Russian Federation: Selected Issues
Author: International Monetary Fund
Publication Date:
06 Dec 2006
eISBN:
9781451975307
Language:
English
Keywords:
ISCR; CR; FSU country; CIS country; oil revenue; oil price; consumption rule; exchange rate; oil GDP; high-price scenario; oil price assumption; gas revenue
Download PDF (2.2 MB)

Russia’s large oil and gas reserves play a key role in its economic development. As with many other large oil exporters, Russia’s energy wealth is also posing numerous challenges to macroeconomic management. Although fiscal policy has saved a large part of the oil windfall in the Oil Stabilization Fund (OSF), this has not been guided by a consistent long-term framework. The framework is illustrated with numerical simulations of different fiscal spending rules that are consistent with sustainable paths of consumption out of oil and gas wealth over time.

Abstract

Russia’s large oil and gas reserves play a key role in its economic development. As with many other large oil exporters, Russia’s energy wealth is also posing numerous challenges to macroeconomic management. Although fiscal policy has saved a large part of the oil windfall in the Oil Stabilization Fund (OSF), this has not been guided by a consistent long-term framework. The framework is illustrated with numerical simulations of different fiscal spending rules that are consistent with sustainable paths of consumption out of oil and gas wealth over time.

I. Managing Russia’s Oil Wealth: An Assessment of Sustainable Expenditure Paths1

A. Introduction

1. Russia’s large oil and gas reserves play a key role in its economic development. Having the world’s largest gas reserves and seventh-largest oil reserves, Russia is one of the world’s main energy exporters. The share of the oil and gas sector in GDP is estimated at about 20 percent, generating more than 60 percent of Russia’s export revenues and more than 30 percent of fiscal revenues. In addition, rising oil prices have generated large terms of trade and real income gains, which have fueled strong domestic demand growth. Apart from their contemporaneous contribution to growth, Russia’s oil and gas reserves, if spent well, also offer an opportunity to raise long-run growth and accelerate Russia’s transition to a high-income, market economy.

2. As with many other large oil exporters, Russia’s energy wealth is also posing numerous challenges to macroeconomic management. How should monetary and fiscal policy deal with the swings in large external inflows associated with volatile oil prices? How much can be spent out of current oil revenues without risking a large fiscal correction in the future when oil and gas reserves have been depleted? How large is Russia’s oil wealth? How much should be saved for a “rainy day” and how much for future generations? While these are all important questions, this chapter is more limited in scope and will focus mainly on the challenges to fiscal policy, particularly those with a bearing on fiscal sustainability and, to a lesser extent, those that affect the contribution of fiscal policy to macroeconomic stabilization.

3. Fiscal policy has been key in recent years to maintaining macroeconomic stability amid rapidly rising oil prices. Russia is taxing and saving a large share of the oil windfall as the economy is running increasingly close to full capacity. Without this policy, the economy might have overheated, and a considerably faster real ruble appreciation would have risked choking the economic recovery under way since the 1998 crisis. The practice of using conservative macroeconomic assumptions in preparing the budget and the introduction of an Oil Stabilization Fund (OSF) based on a price trigger mechanism have been helpful institutional arrangements that have allowed the government to save a large share of what is in effect a permanent income gain.

4. Although fiscal policy has saved a large part of the oil windfall in the OSF, this has not been guided by a consistent long-term framework. Long-term sustainability considerations are lacking in Russia’s budget process. This chapter will highlight the importance of sustainability analysis for fiscal policymaking. It will argue that Russia’s fiscal framework could be strengthened by (i) using sustainability analysis to back fiscal targets; (ii) adopting a medium-term budgetary framework in which the non-oil balance (i.e., the balance net of oil-related revenues) plays a key role; and (iii) defining a clear rationale and transparent rules for the oil fund to gather public support for prudent fiscal policies. The chapter will also emphasize the need for periodic reassessments of long-term sustainability in the face of shocks affecting estimates of oil wealth, but will caution against abrupt policy adjustments following such reassessments.

5. This chapter presents a framework for fiscal sustainability analysis applied to Russia. The framework is illustrated with numerical simulations of different fiscal spending rules that are consistent with sustainable paths of consumption out of oil and gas wealth over time. Using a neoclassical growth model calibrated to match the growth experience of the Russian economy since 2000, the chapter provides a range of estimates for sustainable fiscal expenditure paths. The estimates generally indicate that government spending in Russia is below levels that would be unsustainable over the long run. This result is robust to oil price shocks of two standard deviations, based on historical prices. However, the model is very stylized, and several caveats should be noted. It does not consider short-run macroeconomic stability issues, nor does it have real exchange rate or monetary variables. Therefore, it does not take into account the potential for major negative effects on growth arising from rapid real effective exchange rate appreciation resulting, in turn, from heavy spending out of oil wealth (e.g., the Dutch disease). In addition, the model is deterministic and is not designed to shed light on how fiscal policy should respond to unexpected shocks. The main objective of the model simulations is to show how fiscal targets can be derived based on long-term sustainability analysis and to provide some illustrative scenarios for different fiscal rules.

6. The rest of the chapter is organized as follows. Section B discusses approaches to determine sustainable fiscal policy targets for oil-exporting countries and the institutional supports that may facilitate the implementation of policies based on such targets. Section C assesses Russia’s recent fiscal performance and its institutional framework. Section D presents estimates of sustainable expenditure paths for Russia. Section E concludes by summarizing the policy implications of the analysis.

B. Dealing with Oil and Gas Wealth

Maintaining fiscal sustainability

7. The finiteness of revenues from oil and gas reserves raises the question of how to avoid a large fiscal correction once these resources have been depleted. To assess the sustainability of given levels of expenditure and non-oil revenues, it is helpful to consider oil and gas reserves as assets that are part of the government’s financial wealth.2 The value of these assets is simply the net present value of the future stream of revenues they are expected to generate. If the revenues are consumed when they materialize, the wealth of the government declines. If, on the other hand, the revenues are invested in a financial asset, the composition of the government’s asset portfolio changes, but not its total value. By investing oil and gas revenues in financial assets, the government is preserving its wealth as it is merely converting one asset, natural resource wealth in the ground, into another one, a financial claim.

