debt indexation. The reason for this lies in the incentive aspects discussed in this section, which suggest that it is precisely the lack of debt indexation which may weaken the resolve of policymakers to address the problem of inflation.43
Second, it is sometimes argued that debt indexation may end up leading to other forms of indexation, such as that of wages. This could occur because debt indexation imposes pressure on policymakers to protect both wage earners and asset holders from the detrimental effects of inflation. While there may be some substance to this argument, it is worth noting that, in certain circumstances, debt indexation may actually lead to less wage indexation. This would occur, for example, if, for reasons noted earlier, debt indexation enhanced the stability of monetary policy and, hence, reduced the uncertainty concerning shocks having a monetary origin. If monetary shocks were to become less prevalent, wage indexation might be reduced, since the less is the uncertainty regarding the future course of monetary or real variables, the less will be the perceived need for wage indexation.44
Given the diversity of factors noted above, it is not surprising that the empirical evidence on the relationship between debt indexation and inflation is not clear cut. In a detailed analysis focusing on the inflationary consequences of the 1973–74 oil shock, Fischer (1983 a) concluded that there was no systematic empirical relationship between inflation and various forms of indexation—in particular, debt indexation. There are no systematic data on debt indexation for the G-15 countries for the review period, but it appears that in three of the countries (Brazil, Chile, and Uruguay) most of the domestic public debt was indexed—that is, on average, more than 75 percent of the domestic public debt was indexed to the exchange rate or to the consumer price index during 1982–88. The group of G-15 countries with low indexation (with less than 25 percent of the domestic public debt indexed) includes four countries: Argentina, Colombia, Mexico, and Morocco.
As far as inflation is concerned (both in terms of its average rate and its variability), the experience across these countries has been mixed, reflecting differing fiscal stances during the review period. As noted, given the large number of factors determining inflation, it is not possible to ascertain unequivocally the extent to which debt indexation may have contributed to dampening inflationary expectations. Among the countries with high debt indexation, on the one hand, Brazil experienced extremely high rates of inflation during the review period, owing largely to inadequate fiscal adjustment. On the other hand, Uruguay and, especially, Chile experienced substantially lower average rates of inflation, with no hyperinflation experienced during 1982–88. Also among the countries with low debt indexation, the experience with inflation was not uniform: Mexico and Argentina experienced episodes of high inflation, while a number of other countries showed substantially lower average inflation rates.
To sum up, two main issues have been highlighted in this section: (1) there is a potentially destabilizing interrelationship between nominal public debt and inflation; and (2) in certain circumstances, debt indexation may complement fiscal adjustment in stabilizing inflation.
2. Debt Maturity
Maturity structure refers to the amortization schedule associated with an outstanding stock of debt. This structure is determined by the average maturities of the different types of paper issued by the government. Maturities can vary considerably, from those of overnight paper to those of longterm bonds. Among longer-term bonds, the maturities on individual instruments may vary widely. For example, a ten-year bond with a three-year grace period, and amortization in equal installments from then on, has quite a different maturity structure than a ten-year bond with no grace period, or a bond that is fully amortized at the end of the tenth year. For the purposes of the present discussion, a bond with an n-year maturity is defined as a promise to amortize the bond entirely at the end of the nth year. This definition assumes that the interest payment on a bond with an n-year maturity is determined when the bond is issued.45
The discussion below shows that debt maturity is a key characteristic in the management of public debt for two reasons.46 First, because debt maturity can play a crucial role in preventing “confidence crises” and, second, because it has important implications for the relationship between nominal debt and inflation. Before we address these issues, it is worth reviewing the recent evolution of debt maturities in highly indebted countries.
a. Evidence on Maturity Structure
Table 7 presents information on the debt maturity structure for countries in the sample for which data were available. It reports the proportion of short-term debt (maturing in less than one year) to total debt for 1983–88 and for 1987–88. As can be observed, countries with high external debt, with the exceptions of Chile and Uruguay, tend to have domestic public debt with very short maturities. Moreover, in a number of countries, notably Nigeria and the Philippines, debt maturity has shortened even more in recent years.
Central bank debt at floating rates is included as short-term debt.
Central bank debt at floating rates is included as short-term debt.
