III An Analytical Framework for Fiscal Policy and Public Debt
- Manmohan Kumar, and Pablo Guidotti
- Published Date:
- March 1991
The preceding discussion suggests that developments related to external and domestic debt have to be viewed within a unified framework that integrates several different aspects of the fiscal situation. The framework should include considerations related not only to current government revenues and expenditures but also to expectations regarding their future magnitudes. This section develops such a framework in which the basic operations of the public sector are analyzed in terms of a balance sheet listing its assets and liabilities. This balance sheet is used to obtain a measure of public sector solvency and of the fiscal adjustment required to service the outstanding public debt fully. In computing this measure, the links between debt and (current and future) taxes, subsidies, and government spending are examined. The framework itself can be viewed as complementing the use of medium-term scenarios to provide insights into problems of policy sustainability and adjustment.
I. Balance Sheet of Public Sector
Government assets can be thought of as being composed of two parts: the first is the current stock of assets (the conventional definition); the second is the present value of anticipated future revenues from tax and non-tax sources. Current assets include both domestic and foreign assets.16 The present value of future revenues is the expected stream of revenues discounted back to the present using a given discount rate. Thus, revenues expected in the future can be analyzed in terms of a current asset. The notion of present value is important, because it underscores a fundamental equivalence between the stock and the flow dimensions of fiscal policy: that is, it makes it clear that revenues expected in the future are as relevant in determining a government’s ability to meet its liabilities as current revenues.
Similarly, corresponding to government assets, government liabilities can also be thought of as composed of two parts: (1) the current outstanding stock of debt and other current obligations (the conventional definition), and (2) the present value of future expenditures, including subsidies. This way of looking at government assets and liabilities is useful, because the determinants of both shortterm and long-term fiscal performance are integrated into a single forward-looking balance sheet.17
The difference between the government’s assets and liabilities is its net worth. On the one hand, if assets exceed liabilities, then net worth is positive and the government is regarded as being solvent—that is, it is able to meet both its current and future obligations. On the other hand, if net worth is negative, then the government is insolvent and, without an increase in its assets, it is not able to meet its current contractual debt obligations. With reference to sovereign governments, however, the above definition of solvency appears to be somewhat simplistic, because not all government assets can be used to service government liabilities. Therefore, in the ensuing discussion, it is assumed that the flow of government revenues from taxes and other sources constitutes the main source of funds available to service public debt.18 The issue of public sector solvency is of particular importance when one is dealing with highly indebted countries, and it will be examined at length below.
The implications of the above framework can be seen clearly by using the following identity, which schematically presents a government’s balance sheet in terms of domestic currency:
where G, S, and R denote the present values of (expected) government expenditures, subsidies, and (tax and non-tax) revenues; A* denotes the stock of foreign exchange reserves; and B and B* denote domestic and external government debt, respectively.19E denotes the exchange rate. By assumption, B is denominated in local currency, while A* and B* are denominated in foreign currency. K denotes the government’s net worth. Consequently, government assets appear on the left-hand side of its balance sheet while government liabilities appear on the right-hand side. Because the balance sheet focuses only on those assets which are considered most likely to be used to service the public debt, K provides a measure of the government’s net worth which is relevant in assessing the public sector’s ability to service its liabilities. Therefore, it is important to stress that for the purposes of the discussion below, solvency is defined only with respect to this notion of government net worth.
2. Implications of Balance-Sheet Approach
The forward-looking balance sheet provides several interesting insights. First, for a given net worth to be maintained, any increase in debt has to be matched by one or both of the following: (1) an increase in government revenues or current assets; and (2) a decrease in expenditures. The changes in revenues and/or expenditures refer to their present values and, hence, comprise not only current fiscal adjustments but also any expectations of future improvements.
Second, domestic and external debt appear to enter the statement of public sector liabilities on an equal footing—that is, domestic and external debt have equal claims on government resources. Therefore, if an external debt-servicing problem exists, then it is likely that a domestic debt-servicing problem also exists. This simple observation suggests that departures from the equivalence of claims should be based on a recognition that the characteristics of the two types of debt may be quite different and that, as a result, government policies for managing the two types of debt may also differ significantly.