8. Under certain conditions, the optimal distribution of spending over time follows the so-called permanent consumption rule. Assuming that the taxation of the non-oil sector is constant at a given level and that utility is only a function of government spending, the maximization of the sum of discounted utility over an infinite horizon subject to a present-value budget constraint yields a constant level of spending and, therefore, a constant non-oil deficit.3 The optimal non-oil deficit is then equal to the return on the present discounted value of oil wealth.4 This deficit is less than the annual flow of oil and gas revenues (i.e., there is an overall surplus), thus allowing enough financial assets to be built up to finance the same deficit once oil and gas reserves have been depleted. This permanent consumption approach is becoming the standard in the analysis of fiscal sustainability for oil-exporting countries (OECs).5 In the standard formulation of this approach, all variables are scaled by non-oil GDP and, hence, the government targets a constant ratio of expenditure to non-oil GDP.

9. A practical alternative formulation of the permanent consumption rule obtains if variables are scaled by overall GDP and a constant expenditure-to-GDP ratio is targeted. Similar to the standard permanent consumption rule, this alternative formulation also defines a constant expenditure level that can be sustained indefinitely, without the need to increase the level of taxation in the non-oil sector (the ratio of non-oil tax revenues to non-oil GDP) after oil reserves have been depleted. The difference is in the variable by which expenditure is scaled: in this case it is overall GDP, whereas it is non-oil GDP in the standard permanent consumption approach.6 Another difference is that the alternative formulation tends to generate more front-loaded expenditure paths. Front-loading, of course, means that, under the alternative formulation, less is saved than under the standard formulation and that, therefore, a smaller primary deficit can be sustained once oil and gas reserves have been depleted (Figure 1).

Figure 1.
Figure 1.

Standard Permanent Consumption Rule (Left panel) and Alternative Permanent Consumption Rule (Right panel)

Citation: IMF Staff Country Reports 2006, 430; 10.5089/9781451975307.002.A001

Given a constant ratio of non-oil revenue to non-oil GDP (Tno/Yno), keeping constant the ratio of expenditure to non-oil GDP (G/Yno), as recommended under the permanent consumption approach, implies a constant ratio of non-oil deficit to non-oil GDP (Dno/Yno). However, this policy implies a gradually rising expenditure-to-GDP ratio (G/Y) because the share of non-oil GDP in total GDP grows over time. By keeping expenditure constant as a share of overall GDP, expenditure becomes more front-loaded than under the permanent consumption rule, both as a share of GDP—G*/Y—and as a share of non-oil GDP—G*/Yno (compare points G0 and G in the left panel to points G*0 and G*, respectively, in the right panel). The ensuing smaller surpluses in the phase of oil exploitation (compare the white triangles in the two panels) imply that only a smaller deficit (and a lower expenditure ratio) can be sustained after oil depletion.

10. Revenues from oil and gas are also highly uncertain, suggesting the need for additional precautionary savings. There is much uncertainty about the quantity, quality, and cost of extraction of oil reserves, which, in combination with uncertainty about future oil prices, makes it difficult to assess a country’s oil wealth and, correspondingly, its need for savings. In this context, a very conservative approach to dealing with oil wealth uncertainty is the so-called bird-in-hand rule According to this rule, the non-oil deficit must not exceed the real return on oil revenues that have already been saved and transformed into financial assets. This rule mandates saving the entire proceeds from oil and, therefore, generates a more back-loaded spending profile. It has the practical advantage that it does not require estimates of oil wealth. Norway has implemented this rule since 2001,7 but the rule might be less appropriate for countries with significant needs in terms of basic infrastructure and investment in human capital.

11. The more expenditure is front-loaded, the bigger the risk of an excessive real exchange rate appreciation. Rapid appreciation of the real effective exchange rate can cause deindustrialization (the Dutch disease) as it reduces the competitiveness of the country’s manufacturers and, conversely, makes imports more attractive. More generally, heavy dependence on resource revenues tends to skew economic incentives away from competition in the product markets and toward appropriation of rents.

Dealing with volatility

12. A crucial issue for the implementation of prudent fiscal policy is the identification of the temporary and permanent components of oil price changes. Oil prices are subject to permanent shocks, but most oil price changes also have a significant temporary component. Macroeconomic stabilization would require that expenditure policy should not be influenced by the temporary component of oil price changes.8 In that way, fiscal policy automatically tightens during peaks when revenues rise and relaxes during troughs when revenues fall.9 However, the permanent component of price shocks does alter oil wealth and, therefore, calls for a reassessment of the sustainable fiscal position.

13. In light of these price changes, a strong macroeconomic case can be made for decoupling public expenditure from oil revenues. Barnett and Vivanco (2003) present evidence that year-on-year fluctuations in the oil price have a large temporary component and, therefore, only a minor impact on oil wealth.10 The volatility of oil prices generates corresponding volatility in government revenues. If such volatility is transferred to expenditures, significant macroeconomic costs will ensue, including the reallocation of resources to accommodate changes in demand and relative prices, and real exchange rate volatility (Barnett and Ossowski, 2003).

14. A prudent strategy would require minimizing the adjustment of fiscal positions in response to oil price changes. Barnett and Vivanco (2003) argue that, given the cost of expenditure volatility, the risk that expenditure increases become “entrenched” and difficult to reverse, and the uncertainty surrounding the temporary/permanent breakdown, governments would be well advised to undertake only gradual adjustments, so as to avoid overshooting. In the same vein, Wakeman-Linn and others (2004) caution against large and sudden adjustments, as they can strain the government’s institutional capacity for planning, executing, and monitoring expenditures, resulting in substantial waste.

Helpful institutions

15. OECs have implemented a number of institutional arrangements to address the special challenges they face. These arrangements include “institutional supports”, such as fiscal responsibility legislation, fiscal rules or guidelines, medium-term expenditure frameworks, and nonrenewable resource funds (NRFs). With the exception of NRFs, these institutional arrangements are of course not specific to resource-rich economies.

16. A medium-term budgetary framework is a key institutional arrangement for OECs. The emphasis on the medium term allows annual deviations from average fiscal targets but at the same time calls for clear justifications for such deviations and provides limits to their extent. Such a framework can also be instrumental in gradually adjusting to permanent changes in oil prices.

