III. Medium-Term Fiscal Sustainability Analysis20
70. The need for a medium-term fiscal sustainability analysis is evident from the current threat to Ghana’s economic performance posed by external shocks in 1999, which contributed to an excess in net domestic financing over projected levels and consequent increase in domestic debt. The medium-term fiscal model used to carry out this analysis is consistent with the macroeconomic framework and budget forecasting methodology underlying the ongoing ESAF-supported economic program. It is used to assess whether the fiscal position of the government remains sustainable after the external shock, and if not, whether planned adjustments in fiscal policy are able to restore domestic financing and debt to sustainable levels.
B. Stylized Model
71. Analysis of fiscal sustainability is based on projections of revenue, expenditure, and financing within a consistent macroeconomic framework. Given projections of real GDP growth, inflation, trade, and external financing, the macroeconomic model generates estimates of interest rates and exchange rates, as well as the fiscal position, including detailed projections of revenue and expenditure based on the tax structure and expenditure programs. Simulations can then be used to analyze the effects of different policies on fiscal sustainability. The following indicators of fiscal sustainability are examined: (a) the domestic revenue-to-GDP ratio; (b) the ratio of interest payments to revenue; (c) the overall deficit as a percent of GDP; (d) the domestic primary balance as a percent of GDP; (e) net domestic financing as a percent of GDP; and (f) the ratio of domestic debt to GDP. The stylized macroeconomic model underlying the fiscal sustainability analysis is comprised of the following components:
72. Ghana’s domestic revenue is composed of domestic tax and nontax revenues and trade taxes. Both domestic taxes and nontax revenues are predominantly a function of nominal GDP. Direct taxes, VAT on domestic goods, excise taxes, petroleum taxes, and nontax revenue are assumed to be proportional to GDP. VAT and tariff collections from imports are a function of imports and the exchange rate. Cocoa taxes depend on a number of factors, such as world and producer prices, which are exogenous to this stylized model, and the exchange rate. Therefore, the domestic revenue-to-GDP ratio can be expressed as:
where a represents domestic taxes and nontax revenues, E is the exchange rate, τ represents the effective tariff, m* is imports and c* represents cocoa taxes.21 In the absence of tax changes, the revenue-to-GDP ratio tends to decline over time because the cocoa tax is expected to fall as a proportion of GDP over the medium term. Although imports are expected to grow broadly in line with GDP, the effective tariff is also to expected to decline over time as a result of tariff reforms. The remaining source of revenue is external grants:
where the amount of grants, g*, is exogenous in terms foreign currency.
73. There are four types of expenditures: (a) domestic primary expenditure; (b) domestic interest payments; (c) foreign interest payments; and (d) foreign-financed capital expenditures. Primary domestic expenditure, s, is a function of nominal GDP. The main components of primary domestic expenditures do not vary proportionally with GDP over the medium term. For example, the nominal income elasticity of wage payments is less than unitary given the government’s policy of reforming the public service to make it more efficient. Domestic capital expenditure is projected to rise as a percent of GDP to compensate, but not fully, for the decline in external project financing. Recurrent spending on goods and services rises slightly in 2006–07 as a consequence of higher domestic capital expenditure.
74. Domestic and foreign interest payments are given by:
where I and I* are the domestic and external interest rates, respectively, and d and d*, the stock of domestic and foreign debt, respectively. Foreign-financed capital expenditures, f, are assumed to be equal to total disbursements of external project grants and loans.
75. The ratio of domestic interest payments to revenue is given by:
Overall and primary domestic deficits
76. The overall deficit, b, is the difference between revenue and expenditure:
77. The primary domestic deficit, z, is the difference between domestic revenue and primary domestic expenditure:
78. The domestic debt stock can be derived from the government budget constraint, as the overall budget deficit has to be financed either externally or domestically. Therefore:
Nominal GDP, exchange rate, nominal domestic interest rate
79. Nominal GDP, Y, is obtained from exogenous projections of real GDP growth and inflation.
where P is the GDP deflator.
80. The exchange rate depreciation is assumed to adjust according to purchasing power parity:
where gE is exchange depreciation, gp is domestic inflation gp* is foreign inflation.
81. The nominal domestic interest rate is given by:
where K is the real interest rate which is assumed to be related to risks associated with uncertainty about the sustainability of the fiscal position and, more generally, of macroeconomic policies. Therefore, any change in projected domestic inflation affects nominal GDP, the exchange rate, and nominal interest.