From the above observation, it follows that a domestic debt problem may occur where the ratio of domestic debt to GDP is low by international standards if, at the same time, either the ratio of external debt to GDP is high or the current or expected future fiscal position is weak. As noted in Section II, both of these conditions prevail in a number of the G-15 countries.
Third, the present value of the anticipated future stream of government expenditure and subsidies is, to the extent that the stream is perceived as a “permanent” obligation, also a form of government debt. This equivalence appears to be particularly relevant for subsidies and transfers, because they may be thought of as representing promises to provide flows of payments in much the same way governments agree to make contractual interest payments.20 By the same token, the stream of future taxes can be regarded as a form of government asset. For this reason, “solving the debt problem” may involve, for example, a cut in subsidies or an increase in taxes. In this sense, a reduction in subsidies represents a reduction of government liabilities just as a reduction in debt would.
An illustration of the equivalence between debt and the discounted present value of government expenditures and subsidies is provided by remuneration paid on bank reserves—a government transfer quite often employed in developing countries. Paying interest on reserves held against bank deposits is an interesting example, because whether such payments are included in government expenditure or are considered to be interest on part of the domestic debt depends on the government’s accounting practices. For example, on the one hand, a straight transfer would take place if the central bank just paid remuneration on commercial banks’ reserves without acknowledging the latter as part of its domestic debt. On the other hand, some countries (for example, Argentina until recently) consider part of these reserves to be government obligations. In this situation, the central bank’s transfer would become an interest payment.
Finally, it follows from the forward-looking balance sheet that whether a government is solvent or not depends on the amount of its expenditures (including subsidies), its total revenues, and its debt. On the one hand, a government is solvent if its net worth is non-negative, or, to put it differently, a solvent government does not have a debt problem.21 On the other hand, if a government’s net worth is negative, it will not be able to service fully its debt obligations.
What happens when a government is insolvent or nearly insolvent? Consider first the case where total contractual debt has reached the maximum level that can be serviced and, therefore, the government’s net worth is zero. If solvency is to be maintained, there cannot be any further increase in liabilities without a corresponding increase in assets. Thus, any increase in domestic debt not matched by an equivalent increase in assets should be met either by a decrease in the contractual value of external debt (for instance, by debt relief) or by a decrease in the present discounted value of expenditures and subsidies. Otherwise, the government will become insolvent.
A government’s insolvency has two main implications. First, if the government has no other assets to draw on in order to cover a negative net worth, the market value of its contractual obligations will have to fall; the market value, then, reflects the government’s perceived debt-servicing capacity. The market value can only fall if domestic and external debt trade at less than their contractual values—that is, if they trade at a discount. In fact, those discounts will be set by the market at levels where a government’s net worth will not be negative. Recalling the forward-looking balance sheet, this implies that
where the prices of domestic debt and foreign debt, which are denoted by q and q*, respectively, are less than unity, so that the discounts are 1-q and 1-q*, respectively.
While the market value of total debt (i.e., qB+Eq*B*) has to decline when there is the perception of insolvency, the shares of the burden that fall on domestic and foreign debt can differ markedly if the characteristics of the two debts are different. For example, if domestic debt were perceived as having, in some sense, seniority over external debt, q might not fall; hence, the brunt of the adjustment would be borne by the market price of external debt, q*.
The second implication of government insolvency concerns the issuance of new debt. If an insolvent government is able to issue new domestic debt without a corresponding improvement in its debt-servicing ability (e.g., by increasing assets or by strengthening its primary fiscal stance), then any new debt issues are likely to induce capital losses on previous creditors by depressing even more the market value of outstanding debt. The question then becomes, how can this debt be issued?
One possible explanation is that the domestic debt provides liquidity services to its holders. Such would be the case, for example, if banks were allowed to hold new debt as part of their legal reserves against bank deposits.22 New debt could then be taken up even if its market value were less than its contractual value. This is similar to the case in which the banking system is forced to hold debt at interest rates lower than market rates. The issuance of debt is feasible because it is likely to result in lower rates being paid on bank deposits, which continue to be held voluntarily by the public because of their superior liquidity relative to other financial assets.