17. Setting policy in terms of non-oil fiscal indicators is also helpful. Compared with traditional headline indicators, the non-oil balance is better suited to focus the political process governing the budget on fiscal sustainability. In addition, the non-oil balance provides information about the effect of fiscal policy on aggregate demand that complements the information gleaned from traditional indicators based on the overall balance (Box 1).

18. NRFs, if appropriately designed, are another example of “supporting” fiscal institutions. They are not a substitute for a strong commitment to sound policymaking, but a well-designed fund can help “sell” to the public and policymakers the importance of saving oil revenue and garner public support for prudent fiscal policies. Two critical ingredients of good fund design are: (i) integration with the budget in the context of a medium- and long-term fiscal framework; and (ii) stringent mechanisms to ensure transparency, good governance, and accountability that help prevent the misuse of resources.

19. NRFs can take various forms, ranging in scale from separate bodies to just government accounts. They can be set up as “stabilization funds,” aimed at reducing the impact of volatile revenues on the budget by transferring uncertainty and volatility from the budget to the fund. Or they can be set up as “savings funds,” aimed at addressing the issue of the exhaustibility of oil and creating a vehicle to store wealth for future generations. NFRs can also be arranged so as to pursue both objectives.

Fiscal Indicators

What is the “final effect” of fiscal policy on economic activity? is a question far beyond what can be asked of any summary indicator. Only simulations of full-scale macroeconomic models can shed light on final effects. On the contrary, summary indicators can provide only an indication of the impact of fiscal policy on aggregate demand (albeit ignoring the distortions to individual decisions caused by the tax system).

Among deficit/surplus measures, the overall balance is arguably the best indicator of the impact of fiscal policy. Blanchard (1990) argues that any improvement on the inflation-adjusted overall balance as an indicator of the impact of fiscal policy would involve estimating marginal propensities to consume, taking into account consumers’ expectations, and forecasting fiscal and macroeconomic variables.

The issue, however, arises of the appropriate measurement of the overall balance. The distinction between transactions “above the line” and those below it—that is, between non-financial and financial transactions—has a direct bearing on the size of the measured balance and entails some unavoidable degree of arbitrariness (Blejer and Cheasty, 1993).

In this respect, the special nature of oil-related revenues suggests the use of the “non-oil balance” as a supplementary indicator in the analysis of fiscal policy in OECs. Oil resources represent government wealth. Therefore, oil revenues can be seen as the result of the transformation of oil wealth into financial wealth.1 As such, they should not be included among income items in the government budget. They should be considered as financing items (i.e., “below the line”) to the extent that they are used to finance the excess of government spending over non-oil revenues.

The possibility of changes in oil taxation introduces further complications and suggests the need to monitor a third indicator, the “balance at constant oil price.” Pursuing further the analogy between oil and financial wealth, changes to oil taxation can be seen as determining changes in government wealth. By decreasing (increasing) tax rates on oil extraction, the government is in fact making (receiving) a capital transfer to (from) the private companies that have acquired the right to extract and sell oil. Such “transfers” will have an impact on aggregate demand. By definition, however, the non-oil balance is not affected by changes in oil taxation. The balance at a constant oil price will, on the contrary, generally highlight revenue changes due to changes in legislation.

Nevertheless, changes in the balance at a constant oil price provide very imprecise indications and should be used with care. The size of the change is not necessarily invariant to the assumed reference price. In extreme cases, where new legislation affects only revenues if the oil price is above a given threshold, the indicator may or may not signal a change in the impact of fiscal policy, depending on whether the reference price is above or below that threshold.

1

The definition of oil revenues will depend on the specific arrangements in place in each country. In general, it should include all revenues from extractive industries: dividends from the government’s participation in the sector, profit taxes, royalties, and export duties.

20. NRFs need to be integrated into the budget. In that way, the link between fiscal policy and asset accumulation is made clear and the emergence of two budgets and related fiscal management problems are avoided. Integration is realized, for instance, if the NRF is set up as a “financing fund.” Under such a fund, the budget is required to transfer oil revenues to the fund. In turn, the fund provides the budget with a reverse transfer equal to all budgetary outlays minus non-oil revenue (i.e., the non-oil balance). Hence, if oil revenue is larger than the non-oil deficit (i.e., if there is an overall surplus) resources are transferred to the fund; otherwise, the fund finances the overall deficit. This arrangement allows for easy monitoring of the relations between the relevant policy variables within a consistent framework. Assets are accumulated to the extent there is an overall surplus and the role of the non-oil balance in determining such a surplus is brought into sharp focus.

21. Integration of NRFs into a longer-term fiscal framework is also necessary because NRFs are tools of fiscal policy and do not constitute fiscal policy per se. For instance, in the absence of a commitment to a target for the non-oil balance, a fund would not directly curtail government spending. Since resources are fungible, governments could even borrow or run down other assets while accumulating resources in the fund, thus leaving government savings unchanged and failing to stabilize spending despite more stable budgetary revenues. Indeed, governments are likely to find borrowing particularly easy when resource prices are high.

22. Finally, transparency, good governance, and accountability are needed for the public to see how oil revenue is managed and spent. While a financing fund is little more than a government account, its operation will entail the accumulation of assets the effective management of which will be key to the success of the fiscal framework. Responsibilities concerning the management of the assets accumulated in a financing fund need to be assigned to a dedicated body and subject to specific guidelines. An asset management strategy would need to be defined, including prudential investment rules targeting the desired levels of risk, liquidity, and return. A clear allocation of responsibilities is important to ensure that those who manage and oversee the operation of the funds are held accountable. Transparency and freedom from political interference are key, as is regular and audited reporting.

23. Most oil funds are not set up as financing funds, because few OECs highlight the non-oil balance in their budgets or use a medium-term budget framework. In a sample of 17 OECs with NRFs, only 1 has a financing fund, and only one country follows a rules-based fiscal policy integrated in a medium-term framework (Table 1). Withdrawals from NRFs are typically managed on a discretionary basis. The prevailing purpose of NRFs in this sample of countries is stabilization. A savings motive is considered in only 6 countries. As a consequence, budgetary targets seldom reflect sustainability considerations.