C. The Original Medium-Term Outlook
82. The government’s original program for 1999–2001 was designed to support real GDP growth of 6 percent and, through adherence to prudent financial policies, reduce inflation to no more than 5 percent over 1999–2001. The program moved the overall fiscal deficit toward balance in line with the Ghana—Vision 2020 strategy, and domestic primary surpluses to at least 3.5 percent of GDP by 2001 so as to reduce domestic debt. Net domestic financing of the budget was projected to fall to 1.3 percent of GDP in 2001 and the stock of domestic debt to begin declining relative to GDP by 2001. To achieve these objectives, expenditure growth would be restrained, although expenditures in health and education were to be protected.
83. Over the medium-term, the government plans to reform its tax system, shifting from reliance on trade taxes toward taxation of domestic consumption, thereby raising revenue more efficiently to support expenditure programs while improving equity. The tariff policy reforms are expected to start next year and continue through the medium term. Import tariff reform is guided by regional integration and the overall objective of reducing tariffs to improve the efficiency and competitiveness of Ghana’s productive activities. Specifically, the government plans to reduce the top import tariff rate from 25 percent to 20 percent by the beginning of 2000 and shift goods to lower rates over the medium term, thereby reducing the average tariff rate below 10 percent and the effective rate, τ, from 7.3 percent in 1999 to 6 percent in 2007. The government’s original plan under its medium-term cocoa strategy included the gradual reduction of the cocoa export tax to 15 percent of the world f.o.b. price by the 2004/05 crop season. This was designed to raise the incentive to produce cocoa and improve incomes of producers who comprise a large segment of the rural population engaged in small-scale cocoa production.
84. To compensate for the revenue losses from trade tax reform and meet spending requirements, the VAT rate was to be raised by annual increments of ½ percentage point over 2001–06, bringing the rate to 13 percent. To further improve efficiency, petroleum taxes would be increased. By 2007, tax revenue was projected to rise above 17 percent of GDP, the overall budget to be in surplus by 1 percent of GDP, net domestic financing negative, and the stock of domestic debt below 10 percent of GDP. The domestic debt stock was expected to decline by about one-half relative to GDP over the medium term, allowing real interest rates to decline and making the fiscal position significantly stronger than at the present time.
D. The Baseline Scenario
85. The delays and shortfalls in external assistance in 1999 and deterioration in the terms of trade by over 13 percent in 2000 substantially changed the medium-term outlook. Real GDP growth for 1999 is projected to be lower than originally projected, specifically 4.5 percent rather than 5.5 percent, owing to tighter financial policies and external factors. The decline in the price of cocoa is expected to have greater impact next year, with external financing gaps opening up in 2000 and 2001 that can only be partially closed with external assistance. To fully close these gaps, imports are expected to be lower than originally projected under the program, thereby slowing GDP growth. In particular, real GDP growth is projected to be 4 percent in 2000, 4.5 percent in 2001, and reach 6 percent only in 2003.
86. The baseline for examining fiscal sustainability is the medium-term projection of fiscal accounts based on the revised macroeconomic and external projections, and tax and expenditure policies in 1999, with, in particular, the VAT rate remaining at 10 percent.22 The ratio of direct taxes, indirect taxes (except VAT on imports), and nontax revenue to GDP remains constant through the medium term given unchanged domestic tax rates and a projected unitary elasticity of these revenue sources with respect to GDP. The VAT on imports and trade taxes are based on the medium-term projections of imports and cocoa exports and the effective tariff rates implied by tariff reform and the medium-term cocoa strategy.
87. Given the tax burden, expenditures will need to be curtailed over the medium term. Capital expenditures decline over the medium term as external project assistance is projected to decrease and the role of the private sector in the economy is expanded. Domestic capital expenditure is expected to increase only slightly. Transfers to household are determined by mandatory programs and are projected to increase slightly as a percentage of GDP over the medium term. Domestic interest is calculated based on domestic debt stocks and their associated interest rates. Net domestic financing is added to domestic debt stocks, thereby providing the base for carrying the interest calculation into subsequent periods (Equation 8).