Even if new debt did not provide liquidity services to the holder, it might still be issued if its yield were high enough to compensate bondholders for the anticipated capital loss owing to government insolvency. A third possibility is that the new domestic debt might be issued if it were perceived by the public as being “senior” relative to external debt—in other words, if holders of domestic debt believed that whatever government resources were available would be used to service their debt first.23 Such a perception could reflect a belief that the costs associated with default are substantially higher when the holders of government paper are domestic residents.24 The public perception of the seniority of domestic debt compared with external debt, however, is likely to vanish rapidly if the government faces problems in servicing its domestic debt.
3. Illustrative Examples of Government Net Worth Calculation
The above framework for analyzing government assets and liabilities provides us with a convenient method for obtaining estimates of a government’s net worth, as defined earlier. These estimates can provide an indication of the ability of a government to service its debt in full. In those cases where net worth is negative, the estimates can also be used to obtain a measure of the fiscal stance required to reestablish solvency.
To illustrate the relative contribution of the individual components of the government’s balance sheet to its net worth, four hypothetical cases are presented in Table 5. It should be noted that these four cases, although they are hypothetical, reflect orders of magnitude which are not atypical of several of the G-15 countries and have been designed to show the interaction of alternative fiscal stances with different levels of public debt.
|Case||Primary Surplus or Deficit (-)||Present Discounted Value of Primary Balance (r = 0.05)2||Foreign Exchange Reserves||Total Public Debt||Net Worth|
Computations are based on equation (1).
r = discount rate.
Computations are based on equation (1).
r = discount rate.
The starting point for the calculations is an estimate of the primary fiscal balance expected to prevail in the future. For instance, it could be assumed that the primary balance in the recent past provides a reasonable indication of the balance likely to prevail in the future. In the first two cases, expected primary surpluses of 1.5 percent and 1.0 percent of GDP, respectively, are assumed. In the two remaining cases, expected primary deficits of 1.5 percent and 2.5 percent of GDP, respectively, are assumed. The present discounted value of these primary balances, computed using an illustrative discount rate of 5 percent per annum, can thereby be obtained.25 For instance, the present value of a primary surplus of 1.5 percent of GDP is equivalent to that of a current asset amounting to 30 percent of GDP, which clearly would increase the government’s net worth. Similarly, the present value of a primary deficit of 2.5 percent of GDP is equivalent to that of a debt amounting to 50 percent of GDP, which clearly would decrease the government’s net worth.
The two other variables that enter the net worth calculations are the ratios of foreign exchange reserves to GDP and of total (domestic plus external) public debt to GDP. Foreign exchange reserves are assumed to range between 5 percent and 10 percent of GDP, in line with the observation that foreign exchange reserves tend to be small relative to the other magnitudes involved in the calculation of government net worth. Total public debt levels range from 30 percent to 80 percent of GDP. Government net worth is obtained by adding together the present value of the primary balance, the foreign exchange reserves, and the total public debt. Government net worth is negative in three of the four cases, reflecting the impact of a high level of public debt and/or a large primary fiscal deficit. The values of net worth vary significantly across the different cases, with the government in Case A being solvent—a situation in which the expected primary surplus is adequate to meet outstanding government liabilities—with the governments in the remaining cases having substantial negative net worths. It is interesting to note that the government in Case D, while having the lowest ratio of public debt to GDP—similar to that of its counterpart in Case A—shows a significantly negative net worth—similar to that of its counterpart in Case C—because of its inadequate fiscal adjustment. This example illustrates the fundamental equivalence between the consequences of, on the one hand, inadequate fiscal discipline and, on the other hand, an excessive debt burden.
An important implication of the balance-sheet approach is that government net worth—by summarizing the broad fiscal situation—provides an indication of the government’s ability to service its current debt. It was also noted earlier that secondary-market discounts on external debt reflect uncertainty about a government’s ability to meet its current liabilities. Therefore, one should expect to observe a relationship between an estimate of government net worth and the secondary-market discount on the external debt. In particular, the higher government net worth is, the greater should be its perceived ability to service its external debt and, hence, the lower the secondary market discount on that debt.
It is interesting to explore the extent to which the above relationship is borne out in the G-15 countries. To this effect, the previous methodology was used to obtain a measure of the public sector’s ability to service its outstanding debt for nine of the G-15 countries.26 This measure was then correlated, across the nine countries, with the corresponding secondary-market discounts for 1987–88. The findings suggest that there is, indeed, a close statistical association between secondary-market prices and the public sector’s perceived ability to service its outstanding liabilities, with the correlation coefficient between these two variables being 0.85—statistically a highly significant result.