Table 1

- Selected Oil Funds: Main Features

article image

24. Unsurprisingly, therefore, the experience with oil funds is mixed. Funds have been associated with a variety of fiscal policy outcomes. Davis and others (2003), for instance, analyze a sample of 12 OECs and conclude that “in some countries with NRFs, expenditure has tended to be less correlated with changes in the price of the resource” (p. 299) but add that “the establishment of the NRF did not have an impact on [the level of] government spending” (p. 302).

C. The Situation in Russia

25. The general government surplus has increased sharply, as oil prices have more than doubled since 2001. The price for the Urals oil blend rose from $23 per barrel in 2001 to almost $50 per barrel in 2005, contributing to robust real GDP growth of about 6 percent per annum (Table 2), and a strong fiscal position. The headline surplus increased by 5.4 percent of GDP during 2001–05; meanwhile, the primary surplus increased somewhat less (3.8 percent of GDP), because the reduction in outstanding liabilities and a negative real interest rate lowered interest spending from 2.7 percent of GDP in 2001 to 1.1 percent of GDP in 2005. With oil revenues rising from 6 percent of GDP in 2001 to 14 percent in 2005 (Box 2), the budget would still have balanced at an oil price of $26 per barrel in 2005 (against $19 per barrel in 2001).

Table 2.

Russian Federation: Selected Indicators, 2001–05

(In percent of GDP unless otherwise indicated)

article image

U.S. dollar per barrel of Urals blend, year average.

Change in the cyclically adjusted primary balance.

Change in the cyclically adjusted non-oil primary balance.

26. The strong fiscal performance is partly a reflection of certain key features of Russia’s fiscal framework. The law establishing the OSF mandated that revenues from the two major oil taxes (the extraction tax and the export tariff) should be deposited in the fund for the part corresponding to oil prices above $20 per barrel (Box 3).11 This requirement, together with the practice of using conservative oil price assumptions in the formulation of budget plans (the assumption for the 2004 budget was $20 per barrel), helped to set apart a large part of the revenue increase due to the rapid rise in oil prices during 2004.

27. However, in 2005 the increase in the overall surplus was accompanied by a deterioration of the non-oil deficit. The improvement in the headline balance of 3.2 percent of GDP with respect to 2004 was lower than the rise in oil revenues, which increased from 9.2 percent of GDP to 14 percent over the same period. Meanwhile, the non-oil primary deficit widened by 1.8 percent of GDP, largely because of a decline in non-oil revenues. Further underlying fiscal relaxation is expected in 2006–07.

Oil Taxation in Russia

Oil operations are governed primarily by a tax/royalty regime. Oil revenues accrue through three main instruments: (i) the corporate income tax (CIT) and dividends; (ii) the natural resources extraction tax (RET); and (iii) the export tariffs (ET).1 In 2005, oil revenues amounted to about 14 percent of GDP, of which 6 percent from ETs, 4½ percent from RET, and 3½ percent from CITs and dividends. The Russian tax system has been subject to frequent and unpredictable changes, which have had a negative impact on the business climate. Recently, with the gradual enactment of a comprehensive Tax Code, changes have aimed at streamlining the system.

The current CIT rate is 24 percent. Prior to the enactment of Part II of the Tax Code in 2002, the statutory rate was 35 percent. The Tax Code has improved the structure of corporate taxation in several respects, including (i) the definition of profit, as many expenses were allowed to be deducted that previously were not (e.g., interest on long term loans); (ii) the rules governing depreciation, as rates have become more closely linked to the economic life of the related asset; and (iii) the rules controlling “transfer pricing” (though these are not yet in line with international best practice). The CIT is mainly a regional tax. While there is a single profit tax system, the statutory tax rate of 24 percent is made up of federal component (5 percent), regional (17 percent), and local (2 percent) components.

The RET is a “royalty” levied on all extracted oil at a rate of 22 percent on the excess of c.i.f. Urals price over $9 per barrel. This has been effective since July 1, 2005. The tax rate has been gradually increased over the last three years, rising from 18.5 percent in 2003, to 18.8 percent in 2004, and 21.0 percent in the first half of 2005. The extraction tax mainly accrues to the federal budget (roughly 80 percent).

ETs are levied on oil, at increasing rates, on the excess of c.i.f. Urals prices over $15 per barrel.2 The rate is 35 percent for the excess over $15 per barrel up to $20 per barrel; 45 percent for the excess over $20 per barrel up to $25 per barrel; and 65 percent for the excess over $25 per barrel). This schedule has been effective since August 2004. Previously, the rate was 35 percent for the excess over $15 per barrel up to $25 per barrel, and 40 percent for the excess over $25 per barrel. ETs are an exclusively federal revenue source.

The tax/royalty mix appears rather unbalanced, with production-based instruments (the RETs and ETs) providing about two-thirds of revenues. However, the progressive rate schedule of ETs allows the government to share in the upside of oil price cycles. At the same time, the deduction of prices below $9 per barrel from the RET base and of prices below $15 per barrel from the ET base limits the fixed cost imposed on firms. The low reliance on the CIT reflects concerns with tax avoidance in a context of much intrafirm trade and relatively high corruption. The deduction and the sliding-scale elements of the Russian royalty system attempt to reconcile the resilience to tax avoidance of a quantity-based system with the investment incentives provided by a profit-based system. Figure A plots the marginal (MTR) and average (ATR) tax rates at different oil prices under the old (OE) and new (NE) RET and ET regimes. Figure B plots the producer’s take per barrel (PTPB) in dollars using a weighted average of the rates applying to exported and “domestic” oil under the new (AN) and old (AO) tax regimes.

Figure A.
Figure A.

Combined RET and ET Rates

Citation: IMF Staff Country Reports 2006, 430; 10.5089/9781451975307.002.A001

Figure B.
Figure B.

Producer’s Take Per Barrel

Citation: IMF Staff Country Reports 2006, 430; 10.5089/9781451975307.002.A001

1

Revenues from VAT on oil products are not considered here since the focus is on taxes on production.

2

Oil products are also subject to ETs, with rates linked to those applying to crude but lower.

