This section provides an evaluation of the effectiveness of NRFs in two ways. First, it considers the effect of NRFs on government expenditure and its relationship to resource export receipts. The empirical evidence draws on econometric estimates, using data for a number of countries with NRFs.18 Second, it provides a review of selected country experience with NRFs to assess whether, and in what way, the rules of the NRF may have constrained government behavior.
Empirical Results
The experiences of selected nonrenewable resource-producing countries with and without NRFs are examined below; for a given country with an NRF, pre- and postfund performance results are compared.
A sample of 12 countries is examined, including 5 countries that have had NRFs for a significant period and 7 others that are used as a comparator group (Table 5.1). The statistical analysis focuses on central government expenditure and nonrenewable resource export earnings, both in real per capita terms. The use of export earnings may be preferable to government resource revenue because the latter may accrue with a lag and could even be endogenous.19
Selected Countries: Relationship Between Nonrenewable Resource Export Earnings and Central Government Expenditure
See Annex II f or more details.
Based on preferred equations, as described in Annex II. “ns” indicates chat the variable is not statistically significant at the 10 percent level. A “+” Or “-” indicates the sign on the estimated coefficient, if statistically significant.
Venezuela introduced an NRF in late 1998.
Selected Countries: Relationship Between Nonrenewable Resource Export Earnings and Central Government Expenditure
| Country | Sample Period |
Fund Start |
Ratio of Average Nonrenewable Resource Export Earnings to Government Expenditure (percent) |
Percent Change in Government Expenditure Associated with a 50 Percent Nonrenewable Resource Price Increase1 |
Impact of Nonrenewable Resource2 |
|
|---|---|---|---|---|---|---|
| Lagged | Contemporaneous | |||||
| Countries with nonrenewable resource funds | ||||||
| Chile | 1969–99 | 1936 | 31.4 | −6.2 | − | ns |
| Kuwait | 1965–90 | 1976 | 141.0 | −0.2 | + | − |
| Norway | 1976–99 | 1991 | 30.7 | −1.7 | ns | ns |
| Oman | 1965–99 | 1930 | 73.2 | 24.2 | + | + |
| Papua New Guinea | 1974–99 | 1974 | 71.4 | 3.0 | + | ns |
| Countries without nonrenewable resource funds | ||||||
| Algeria | 1979–99 | — | 68.2 | 3.7 | + | ns |
| Bahrain | 1965–99 | — | 79.3 | 9.9 | + | ns |
| Mexico | 1980–99 | — | 16.4 | 10.2 | ns | + |
| Saudi Arabia | 1965–99 | — | 83.0 | 11.5 | ns | + |
| United Arab Emirates | 1980–99 | — | 101.0 | 9.3 | + | + |
| United Kingdom | 1983–99 | — | 5.0 | 2.8 | ns | ns |
| Venezuela3 | 1965–98 | — | 86.2 | 10.3 | + | ns |
See Annex II f or more details.
Based on preferred equations, as described in Annex II. “ns” indicates chat the variable is not statistically significant at the 10 percent level. A “+” Or “-” indicates the sign on the estimated coefficient, if statistically significant.
Venezuela introduced an NRF in late 1998.
Selected Countries: Relationship Between Nonrenewable Resource Export Earnings and Central Government Expenditure
| Country | Sample Period |
Fund Start |
Ratio of Average Nonrenewable Resource Export Earnings to Government Expenditure (percent) |
Percent Change in Government Expenditure Associated with a 50 Percent Nonrenewable Resource Price Increase1 |
Impact of Nonrenewable Resource2 |
|
|---|---|---|---|---|---|---|
| Lagged | Contemporaneous | |||||
| Countries with nonrenewable resource funds | ||||||
| Chile | 1969–99 | 1936 | 31.4 | −6.2 | − | ns |
| Kuwait | 1965–90 | 1976 | 141.0 | −0.2 | + | − |
| Norway | 1976–99 | 1991 | 30.7 | −1.7 | ns | ns |
| Oman | 1965–99 | 1930 | 73.2 | 24.2 | + | + |
| Papua New Guinea | 1974–99 | 1974 | 71.4 | 3.0 | + | ns |
| Countries without nonrenewable resource funds | ||||||
| Algeria | 1979–99 | — | 68.2 | 3.7 | + | ns |
| Bahrain | 1965–99 | — | 79.3 | 9.9 | + | ns |
| Mexico | 1980–99 | — | 16.4 | 10.2 | ns | + |
| Saudi Arabia | 1965–99 | — | 83.0 | 11.5 | ns | + |
| United Arab Emirates | 1980–99 | — | 101.0 | 9.3 | + | + |
| United Kingdom | 1983–99 | — | 5.0 | 2.8 | ns | ns |
| Venezuela3 | 1965–98 | — | 86.2 | 10.3 | + | ns |
See Annex II f or more details.