4. Required Fiscal Adjustment
Once a negative estimate of government net worth (as shown in three of the four cases in Table 5) is obtained, this immediately poses the question of what would be the primary fiscal adjustment required to re-establish solvency—that is, how large a permanent primary surplus would be required to enable the government to service its current public debt? The surplus must be large enough to prevent government net worth from being negative. The required primary surpluses (as a percentage of GDP), computed using discount rates of 3 percent and 5 percent per annum for the four hypothetical cases discussed above, are reported in Table 6. When a 5 percent discount rate is used, the primary surplus needed is quite sizable, especially in Cases B and C, for which it exceeds 3 percent of GDP. More importantly, with the exception of Case A, the calculations imply a substantial permanent fiscal adjustment compared with current expected performance. It is important to stress that while some of these numbers may appear to be manageable for governments, they represent a significant adjustment, given that the improvement has to be maintained into the future.
|Case||Required Fiscal Surplus (r = 0.05)1||Required Fiscal Surplus (r = 0.03)1||Expected Fiscal Balance|
r = discount rate.
r = discount rate.
These required surpluses could fall markedly if the countries concerned experienced higher longrun rates of growth. For instance, if the rate of growth were 2 percentage points higher than we assumed previously, so that the corresponding discount rate were 3 percent rather than 5 percent per annum, then the required primary surplus would fall in all four cases. A substantial permanent fiscal adjustment, compared with the government’s expected primary fiscal stance, would still be required in the last three cases, however.
The balance-sheet framework discussed above could be of considerable use in the design and evaluation of adjustment programs. This is so because the public sector’s net worth and the calculations pertaining to the required fiscal adjustment may be regarded as central to the issue of policy sustainability. In this respect, the use of the balancesheet framework could be viewed as a complement to the use of medium-term scenarios to evaluate the impact of changes in economic policy and exogenous variables on the major macroeconomic targets. Unlike medium-term scenarios, calculations of net worth and the required fiscal surplus do not require one to project the specific path of policy and macroeconomic variables. Since this framework relies only on discounted present values, it provides a measure of the sustainable fiscal stance that could be consistent with alternative assumptions about the time path of exogenous and policy variables.27
The above framework could also be used to shed light on one of the questions raised in Section II, where it was noted that the ratio of total public debt to GDP in many of the G-15 countries is similar to those in a number of industrial countries. The question was raised concerning the extent to which the fiscal stance of the two groups of countries would explain why only the developing countries have faced a debt crisis. If one considers the recent primary fiscal performance of a number of the industrial countries with the highest levels of public debt, the difference in fiscal performance does not appear to justify the presence of a debt crisis only in the case of developing countries.28
One possible explanation for the crisis occurring only in developing countries is that future expected primary surpluses are larger in industrial countries. In fact, it has been argued that industrial countries’ ability to raise additional revenues may be substantially greater than that of developing countries, because of the former’s more efficient tax collection as well as structural factors which imply differences between the tax bases in the two groups of countries.29 If this is so, it would suggest that in a developing country, structural reforms in the public sector, as well as policy measures which made the market more optimistic about the country’s future fiscal performance, could have a significant impact on the public sector’s creditworthiness, even if the effects of those reforms might be felt only after a lag.30
Examples of domestic assets are land and buildings owned by the government, and domestic government loans made to the private sector. Foreign assets include foreign exchange reserves, foreign loans made by the government, and fixed foreign assets, such as embassies.
See Buiter (1983) for a similar treatment of public sector accounts.
One can think of foreign exchange reserves as a current asset which, under certain circumstances, may be used for servicing debt. This assumption is made in the ensuing discussion.
Seigniorage owing to inflation can be thought of as a tax and is therefore included in government revenues, R. The operational surplus or deficit of public sector enterprises is included in government revenues, R. It is worth noting that the privatization of public sector enterprises has two effects on the government’s balance sheet. On the one hand, it will affect the discounted present value of revenues. On the other hand, it affects current assets. If assets are valued at market prices, these two items will tend to cancel each other out. However, the decision to privatize may represent a change in government policy—say, with respect to pricing—and could have an effect on net worth. Similarly, it may be argued that, over time, the government could levy taxes on the newly created private enterprises, and to the extent that this possibility is not reflected in the sale price, the present value of revenues could increase.