The Russian Oil Stabilization Fund

The Oil Stabilization Fund (OSF) was established in 2004 with the objective of reducing the impact of fluctuations in oil prices on the resources available to the budget. The design of the fund—including accumulation, withdrawal, and investment rules—is subject to government regulation and the Budget Code of the Russian Federation. The government reports to the Parliament quarterly and annually. As of end-2005, the balance of the OSF stood at Rub 1,237 billion (5.7 percent of GDP).

Deposits. The OSF receives revenues from the export duty on crude oil and the resource extraction tax on oil. Specifically, the OSF receives the share of those taxes that is estimated to be due to the difference between actual oil prices and a threshold level, currently set at US$27 per barrel (Urals blend).1 Revenues from the export of gas and oil products are not included. In addition, federal budget surpluses are transferred to the fund at the end of the fiscal year. The taxes financing the OFS accounted for roughly 40 and 50 percent of overall oil and gas revenues in 2004 and 2005, respectively. The amounts derived from such taxes and deposited to the OSF in 2004 and 2005 (Rub 533 and Rub 1,393 billion rubles, respectively), corresponded approximately to one-fourth and one-third of overall oil and gas revenues in those years. In addition, the OSF received the unspent surpluses from 2003 and 2004 (Rub 106 and Rub 218 billion, respectively; in January 2006, the unspent surplus from 2005, Rub 222 billion, was deposited in the OSF).

Withdrawals. Money from the fund can be used to cover the budget deficit when the prices for oil in the world market are below the cutoff price. In addition, if the fund’s balance exceeds a predetermined threshold (currently set at Rub 500 billion), the difference between the actual balance and the threshold can be used for purposes specified in the budget law. As the fund had exceeded the threshold level in 2005, its surplus resources were used to prepay foreign debt (Rub 94 billion to the IMF; Rub 430 billion to Paris Club creditors; Rub 124 billion to Vnesheconombank), as well as to provide funding for the Russian Pension Fund (about Rub 30 billion).

Investment. The OSF is held in an foreign currency denominated account at the Central Bank with a remuneration determined by the yield on a basket of first-rate foreign sovereign bonds.

OSF: Deposits and withdrawals 2004-2005

(Billion Rubles)

article image

28. These developments suggest that despite a satisfactory performance so far, the Russian fiscal framework may not be robust to spending pressures from high oil prices. With respect to the discussion in Section B, several areas for improvements can be identified. First, budgetary targets are not grounded in sustainability analysis, which makes them susceptible to being seen as arbitrary. The assumptions concerning oil prices underlying the budget formulation have typically been conservative. However, lacking a formal link with sustainability, they can be (and have been) criticized as unrealistic. Second, budgetary targets are not set within a medium-term framework. This implies that there is no provision for gradual adjustment in the face of structural changes, and that pressure can mount for large, potentially wasteful, year-on-year adjustments.12 Third, the non-oil balance plays no official role in the budget process,13 and the OSF is not structured as a financing fund; this creates the impression that resources accumulated in the fund are “somewhat” additional with respect to those on which the budget is based, and therefore expendable.

D. Assessing Sustainable Expenditure Paths for Russia

29. To provide firmer foundations for medium-term budgetary targets, we use numerical simulations based on a long-run neoclassical growth model. Appendixes I and II provide full details of the model, while Appendix III discusses how it was calibrated to replicate the main trends in the Russian economy since the recovery from the 1998 financial crisis. Simulations are run to assess the sustainability of current fiscal policies and derive policy indications from the alternative fiscal rules discussed in Section B. The scenarios are purely illustrative, designed to show the effect of different fiscal rules on the long-term spending envelope.

30. This calibration and simulation exercise is based on assumptions about several exogenous variables. These exogenous variables include the rate of extraction of hydrocarbons (i.e., oil and natural gas) and the real rate of return on foreign financial assets. On the basis of current data on reserves and extraction rates, we project that oil will be depleted in 50 years and gas reserves in 300 years. The long-run real rate of return on foreign financial assets is assumed to be 3 percent. Appendix III provides further details about the way other exogenous variables are set.

31. The calibrated model is used to conduct long-run simulation exercises based on different spending rules. The simulated policy scenarios include (i) unchanged policy; (ii) the bird-in-hand rule; (iii) the “standard” permanent consumption rule scaled by non-oil GDP; and (iv) the “alternative” permanent consumption rule, with variables scaled by overall GDP. All simulations cover a 200-year span and are constrained to converge to a balanced growth path in the long run. Tax policy is invariant across the policy scenarios. Specifically, the average tax rate on non-oil income is the same across scenarios (35 percent, the 2005 level). Moreover, tax legislation applying to the oil and gas sector is assumed to stay the same as in 2005, so that the absolute level of oil and gas revenues is the same across all scenarios.14 These assumptions imply that in all scenarios overall revenues (excluding interest earnings on accumulated financial assets) decline over time as a share of GDP, eventually converging to 35 percent after oil and gas resources have been depleted. The assumptions also imply that the policy choice variable in all simulations is primary expenditure.

32. Each policy scenario is simulated for three different oil price assumptions. The central price scenario is based on the World Economic Outlook (WEO) world oil price forecasts: $66.5 per barrel in 2006, and $69.75 per barrel in 2007, followed by a gradual decline to $66 per barrel by 2011. Oil prices are assumed to remain constant in real terms from 2012 onward. Starting at the end of 2006, low- and high-price scenarios differ from the central one by ±$21.6 per barrel, or 1.96 times the standard deviation of nominal world crude oil prices over 1970–2005. In 2007, under the low-price assumption, revenues would be 1.5 percent of GDP lower than under the central price assumption; under the high-price assumption, revenues would be 1.3 percent of GDP higher than under the central price assumption (both in real terms). The differences across price scenarios decline monotonically from 2007 onward and gradually disappear as hydrocarbon resources approach the point of depletion. Because under each price scenario tax revenues from gas are projected to decline to less than 1 percent of GDP by 2100, we simplify computations by assuming that oil and gas revenues become nil after 100 years in all simulations (recall that oil is projected to be depleted in 50 years).