Based on preferred equations, as described in Annex II. “ns” indicates chat the variable is not statistically significant at the 10 percent level. A “+” Or “-” indicates the sign on the estimated coefficient, if statistically significant.
Venezuela introduced an NRF in late 1998.
In some countries with NRFs, expenditure has tended to be less correlated with changes in the price of the resource. Three of the five countries with an NRF surveyed (Chile, Kuwait, and Norway) share some distinctive characteristics. Along with the United Kingdom, these are the only countries in which nonrenewable resource export earnings do not bear a statistically significant positive relationship to expenditure (see Table 5.1). In Chile the relationship is actually negative, which could be due to the pro-cyclicality of copper export earnings (Spilimbergo, 1999) combined with a countercyclical fiscal policy. Similar results, although not statistically significant, are found for Norway.
The case of Oman is more difficult to assess. Over the longer run, the relationship between oil export earnings and expenditure appears similar to that in Bahrain and Saudi Arabia, two comparable countries without an NRF (Figure 5.1). Moreover, among the countries in the sample, in Oman nonrenewable resource export earnings have the strongest influence on spending. Inspection of the data, however, suggests that expenditure may have been driven less by revenue availability in recent years, although formal tests do not find evidence of a change in the relationship. Finally, Papua New Guinea exhibited a small but positive response of spending to resource export earnings.
Nonrenewable Resource Export Earnings and Central Government Spending
(In local currency, real per capita) 1
Sources: IMF, World Economic Outlook database (Washington, various years); and IMF staff estimates.1 In thousands of local currency per capita; deflated by the CPI, 1995 = 100.2 For Chile, copper; for Papua New Guinea, copper, gold, and oil; for other countries, oil.3 Excludes information for 1991–93 owing to regional conflict and subsequent reconstruction period.The empirical evidence suggests that the establishment of the NRF did not have an impact on government spending. This result is quite robust and holds for the four countries in which tests could be performed.20 Tests were performed to assess whether (1) there was a significant change in the process driving central government expenditure after the establishment of the NRF; (2) the specific relationship between central government expenditure and nonrenewable resource export earnings changed after the establishment of the fund—for example, a stabilization fund might be expected to reduce the sensitivity of expenditure to resource export earnings; and (3) the establishment of the NRF coincided with a downward shift in the path of expenditure, as might be expected from a savings fund. For all four countries, none of the three tests suggested that the creation of the fund had a statistically significant impact on expenditure.
While the empirical results are quite strong, their interpretation should be qualified. The results are constrained by limitations on the availability and quality of data and the small sample. A particular concern is that government spending, including possibly direct spending by the NRF itself, may not be fully captured in the expenditure data—this would tend to bias the results in favor of finding a positive effect of NRFs on fiscal policy. Moreover, it is difficult to determine the optimal relationship between nonrenewable resource export earnings and expenditure, and, in comparing the results across countries, the size of nonrenewable resource export earnings relative to the economy (or expenditure) should not be ignored. For example, developments in nonrenewable resource exports are more important to Venezuela than to Chile.
The empirical evidence points toward two general results. First, it suggests that some of the countries with an NRF had a more limited expenditure reaction to changes in resource revenues than those without such funds—expenditure is found not to have been positively related to resource export earnings in the former. In Oman, however, prior to the 1990s, government expenditure was sensitive to oil export earnings despite the existence of a fund, and this has also been the case, to a lesser extent, in Papua New Guinea. Second, the evidence suggests that the relationship between government spending and resource export earnings was not affected by the establishment of NRFs in the four cases considered. This may suggest that countries with more prudent expenditure policies tended to establish an NRF, rather than that the NRF itself led to more prudent expenditure policy, though in some cases the fund may have helped maintain cautious policies.