88. The baseline scenario has tax revenue declining to 15.3 percent of GDP in 2000 mainly because of the sharp decline in cocoa taxes and reductions in the effective tariff, τ, owing to tariff reform (Figure 11 and Table 13). The tax ratio recovers to 15.8 percent of GDP in 2002–03 with the recovery in cocoa prices, then declines to 15.6 percent of GDP in 2006–07 as the effective tariff declines and the government’s share of cocoa export receipts stabilize at 10 percent. The ratio of domestic interest payments to revenue, equation (5), decreases from 23 percent in 2000 to 12 percent in 2007, but is above projected levels under the original program. The overall budget on a commitment basis, including grants, never achieves balance as under the original program, remaining 2 percent of GDP in deficit in 2007. Net domestic financing, although declining, remains positive throughout the medium term, reaching 0.7 percent of GDP in 2007. The domestic primary balance never reaches 3.5 percent of GDP, an important target for reducing the stock of domestic debt. In that regard, the domestic debt stock only begins to decline relative to GDP by 2004 and is over 17 percent in 2007, twice the level targeted under the original program.
Figure 11.Ghana: Benchmark and Revised Program, 2000–07
Source: Fund staff estimates.
|Stock of domestic debt||21.1||20.5||20.9||20.9||20.3||19.4||18.3||17.2|
|Overall balance (commitment basis)||−7.6||−6.6||−5.2||−4.2||−3.4||−2.7||−2.3||−1.9|
|Domestic interest to revenue ratio||23.1||20.6||18.7||16.6||14.9||13.6||12.7||12.1|
|Net domestic financing||3.3||2.3||2.7||2.2||1.6||1.2||0.9||0.7|
|Domestic primary balance||2.5||3.0||3.2||3.3||3.3||3.4||3.4||3.3|
|Stock of domestic debt||20.2||18.9||18.6||17.8||16.5||14.7||12.9||11.0|
|Overall balance (commitment basis)||−6.7||−5.7||4.3||−3.2||−2.2||−1.6||−1.0||−0.5|
|Domestic interest to revenue ratio||21.9||18.8||16.4||14.0||11.8||10.2||8.9||7.8|
|Net domestic financing||2.4||1.4||1.8||1.1||0.5||0.0||−0.3||−0.7|
|Domestic primary balance||3.4||3.7||3.8||4.0||4.0||4.0||4.0||4.0|
|Domestic debt fund|
|Stock of domestic debt||19.9||18.2||17.6||16.7||15.3||13.6||11.7||9.7|
|Overall balance (commitment basis)||−6.4||−5.4||−3.9||−3.0||−2.1||−1.4||−0.9||−0.4|
|Domestic interest to revenue ratio||21.9||18.4||15.7||13.1||11.0||9.4||8.0||6.9|
|Net domestic financing||2.0||1.1||1.4||1.0||0.4||−0.1||−0.5||−0.8|
|Domestic primary balance||3.0||3.4||3.5||4.0||4.0||4.0||4.0||4.0|
|External assistance shortfall|
|Stock of domestic debt||21.4||20.0||20.1||19.8||18.8||17.3||15.8||14.1|
|Overall balance (commitment basis)||−6.2||−5.4||−4.4||−3.3||−2.7||−2.0||−1.6||−1.2|
|Domestic interest to revenue ratio||22.2||21.3||21.1||18.8||17.1||15.0||13.4||12.1|
|Net domestic financing||3.4||2.6||3.0||2.3||1.6||1.1||0.6||0.3|
|Domestic primary balance||3.1||3.5||3.6||3.8||3.8||3.7||3.7||3.7|
|Stock of domestic debt||19.9||18.9||18.6||17.9||16.9||15.6||14.1||12.4|
|Overall balance (commitment basis)||−7.0||−6.0||−4.6||−3.5||−2.6||−1.8||−1.3||−0.8|
|Domestic interest to revenue ratio||23.9||21.3||18.8||16.5||13.9||11.9||10.2||9.0|
|Net domestic financing||2.8||1.9||2.3||1.7||1.0||0.5||0.1||−0.2|
|Domestic primary balance||3.2||3.5||3.5||3.6||3.5||3.6||3.5||3.5|
|Higher real interest rates|
|Stock of domestic debt||20.4||19.4||19.5||19.0||18.0||16.6||15.1||13.4|
|Overall balance (commitment basis)||−7.0||−6.1||−4.7||−3.6||−2.7||−2.0||−1.5||−1.1|
|Domestic interest to revenue ratio||23.1||20.7||18.6||16.4||14.4||12.9||11.6||10.6|
|Net domestic financing||2.6||1.8||2.2||1.6||1.0||0.5||0.2||−0.1|
|Domestic primary balance||3.4||3.7||3.8||4.0||4.0||4.0||4.0||4.0|
E. A Fiscal Strategy to Address the 1999 External Shock
89. The sharp decline in international cocoa prices in 1999 has accelerated the government’s cocoa reform agenda, with the producers’ share of the price rising faster than originally planned and government’s cocoa export tax revenue falling. The decline in world cocoa prices is expected to have its greatest impact on cocoa tax revenue in 2000, with projected revenue less than 1 percent of GDP. Measures are needed to address this loss and subsequent losses following from import tariff reform, so as to return to the extent possible to the original fiscal adjustment path.