Unchanged policies

33. The unchanged policy scenario assumes that primary expenditures remain constant as a share of GDP at the 2005 level (32 percent; Table 3 and Figure 2). Reflecting the assumed path for revenues, the primary surplus would gradually decline from 9 percent of GDP in 2005 to 6–8 percent of GDP in 2010 (depending on the oil price assumption). However, the non-oil primary deficit would also decline, from 5 percent of GDP in 2005 to 0–4 percent of GPD in 2010, as non-oil revenues increase as a share of GDP. The two balances would eventually converge to the same value (a surplus of 4 percent of GDP) once oil and gas resources have been depleted.

Table 3.

Russia: Simulation results for different fiscal rules

(In percent of GDP)

article image
Figure 2.
Figure 2.

Unchanged Policy Scenario

Citation: IMF Staff Country Reports 2006, 430; 10.5089/9781451975307.002.A001

34. The unchanged policy scenario shows a very large buildup of international assets. The primary surpluses sustained throughout the simulation horizon, under all price assumptions, generate what seems an excessive accumulation of financial assets. The ratio of net foreign financial assets to GDP surpasses 500 percent by end-2065. The non-oil primary deficit would narrow rapidly, turning into a surplus even before the depletion of oil and gas resources. Such a policy would likely sooner or later become untenable as spending pressures rise, given the need for improving infrastructure and social services in Russia.

Bird-in-hand rule

35. Implementing the bird-in-hand rule would call for an immediate fiscal tightening, which would then be gradually unwound over the medium term. This rule stipulates up-front savings of oil revenues and spending only out of the return on accumulated financial assets. As Russia is running a non-oil deficit in excess of the return on the assets in the OSF, this rule would stipulate an initial tightening of the fiscal stance: primary expenditures are brought into line with non-oil revenues before they are allowed to rise following the increase in accumulated financial assets. In particular, primary expenditures would first decline by about 5 percent of GDP in 2006 from the 2005 level under all price scenarios (and by an additional one percent of GDP in 2007 under the high-price scenario; Figure 3). Subsequently, they would rise to 29–32 percent of GDP—depending on the price assumption—by 2010. After the tightening in 2006 (and also in 2007 under the high-price scenario), the primary surplus would be reduced rapidly, reflecting both higher spending and lower revenues. By 2010, it would be below the 2005 level under the central and low price scenarios. The non-oil primary deficit would increase more slowly (recall that the share of non-oil revenue in GDP is assumed to increase over time) and remain higher than the 2005 level in 2010 by about 4 percent of GDP.

Figure 3.
Figure 3.

Bird-in-Hand Rule

Citation: IMF Staff Country Reports 2006, 430; 10.5089/9781451975307.002.A001

36. However, the bird-in-hand rule would lead to a significant fiscal relaxation in the steady state, accompanied by a large accumulation of financial assets. The large up-front savings generate a significant accumulation of assets—already at 36–62 percent of GDP by 2010, and ranging between 164 and 402 percent of GDP in the steady state. This, in turn, allows the primary expenditure-to-GDP ratio to rise by 5–10 percent of GDP during 2010–55 to eventually reach 38–41 percent of GDP in the steady state. Over the same period, both the primary and the non-oil primary balance would continue to deteriorate (by 7–14 and 2–4 percent of GDP, respectively), eventually converging to a steady state primary deficit of 2–6 percent of GDP, to be financed out of the return from the large positive net asset position.

Standard permanent consumption rule (scaled by non-oil GDP)

37. Simulation of the standard permanent consumption rule suggests tightening of the fiscal stance over the short run but a mild relaxation in the medium term. The mechanics of the rule would call for a short-lived tightening under all price scenarios. The primary expenditure-to-GDP ratio is lower than the 2005 level by 2–4 percentage points in 2006 under all price scenarios (and by additional 1½ percentage points in 2007 under the high-price scenario; Figure 4). Thereafter, the primary expenditure-to-GDP ratio would rise by 0.5–1.5 percent of GDP per year, depending on price assumptions, to reach 31–33 percent of GDP in 2010—roughly the same level as in 2005. Given the declining path of overall revenues, the primary surplus would fall to 4–9 percent of GDP by 2010 from 9 percent of GDP in 2005; meanwhile, the non-oil primary deficit would fall to a range of 2–3 percent of GDP by 2010 from 7 percent of GDP in 2005.

Figure 4.
Figure 4.

Standard Permanent Consumption Rule

(Scaled by non-oil GDP)

Citation: IMF Staff Country Reports 2006, 430; 10.5089/9781451975307.002.A001

38. A significant fiscal relaxation—but smaller than under the bird-in-hand rule—materializes in the steady state, reflecting the benefits of a steady accumulation of financial assets. The primary expenditure-to-GDP ratio would rise by 4–7 percent of GDP during 2010–55, eventually reaching levels 4–8 percent of GDP higher than in 2010. Over the same period, both the primary and the non-oil primary balances continue to deteriorate (by 6–11 and 0–1 percent of GDP, respectively), eventually converging to a steady state deficit of 2–4 percent. Financial assets converge to a steady state level of 123–289 percent of GDP.

Permanent consumption rule (scaled by overall GDP)

39. This rule would call for an immediate fiscal relaxation. As estimates of the constant sustainable primary expenditure level range from 36 to 37 percent of GDP, depending on price scenarios (Figure 5), this implies an increase in primary expenditures of 4–5 percent of GDP as of 2006. As expected, expenditures are higher than under the standard permanent consumption rule over the same period (Figure 6). The primary surplus keeps declining also after 2006, reflecting the reduction of revenues. By 2010, the primary surplus would be 6–7 percent of GDP lower than in 2005. The mechanics of the rule also imply a sudden widening of the non-oil primary deficit, which is then gradually reduced. By 2010, under the low-price scenario, the non-oil primary deficit would be roughly the same as in 2005; under the high-price scenario, the deficit would still be 4 percent of GDP higher than in 2005.

Figure 5.
Figure 5.

Alternative Permanent Consumption Rule

(Scaled by overall GDP)

Citation: IMF Staff Country Reports 2006, 430; 10.5089/9781451975307.002.A001

Figure 6.
Figure 6.