Selected Country Experience
The rest of this section examines the specific historical experience of selected countries with NRFs. Particular attention is given to how the design and implementation of funds has addressed the uncertainties underlying nonrenewable resource prices and the extent to which funds may have constrained overall fiscal policy, 21
Norway’s State Petroleum Fund (SPF)
The Norwegian government established the SPF in 1990. However, the fund was activated only in 1995, following the achievement of overall budget surpluses. The fund is designed to manage accumulated budgetary surpluses and does not have specific rules for the accumulation or withdrawal of resources, making its operation flexible. Specifically, the budget transfers to the SPF net oil revenues. In turn, the SPF finances the budget’s non-oil deficit through a reverse transfer (provided there are enough resources in the SPF).22 Thus, the SPF effectively finances the overall budget balance. An overall budget surplus would be transferred to the fund; a budget deficit would be financed by the fund. The accumulation of assets in the SPF, which include the returns on the fund’s capital, thus represents government net financial saving. The amount actually saved depends on oil prices and the fiscal stance as embodied in the non-oil fiscal deficit. SPF assets are under the control of the ministry of finance and are managed by the central bank, with a high level of transparency and accountability (see Box 4.2). The size of accumulated funds is increasing rapidly, and reached close to 20 percent of GDP at end-1999.
A key reason for establishing the SPF was the desire to make more transparent the intertemporal policy choices available to the country, related to the expected secular decline in oil and gas output and increase in pension outlays. In this context, the SPF has helped to provide a long-term framework for the annual process of setting the non-oil budget deficit. There is currently a wide-ranging debate in Norway on the proper use of the rapidly growing assets in the SPF, which focuses on both long-term and cyclical considerations.
The SPF is effectively a government account rather than a fund. Its features ensure integration into a unitary fiscal system and address fungibility issues. The fund does not attempt to deal directly with the problems posed to the budget by the volatility of oil prices—the latter are addressed in the context of the standard budgetary process. The lack of restrictions posed by the SPF has worked well, since Norway has typically followed sound fiscal and macro-economic policies. Moreover, for a country like Norway, with a highly diversified production structure and a broad fiscal revenue base—oil revenues have typically accounted for less than 15 percent of government revenue—the challenge posed by oil price volatility to fiscal management is significantly less than for other oil producers.
Chile’s Copper Stabilization Fund (CSF)
The CSF was established in 1985 following a sustained increase in the international copper price. The CSF’s accumulation and withdrawal rules are based on a reference copper price determined annually by the authorities. No explicit formula is used to calculate the reference price. In practice, however, the reference price followed a ten-year moving average until the mid-1990s; more recently, the reference price has been set somewhat lower than the moving average. When the price of copper exceeds the reference price by between $0.04 and $0.06 a pound, 50 percent of the resulting state copper company’s revenues is deposited in the CSF; above $0.06 per pound, 100 percent. The rules for withdrawals are symmetric. The resources of the CSF are understood to have grown substantially after 1987.23 In 1998-99 there were significant withdrawals, partly on account of a sharp downturn in copper prices; more recently CSF resources have been used to subsidize domestic gasoline prices through credits to the Oil Stabilization Fund.
The impact of fluctuations in copper exports on Chile’s public finances is significantly less than in the case of many major oil producers. Copper revenues in Chile have typically amounted to less than 10 percent of total government revenues. In addition, the volatility of copper prices has been lower than that of oil prices.
Fiscal and macroeconomic policy has generally been sound in Chile. The operation of the CSF may have helped the government to resist expenditure pressures during the upswings in the copper prices in the late 1980s and mid-1990s, a fact consistent with the negative correlation between copper price increases and government spending reported above.
Oman’s State General Reserve Fund (SGRF)
The SGRF was established in 1980 with the objective of saving oil revenue for future generations. In addition, the Contingency Fund was established in 1990 to smooth the budgetary oil revenue stream, but this was replaced in 1993 by the Oil Fund, which finances investment in the oil sector.
Oman’s long experience with funds illustrates the difficulties that the behavior of oil prices and integration with overall fiscal policy can pose. At inception, the SGRF received 15 percent of oil revenues. However, the rules for the operation of the SGRF have since been changed frequently. Since 1998, the SGRF has received oil revenue in excess of the reference oil price set in the annual budget. In any given year, the government may withdraw from the SGRF up to the amount of the budgeted deficit; special procedures exist, however, for additional withdrawals in excess of the budgeted deficit. Under this arrangement, the SGRF accumulates resources when oil revenues in excess of the budgeted reference oil price are larger than SGRF financing of the budget deficit.24
Because resources of the SGRF can effectively be withdrawn at the discretion of the government, the SGRF’s balance has been determined in part by overall budgetary needs. Mainly as a result of withdrawals from the SGRF to finance the budget, the fund does not seem to have been able to fulfill its objective of accumulating resources for the time (probably only 15 to 20 years away) when oil reserves are exhausted.