90. An increase in the VAT rate would be efficient and consistent with the government’s policy of moving toward a tax system that encourages saving and investment. At present, Ghana’s VAT rate is low in comparison with neighboring countries and with other countries that have a VAT. This analysis therefore assumes an increase in the VAT rate to 12.5 percent early in 2000 as the most efficient way to address the substantial revenue loss in that year. This increase is followed by a further increment to 13 percent in 2003 to compensate for revenue losses from tariff reform in the medium term.
91. The VAT rate increases bring the budget nearly into balance over the medium term, achieving an overall deficit of only 0.5 percent of GDP in 2007 (Figure 11). The domestic primary surplus rises steadily from 3½ to 4 percent of GDP in 2003 and stays at that level afterwards. Net domestic financing becomes negative at about ½ percent of GDP in 2006–07, meaning that the budget allows for retirement of outstanding domestic debt stock by the end of the medium term. Tax revenue recovers to 16½ percent of GDP, below the level reached over the medium term under the original program but enough to bring levels of domestic debt stock and interest substantially closer to original targets. Domestic interest payments as a share of revenue drop below 10 percent in 2006 and below 8 percent in 2007. The stock of domestic debt declines to 11 percent of GDP in 2007, about 10 percentage points below its peak in 2000.
F. Domestic Debt Fund
92. Ghana’s domestic interest burden follows mainly from its high domestic interest rates. Ghana’s domestic debt stock as a share of GDP is not out of line with other African countries, but interest payments on this debt are much larger because Ghana’s interest rate is much higher (Box 2).23 In October 1998, the government carried out a National Debt Workshop with the assistance of Debt Relief International. Among its recommendations, the workshop recommended setting up a domestic debt fund with donor contributions to mitigate the burden of domestic debt service.
Box 2.Public Domestic Debt: Selected African Countries, 1998
|Domestic interest||Domestic debt stock||Average|
|(In % of|
|(In % of|
|(In % of|
|(In % of|
|(In % of|
93. The domestic debt fund examined in this analysis comprised an additional contribution of US$50 million annually from donors collectively during 2000–2002. This amount of support could be expected to have a significant impact on domestic debt and poverty reduction, without straining the limits for liquidity creation and absorptive capacity of the economy. It is assumed that 50 percent of the debt service savings are used for additional priority expenditures. In this light, priority spending was increased by half the amount of the fund, that is US$25 million annually over 2000–2002.
94. The saving in domestic interest achieved with the donor-financed domestic debt fund lowers the overall deficit and net domestic financing in all years of the medium term (Figure 12). The stock of domestic debt drops below 10 percent of GDP in 2007, an objective of the original program. Additional priority spending equal to half of the savings in domestic interest reduces by half the improvement in the domestic primary surplus achieved under the program over the baseline in 2000–2002.
Figure 12.Ghana: Domestic Debt Fund, 2000–07
Source: Fund staff estimates.
G. Shortfall in External Assistance
95. The fiscal strategy for addressing the effects of the terms of trade shock in 2000 relies on both domestic tax measures and additional external assistance. An examination of the impact of a shortfall in external assistance on medium-term fiscal performance usefully illustrates the importance of this assistance. This simulation was carried out by assuming shortfalls (against the baseline) of US$200 million each in 2000 and 2001, US$150 million each in 2002 and 2003, US$100 million each in 2004 and 2005, and US$50 million each in 2006 and 2007. These shortfalls were generated by equiproportional reductions of both program and project grants and loans in each year. Because the balance of payments is left unchanged, exchange rates are expected to adjust so that imports decline in line with the shortfall in external assistance. Shortfalls in program assistance (g* in equation (8)), however, create a gap in budget financing, which was assumed to be closed by additional domestic financing. Because a shortfall in project assistance reduces spending by the same amount, it has no effect on domestic financing.24
96. The shortfall in external assistance causes fiscal performance to deteriorate, substantially nullifying the fiscal measures examined above for addressing the effects of the external shock in 1999 (Figure 13). The only exception is the overall budget balance which follows the path under the revised program, owing to the reduction in the deficit from lower foreign-financed capital expenditure. Notwithstanding this exception, net domestic financing is higher than under the revised program because of the need to make up for the shortfall in external program assistance and the loss in import tariff revenue caused by lower imports.25 This additional domestic financing raises the stock of domestic debt above levels under the revised program. The ratio of domestic interest to revenue exceeds levels under revised program and the baseline scenario for most of the medium term because of both higher interest payments and lower tariff revenue. Lower tariff revenue causes the domestic primary surplus to fall short of the levels achieved under the revised program.