There are two possible reasons for the observed shortening of maturities. First, the fiscal situation in these countries may have raised fears of an open default. In such circumstances, the public may have preferred debt instruments that allowed them to be repaid first. However, in the event of a “run” or “confidence crisis,” when everybody wants to be repaid, none can be sure of obtaining full repayment. The possibility of a run may prevent the government from borrowing at all. Even if the government can undertake some borrowing, few people will find it attractive to acquire long-term government bonds, since they are less senior, in an intertemporal sense, than short-term debt. But short-term maturities are risky for the government, because a “confidence crisis” could trigger a run on government assets and the latter could, in turn, force the government to declare a debt moratorium even when sound economic policies were being followed.
A second possible reason for the shortening of maturities is the increase in inflation and inflationary expectations. The relationship between nominal debt and inflation was discussed earlier. How does the maturity structure of nominal debt affect this relationship? If non-indexed debt has a long maturity, then a given reduction in the real debt burden will require relatively small increases in inflation, albeit over a longer period. Conversely, if the debt has a very short maturity—say, only one month—the same real debt reduction would have to be achieved in the course of a month, requiring a sudden burst of inflation.
b. Maturity Structure and Inflation
The effect of different maturities on the potential relationship between nominal debt and inflation can be illustrated by means of a simple numerical exercise. This is done in Table 8, which shows the impact, for different maturities, of unanticipated inflation (lasting until the bond’s maturity) that is equivalent to a given amount of fiscal adjustment.47 Recalling the earlier definition of maturity, the interest rate on a bond which matures in n years is determined at the time of issue for the length of the contract. This implies that on a ten-year bond, unanticipated inflation acts for ten years. Thus, each of the numbers in the table represents the annual inflation rate which has the same effect on reducing the debt as an annual improvement equivalent to 1 percent of GDP in the primary fiscal balance for each of five years and ten years, respectively, for debt having different maturities.48 The exercise is undertaken for debt-to-GDP ratios of 25 percent and 100 percent. For example, if nominal debt is equivalent to 25 percent of GDP and matures in one year, then the unanticipated inflation rate that would have the same effect in one year as a fiscal improvement of 1 percent of GDP for each of the next five years is 20.8 percent. If that same debt matured in three years, then the inflation rate producing an effect equivalent to that of the specified fiscal improvement would be 7.9 percent per year for each of the next three years. When nominal debt is equivalent to 100 percent of GDP, the same example yields unanticipated annual inflation rates of 4.5 percent and 1.7 percent, respectively.49 Thus, a lengthening of debt’s maturity reduces dramatically the yearly rate of inflation associated with a given amount of seigniorage.
|Unanticipated Inflation Rate Corresponding to Maturity (in Years) of:|
|A.||25 percent debt-to-GDP ratio|
|Fiscal improvement equivalent to|
|1 percent of GDP for each of:|
|B.||100 percent debt-to-GDP ratio|
|Fiscal improvement equivalent to|
|1 percent of GDP for each of:|
The above calculations illustrate the basic relationship that exists between debt’s maturity structure and inflation. It is worth noting that there is a close similarity between indexation and debt’s maturity.50 This similarity arises because changes in either of these change the nominal base on which inflation acts. Indexation affects the nominal base at a given point in time by simply changing the amount of liabilities which are nominal. Thus, reducing the extent of indexation increases the nominal base. Changing debt’s maturity affects the nominal base in a more subtle way. Making a debt issue’s maturity longer enlarges the nominal base on an intertemporal dimension, in that inflation can act on it for a longer period.
Because of this similarity, the policy regarding the management of the maturity structure of debt is not independent of the policy adopted regarding debt indexation. On the one hand, as was noted earlier, indexation, in some circumstances, could play a role in controlling inflation, particularly when policy announcements lacked credibility. The advantage of having long-term maturities in such instances may, therefore, depend on the indexation policy. If debt indexation is used to reduce inflationary expectations, then a long maturity could help to control inflation by providing the government with an incentive to smooth it out over time, as is illustrated numerically in Table 8.
On the other hand, if no debt indexation is used, the existence of nominal bonds with long maturities may exacerbate expectations of high inflation. These expectations are fueled by the fact that a longer maturity structure, by allowing inflation to reduce the debt burden over a longer period, could make the inflation route less costly. Therefore, when indexation is not used, it may actually be desirable to have a shorter maturity structure. This would enhance the credibility of a government’s low-inflation policy objective by reducing the temptation for it to tolerate inflation. This temptation is reduced because, as was indicated above, in order for a country to achieve a given debt reduction, inflation needs to be higher and concentrated over a shorter period, thus increasing the cost associated with that debt reduction.
In sum, long-maturity bonds are desirable for two main reasons: first, they help prevent “runs” or “credibility crises,” and, second, they help stabilize inflation if used in conjunction with debt indexation.51 However, when policy lacks credibility and the government is unable to use debt indexation effectively to dampen inflationary expectations, shorter maturities may be called for.