Kuwait’s Reserve Fund for Future Generations (RFFG)
The RFFG was established in 1976. The objective of the fund is to save in order to provide a stream of income for future generations. The accumulation rule of the fund requires the government to deposit 10 percent of total government revenue, irrespective of oil or budgetary developments. In addition, the fund accumulates the return on its assets. The fund’s capital is understood to be invested mainly in foreign assets.25 There are no precise rules governing the withdrawal of assets, although drawing on the fund requires approval by the national assembly. Use of the fund’s resources in the aftermath of the Gulf War helped to finance a large part of the reconstruction effort.
On the basis of available information, Kuwait has followed relatively prudent fiscal policies. Over the past two decades, the consolidated fiscal position has generally been in strong surplus (excluding the reconstruction period in the early 1990s), and expenditure policy has not generally been driven by revenue availability.
Papua New Guinea’s Mineral Resources Stabilization Fund (MRSF)
The MRSF was established in Papua New Guinea in 1974. with the objective of stabilizing revenue to the budget from mining and oil sources. The inflows into the MRSF’s account are the revenues from mineral resources. The MRSF may make transfers to the general treasury account, based on a combination of rules and discretion. The MRSF Board, which is made up of government officials, has to project average annual mineral resource revenue over the next eight years, with price forecasts constrained by long-run historical average prices. On this basis a transfer, considered by the board to be sustainable over the next five years, is made.
The performance of the MRSF has been mixed. First, despite its stated objectives, transfers from the MRSF have varied significantly from year to year, almost as much as the resource revenues themselves. This has been due to the instability of revenue forecasts—itself a reflection of underlying resource price volatility, but also of budgetary needs. An initial provision restricting the amount transferred to the budget in any one year to no more than 20 percent higher than the previous year was found to be unduly restrictive and was relaxed in 1986. Second, the fund does not appear to have been well integrated with overall fiscal policy or to have helped stabilize budget expenditure, which has been partly financed with debt operations outside of the MRSF.26
The MRSF is in the process of being wound down. In late 1999, the MRSF Board was directed to use the remaining balance to reduce government debt held by the central bank. During 2000, all inflows into the MRSF were to be transferred to the general budget account, and the MRSF Act is expected to be repealed in 2001.
Venezuela’s Macroeconomic Stabilization Fund (MSF)
The MSF, established in late 1998 with the objectives of insulating the budget and the economy from fluctuations in oil prices, initially had relatively rigid rules for transfers to and from the fund. Contributions to the fund were specified as the oil revenues above a reference value corresponding to the five-year moving average. Resources could only be drawn from the fund if oil revenues in a given year were below the reference values, or resources in the fund exceeded 80 percent of the five-year moving average of oil export revenues; in the latter case, excess resources were to be used to amortize government debt.
The rules of the MSF were substantially modified in May 1999. The reference values triggering accumulation or withdrawal of resources were fixed (based on an oil price of $9 per barrel). The resources to be deposited in the fund were limited to only half of every dollar over the new reference value, and discretionary withdrawals from the fund with government authorization and legislative approval were allowed. Further changes to the rules of the MSF are currently under discussion.
The integration of the fund’s operations with central government operations has likewise proven problematic. Because the central government remained in deficit in 1999 and early 2000 despite the strong recovery in oil prices, it could only make deposits into the fund with recourse to other financing. In particular, the buildup of gross assets in the fund was financed in part by domestic borrowing. Moreover, the operation of the MSF did not prevent the implementation of an expansionary expenditure policy as oil prices rose in 2000.
Summary of Country Experience
These country experiences highlight some of the practical difficulties posed by the operation of funds. NRFs have been associated with a variety of operating rules and fiscal policy experience. In some cases where the underlying fiscal policies were generally sound to begin with (such as Norway), funds have been better able to address the problems posed by the behavior of nonrenewable resource prices and fungibility issues. At the same time, however, their operation may not have provided significant restrictions to government behavior. In many other cases, however, the funds’ rules have been frequently bypassed or changed, effectively incorporating an element of discretion. In most funds, the rules allow discretionary withdrawals to be made. As a result, often they do not seem to have constrained government spending or the nonresource deficit.
NRF performance appears to be related in part to the size of the volatility faced by governments. In particular, in Chile and Norway the share of volatile resource revenue in total government revenue is significantly lower than in other countries with NRFs. In these countries, the problems that revenue volatility poses for fiscal management are less severe than for countries that have a heavier reliance on revenues from nonrenewable resources. Conversely, the experiences of Venezuela and Oman suggest that where volatile resource revenues are large, funds (and contingent funds in particular) may be difficult to operate, and their effectiveness may be limited.