Comparison of Different Fiscal Rules

Citation: IMF Staff Country Reports 2006, 430; 10.5089/9781451975307.002.A001

40. The overall primary balance deteriorates further in the steady state, while the non-oil deficit declines. In the steady state, the two balances converge to a deficit of 1 percent of GDP. As expenditure remains constant as a share of overall GDP in this variant of the permanent consumption rule, the dynamics of fiscal balances mechanically reflect the assumptions concerning tax policy, whereby overall revenues decline and non-oil revenues increase as a share of GDP until they reach the common steady state value of 35 percent. The equilibrium deficit would be financed by the return on accumulated assets (amounting to 41–98 percent of GDP in the steady state).

A comparison across scenarios

41. The simulated growth paths of the non-oil economy are similar across scenarios, except for the bird-in-hand rule. Growth rates under all scenarios eventually converge to the same exogenously determined rate and differ only during the transition to the steady state.

42. For all rules, the framework underling the simulations provides a measure of welfare. In particular, one way to measure the welfare of the economy is to calculate the sum of a discounted utility stream of a representative household for an equilibrium consumption path.15 Simulation results suggest that welfare tends to be higher the more front-loaded the expenditure path is. The alternative permanent consumption rule (scaled by overall GDP) ranks highest, followed by the standard permanent consumption rule (scaled by non-oil GDP), the unchanged policy scenario, and, finally, the bird-in-hand rule.

43. Both welfare and growth comparisons need to be interpreted with great care. Underlying the welfare ranking is the large catch-up potential of the Russian economy. The initial capital stock of the simulated economy turns out to be a mere 30 percent of the steady state capital stock. During the transition to the steady state, household consumption remains below the steady-state level. Under these conditions, it is not surprising that saving more today for higher spending in the future (bird-in-hand) produces a lower welfare ranking. However, owing to the nature of the model, the welfare ranking discussed above ignores the limitations of absorptive capacity of the economy or administrative capacity of the government, which could constrain the speed of expenditure adjustments. If government spending is adjusted rapidly, resources might be wasted. Furthermore, any cost associated with high inflation or Dutch disease are not considered in the analysis since the model includes only real variables and does not distinguish between tradables and non-tradables.

44. The expenditure paths under the two variants of the permanent consumption rule differ less than it appears at first sight. Once the larger front-loading allowed in the alternative scenario (with scaling by overall GDP) is tempered to take into account the possible risks to the economy of an abrupt spending increase, the resulting medium-term spending envelope is unlikely to be very different from the one implied by the standard scenario (with scaling by non-oil GDP). Since the results are close and the former rule is perhaps more transparent and easier to implement, as it expresses expenditure as a percent of overall GDP rather than non-oil GDP, it would seems to be the preferred rule for Russia at this juncture.

E. Conclusion

45. The sharp rise in oil prices in recent years has boosted Russia’s GDP growth, but has also posed serious challenges to policymakers. With the absorptive capacity of the economy nearly exhausted, taxing and saving a large part of the oil windfall have helped to avoid higher inflation and excessive real ruble appreciation. Furthermore, because administrative capacity is limited, this policy has also prevented the waste that could have resulted from very large increases in government expenditures, especially since structural reforms—an area where the oil wealth might be well spent—have stalled. However, strong pressures to spend more of the oil windfall continue to mount.

46. Russia’s fiscal institutions have helped to insulate oil revenues from spending pressures, but further strengthening is needed. Institutions cannot be a substitute for a commitment to sound fiscal policy, but, if well-designed, strong institutions can help garner support for fiscal discipline. The employment of conservative macroeconomic assumptions in the preparation of the budget and the creation of the OSF in 2004 have been useful in this regard. This chapter has suggested a number of additional improvements. Specifically, it has suggested that fiscal targets should be embedded in a multiyear fiscal plan that is consistent with the government’s long-term view on the use of Russia’s oil wealth. In this context, the OSF could be transformed into a financing fund to which all hydrocarbon revenues accrue. This change would help focus the budget process more on long-term considerations as fiscal policy would be framed with reference to the non-oil balance—which is the relevant measure when assessing fiscal sustainability in an oil-exporting country.

47. To illustrate how fiscal targets can be given firmer foundations, this chapter has estimated long-term spending envelopes based on a calibrated neoclassical general equilibrium model. Simulations assuming unchanged tax policy have been carried out for different spending rules. The alternative permanent consumption rule, based on constant expenditure as a share of overall GDP, would seem to be most appropriate for Russia at this juncture. This rule allows sufficient room to accommodate many of Russia’s spending needs over the medium term without creating a potentially unsustainable fiscal position. Estimates suggest that under this rule Russia could increase primary expenditure by about 5 percent of GDP over the medium term (and sustain the same level over the long run).16 The results are sensitive to several important assumptions, not least the future path of oil prices and world interest rates, and to the preferred degree of front-loading. This sensitivity also highlights the need to update long-term forecasting exercises routinely to ensure that the budget process is based on accurate estimates of oil wealth and takes into account ongoing structural changes in the economy. In addition, because of the highly stylized nature of the model, the estimated spending envelopes should be interpreted as indicative.

48. Any fiscal relaxation implied by a spending rule should be gradual and supported by structural reforms or other expenditures or tax cuts that support long-term growth. Fiscal loosening should allow industries sufficient time to adjust through improvements in productivity and, hence, avoid overshooting the real effective exchange rate. Strengthening the public administration would help to make spending more efficient, while structural reforms and other expenditures (or tax cuts) that promote the accumulation of human capital and improve the investment climate would help to expand the absorptive capacity of the economy, thereby creating room for fiscal relaxation without causing macroeconomic instability.

Appendix I: The Model

1. This appendix builds a long-term analytical framework to study preferred fiscal rules in countries endowed with exhaustible resources. It builds a neoclassical growth model that is augmented to take into account the effects of public capital on productivity of firms. A preferred fiscal rule for economies facing exhaustible resource-related revenues is identified based on the welfare ranking.

2. The analysis in this chapter is based on some simplifying assumptions. In particular, the rate of extraction (i.e., output) of natural resources, the amount of resources in the ground, and the tax system are taken as given and will not be endogenized. Furthermore, for modeling purposes, it is assumed that the government has a savings technology of holding the worlds’ financial assets, while the only savings technology available to households is to accumulate physical capital.