The evidence on interest rates presented in Section II (Table 3) suggests that during 1982–88, the governments of several of the G-15 countries may have been able, on average, to reduce the real value of nominal public debt through negative ex post real interest rates. However, since inflationary expectations increased dramatically and, hence, nominal interest rates increased to compensate bondholders for the expected capital loss, when the reduction was economically significant, the cost has been heavy in terms of higher inflation and, in some instances, hyperinflation.
The primary fiscal improvement of 0.6 percent of GDP per year over a period of five years is obtained as follows. The once-for-all increase in the price level induces a reduction in the real value of the stock debt of 2.3 percent of GDP. This reduction equals the present discounted value of the primary surplus of 0.6 percent of GDP per annum—calculated over a five-year period and using a discount rate of 5 percent per year.
It is important to stress that for purposes of the above discussion, a country with a large stock of nominal public debt need not have a large net stock of public debt. For example, the stock of nominal liabilities might be offset by real or indexed assets, in which case the same issues would arise for a country with a small net total public debt.
In connection with price indexation, it has to be noted that foreign debt is, from a formal point of view, similar to domestic debt indexed to the market exchange rate. Thus, most of the observations which are valid for indexed debt are basically applicable to international debt. Later on in this paper, when the possibility of open default is discussed, reasons why external and domestic indexed debt may be different are mentioned.
Calvo and Guidotti (1989 and 1990 a) have also suggested that because the presence of some amount of nominal debt can provide a degree of flexibility to the policymaker, the public debt should not be fully indexed. The optimal degree of debt indexation derives from the trade-off implied by the fact that while nominal debt provides flexibility, it also increases the risk of higher inflation. The nature of this trade-off depends heavily on the ability of the authorities to commit themselves to a policy of low average inflation. If policymakers can credibly maintain a low inflation rate, on average, then even small fluctuations around this average level may reduce the need to adjust other taxes to finance expenditure shocks. In fact, this argument suggests that the more policymakers are able to credibly precommit policy, the smaller will be the proportion of debt that should be indexed.
The size of the default premium depends importantly on the perceived costs associated with open default. These costs may vary across various types of debt instruments, depending on who is the holder.
It should be pointed out that implicit in the argument that indexation weakens the resolve of governments to fight inflation is the notion that indexation reduces the costs associated with inflation. Hence, the argument does not necessarily lead to the conclusion that higher indexation reduces social welfare. (See Fischer and Summers (1989) and De Gregorio (1991 a).)
In an early work, Bach and Musgrave (1941, p. 823) argued that “by imposing upon the government a contingent liability dependent on its failure to check price inflation, the flotation of stable purchasing power bonds may exert a wholesome pressure upon Congress to adopt aggressive anti-inflationary policies.”
It has also been suggested that real shocks—for instance, shocks to productivity—need to be accommodated by changing the real wage; thus, when shocks are mostly real, wage indexation is undesirable to the extent that, by increasing real wage rigidity, they induce excessive fluctuations in output and employment. The inverse relationship between debt indexation and wage indexation may be of particular relevance when policy lacks credibility, for it is in precisely such a situation that policymakers may find debt indexation most useful for controlling inflation and, hence, for reducing shocks having a monetary origin. These implications follow from Gray’s (1976) analysis of wage indexation and Calvo and Guidotti’s (1990 a) analysis of debt indexation.
The definition of debt maturity is more complex where the interest rate on a long-term bond is indexed to a short-term interest rate. For the purpose of analyzing the interaction between nominal bonds, inflation, and maturity structure, floatingrate bonds are considered short-term liabilities. This is so because while the real value of debt may still be reduced by an unanticipated increase in inflation, the indexation mechanism incorporates any anticipated increase in inflation into nominal interest rates, in much the same way as debt may be refinanced through the issue of short-term bonds.
Calculations in Table 8 are made based upon the assumption that interest is paid at maturity. For low interest rates, the numbers reported in the table would not change significantly if alternative assumptions were made.
It should be pointed out that the 4.5 percent inflation rate lasts for one year, while the 1.7 percent inflation rate lasts for three years.
The link between optimal indexation and the maturity structure of public debt is discussed in greater detail by Calvo and Guidotti (1990 a).
When there are expectations of high inflation, however, it may often be difficult, or at least excessively costly, to issue long-term bonds. If open default is a viable option, it may not even be possible to issue long-term bonds.