Figure 13.Ghana: External Assistance Shortfall, 2000–07
Source: Fund staff estimates.
H. Resilience of Program to Shocks
97. Given the impact of external developments in 1999 on fiscal performance, the medium-term fiscal sustainability analysis was used to examine the resilience of the revised program to shocks—specifically, higher inflation and interest rates.
98. Assuming an inflation path that is 4 percentage points higher than the baseline in 2000–02, and thereafter converges gradually to the baseline path, adversely affects medium-term fiscal sustainability (Figure 14). The ratio of tax revenue to GDP declines from program levels because trade taxes are not linked to GDP and therefore decline as a proportion of GDP owing to inflation.26 Because expenditures decline less as a percentage of GDP with higher inflation, the domestic primary surplus is lower than under the program, remaining fairly level at 3½ percent of GDP from 2001, about ½ percentage point below the revised program. Correspondingly, the overall deficit and net domestic financing are higher, although they are below 1 percent of GDP and negative, respectively, by 2007. The ratio of domestic interest to revenue rises dramatically, exceeding baseline levels through 2002, as increases in domestic interest owing to higher nominal interest rates exceed inflationary increases in revenue. The ratio of the stock of domestic debt to GDP is lower in 2000 than under the program because higher inflation affects GDP immediately. However, higher domestic financing quickly takes over, resulting in a larger stock of domestic debt by about 1½ percentage points of GDP in 2007.
Figure 14.Ghana: Higher Inflation, 2000–07
Source: Fund staff estimates.
99. Real interest rates that are 2 percentage points above those in the program roughly halve the improvement in key fiscal variables achieved under the revised program over the baseline scenario—for example, the overall budget deficit rises from 4.3 percent of GDP in 2002 under the revised program to 4.7 percent of GDP, but remains less than 5.2 percent in the baseline scenario (Figure 15). Similarly in 2002, net domestic financing increases from 1.8 percent of GDP under the revised program to 2.2 percent, remaining below 2.7 percent in the baseline scenario. The stock of domestic debt declines to 13½ percent of GDP in 2007, between 11 percent under the revised program and 17 percent in the baseline scenario. Higher interest rates virtually eliminate any improvement in the domestic interest-to-revenue ratio through 2003 and leave the ratio above 10 percent in 2007. As expected, tax revenue and the domestic primary surplus are unaffected by higher real interest rates.
Figure 15.Ghana: Higher Real Interest Rates, 2000–07
Source: Fund staff estimates.
I. Concluding Observations
100. The medium-term fiscal sustainability analysis provides a useful quantification of the impact of the shocks experienced in 1999 on fiscal performance through 2007. It showed that, without compensating measures, domestic interest and debt absorbed substantial revenue and financial capital, crowding out public expenditures and private investment over the medium-term. It also permitted precise calculation of the magnitude of measures needed to bring fiscal performance back on track for achieving the objectives of the government’s economic program—for example, the VAT rate increases needed to compensate for the sharp revenue decline in 2000 and subsequent revenue losses from tariff reform during the medium term.
Prepared by Anthony Pellechio.
All lower case variables are expressed as a share of GDP. The asterisk represents values expressed in foreign currency and assumed exogenous. Imports in the baseline scenario are a function of GDP, but not in subsequent simulations.
This differs from the original program, which had VAT rate increases of ½ percent over the medium term in order to compensate to reductions in import duty collections owing to tariff reform.
The high real interest rates in Ghana are a result of past expansionary fiscal policies, which placed a heavy burden on monetary policy. Real interest rates are expected to decline gradually as debt levels decline and confidence increases.
Foreign-financed capital expenditure, denoted f in the model, does not appear in equation (8).
To the extent lower imports, which were expected to be financed by foreign financing, are exempt from tariffs, the fall in tariff revenue will be less than assumed in the simulation.
Higher inflation causes the exchange rate to depreciate according to equation (10). A more depreciated exchange rate raises the values in domestic currency of import duties and cocoa export taxes, but not enough to outweigh the effect on their ratio to GDP of higher nominal GDP owing to higher inflation.