It may be argued that the government has significantly more discretion in cutting subsidies than in defaulting on debt, because subsidies are generally not contractual. However, it should be emphasized that, in practice, subsidies are often difficult to remove because of the influence of vested interests.
In this discussion, no distinction is made between ability and willingness to pay. The willingness to pay, in the context of fiscal adjustment, is clearly an important factor, which may vary across countries. Thus, even if a government is technically solvent, its public debt could still sell at a discount.
Banks would voluntarily hold government bonds in these circumstances whenever the spread between the interest rate earned on their reserves and the lending rate was lower than the equivalent rate associated with the discount on government bonds.
Two observations may provide some support for the notion that the new domestic debt issued in the G-15 countries after 1982 may have been perceived as having an implicitly senior status. First, domestic public debt, unlike external debt, continued, in general, to be serviced. At the same time, while almost no new private external loans were made to G-15 country governments, they continued to have access to fresh domestic funds. Second, domestic debt was issued at interest rates that, at least on an ex post basis, were not higher (and were often considerably lower) in real terms than the interest on external debt. The possibility that domestic debt may have been perceived as being senior suggests that the increased reliance on domestic debt financing by most of these countries could have contributed in a substantial way to the capital losses experienced by foreign creditors.
It could be argued, on a legal basis, that the reverse could actually be the case. While the government can always plead sovereign immunity in its own courts, it cannot do so in foreign courts. Furthermore, foreign creditors can penalize the debtor government in a number of ways—for example, suspending trade credits or seizing the debtor’s assets located abroad. Nonetheless, it could also be argued that it is easier to repair damage to a country’s international reputation than damage to its domestic reputation, which could have immediate political repercussions.
The discount rate equals the long-run real rate of interest minus the long-run rate of growth of GDP. The 5 percent discount rate is consistent with a real rate of interest of 8 percent and a GDP growth rate of 3 percent. Considerable empirical evidence indicates that the long-run real rate of return on capital, which may be equated with the long-run real rate of interest, is in the range of 7–10 percent per annum. (See, for example, Boskin (1978).) Thus, a 5 percent discount rate could be consistent with output growth rates ranging between 2 percent and 5 percent per annum. It is worth noting that, in principle, for the purpose of evaluating government solvency, the discount rate used to obtain net worth should be based on an interest rate which excludes a default-risk premium, since the presence of a default-risk premium already embodies an expectation of insolvency. Similarly, it should be noted that for the purpose of evaluating solvency, it is natural to assume that the public sector has access to capital markets. The observation that many G-15 countries face steep risk premiums as well as credit constraints is, itself, a reflection of their perceived insolvency, not the cause of it.
The nine countries include Argentina, Brazil, Chile, Colombia, Ecuador, Mexico, Peru, the Philippines, and Venezuela. These are the countries for which the required data were obtainable from recent unpublished Fund documents. Owing to their confidential nature, the data for individual member countries cannot be included here.
It is worth noting that, operationally, the net worth and required fiscal stance calculations require considerably fewer assumptions as well as less data—over the time horizon covered by the medium-term scenarios—because they do not focus on year-to-year projections.
If one were to perform the calculations applied in Table 5—taking the recent fiscal performance as an indication of the future primary fiscal stance—for both the G-15 countries and a number of industrial countries with high public debt levels, the public sector’s balance sheet (as described by equation (1)) would not look substantially different for countries in the two groups. For both groups of countries, one would conclude that the recent primary fiscal stance was not consistent, over the long run, with the servicing of current public debt.
See Reisen (1989 a); for an alternative view, see Tanzi (1989). In addition, it might be argued that the process of economic integration might contribute, in some cases (e.g., European countries), to a more optimistic view about future fiscal performance.
Another possible explanation could be that the GDP growth of industrial countries may be expected to be higher. However, neither a comparison of growth rates of the two sets of countries over the last two decades nor the growth projections reported in the recent World Economic Outlook: A Survey by the Staff of the International Monetary Fund (Washington: International Monetary Fund, May 1990) suggest that such a growth differential is likely to play a significant role.