3. The economy is a decentralized economy and is populated by three types of economic agents. First, a large number of households whose measure is normalized to unity. They are indexed by ηε[0,1]. The size of each household grows at an exogenous rate, n. Hence, n represents the growth rate of the population. Households own capital stock that can be used for producing output in the nonnatural resources sector. They also own natural resource reserves that generate a stream of output at exogenous rates over a finite period.17 Second, the economy is populated by a single firm in the nonnatural resources sector which takes prices as given in making its decisions.18 Third, the government also engages in economic activity: taxing households and making expenditures. The government announces and commits itself to its fiscal policy for future dates. Households and the firm make their decisions after observing the announced fiscal policy. The government is assumed always to implement the announced policy. Therefore, no time-inconsistency problem arises.

4. Timing of events in the economy is discrete, and no uncertainties exist. At the start of each period, households rent their capital stock and labor services to the firm. The firm employs them and uses public capital available for free to produce a single consumption-capital good. After production, payments for production factors, taxation, government expenditures, sales and purchase are made simultaneously. The firm returns the undepreciated portion of the capital stock to the households and also makes payments for the use of production factors. The natural resources sector also generates output. The generated output is exchanged for consumption-capital goods in international markets, which are then distributed to households. Income from the natural resources sector is taxed at an exogenous rate. Households are also subject to a lump-sum tax. The government purchases investment goods, makes lump-sum transfers to households, and also purchases consumption goods. Any primary fiscal surplus is saved in foreign financial assets, while fiscal deficits can be financed by withdrawing the financial assets already accumulated. Households use their after-tax/transfer incomes to purchase consumption and investment goods from the firm. After all transactions are made, households consume consumption goods. Undepreciated private capital is augmented by investment goods and is carried over into the following period. Undepreciated public capital is augmented by investment goods the government purchased. The government consumes consumption goods.19

5. Some conventions about notation follow. Throughout this appendix, the superscript f indicates quantities chosen by the firm, while the subscript t indicates quantities either in period t (in the case of flow variables) or at the beginning of period t (in the case of stock variables). The subscripts p and g indicate variables chosen by the private sector and the government, respectively. Uppercase and lowercase variables represent aggregate and individual (both firm and household) variables, respectively.

The firm

6. A constant-returns-to-scale production technology is available for the firm to transform labor input, ltf, private capital, kp,tf, and aggregate public capital normalized by aggregate labor input, Kg,t / Lt, into a yt unit of consumption-capital good, where Kg,t and Lt are aggregate public capital stock and labor input, respectively. Inputs ltf and kp,tf are under the direct control of the firm, while an aggregate variable, Kg,t/Lt, is outside the control of the firm.

7. The following Cobb-Douglas production function is assumed:

yt=yt⁢(kp,tf,ltf)=[ϕ⁢(At,Kg,tLt)⋅ltf]1−α⁢(kp,tf)α,(A1)

where αε(0,1) is the substitution parameter and ϕ(∙) > 0 represents the level of total factor productivity (TFP), which is a function of the technology level, At, and the ratio of public capital to labor. The technology level is assumed to grow at a constant rate, y, that is exogenous to the economy.

8. We assume that production depends on the normalized public capital, Kg,t / Lt not on the absolute level of the public capital, Kgt, to capture congestion effects. The assumption implies that, the higher the level of economic activity (approximated by the economy-wide labor input) is, the larger is the public capital stock required to maintain its efficiency in production.20 This assumption also ensures consistency of a particular class of the model with balanced growth in the steady state, one of the stylized facts of economic growth documented by Kaldor (1963). Given a constant ratio of Kg, t to aggregate output, and given a function φ(.,.)that is linear in At and Kg,t / Lt the model can generate a balanced growth path on which private capital and output grow at a constant rate (1 + γ)(1 + n) −, driven by the exogenous productivity growth, 1 + γ, and the exogenous population growth, 1 + n.21 As a result, public capital also grows at the rate (1 + γ)(1 + n) −. Therefore, the model can be redefined using detrended variables. Throughout the chapter, a linear function

ϕ(At,Kg,tLt)=At+θKg,tLt

is assumed. Appendix II discusses the detrending in detail.

9. The firm maximizes the profit in each period, given the price of single consumption-investment goods, pt, the rental rate, pt.rt, and the wage rate, pt.wt. rt and wt are functions of aggregate private capital Kp, and public capital Kg,t, which the firm takes as given. These factor prices are written as rt(Kp,t, Kg,t) and wt(Kp,t, Kg,t), respectively. Specifically, the firm produces a single consumption-investment good by employing labor and capital, and then sells consumption goods cp,tf and investment goods ip,tf to the households at price pt. Therefore, the firm’s maximization problem can be written as

max{cp,tf,ip,tf,kp,tf⁢ltf}pt⋅[(cp,tf+ip,tf)−rt⁢(Kp,t,Kg,t)⋅kp,tf−wt⁢(Kp,t,Kg,t)⋅ltf],(A2)

subject to the constraint

cp,tf+ip,tf≤yt.

This problem can be rewritten as

max(kp,tfltf)pt⋅[yt(kp,tf,ltf)−rt(Kp,t,Kg,t)⋅kp,tf−wt(Kp,t,Kg,t)⋅ltf].(A3)

10. The aggregation of firm variables is straightforward. Since there is only one firm in the economy, yt,kp,tf, and ltf are all aggregate variables.

Households

11. Households are assumed to derive utility by consuming consumption goods and services, denoted by cp,t.22 The discounted sum of period-by-period utility is written as

∑t=0∞βtu(clp,t),(A4)

where β is the discount factor. As a baseline case, we study the following constant relative risk aversion (CRRA) utility function:

u(cp,t)=cp,t1−σ1−σ,σ>0.(A5)

12. The households provide labor services and capital to the firm. Since households’ utility does not depend on labor, households provide all their labor endowment inelastically. We normalize the initial period labor input to unity: l0 = 1. Private capital is solely owned by a household and is denoted by kp,t. The households receive factor payments from the firm equal to wt ·lt + rt ·kp,t. The households also receive income generated by the natural resources sector, op,t ·qt, where op,t is the oil output (in per capita terms, net of taxes) and qt is the export price normalized by the price of consumption-investment goods, pt.

13. Aggregation of household variables is straightforward. Given the capital stock held by individual households at the beginning of period t, aggregate private capital is defined by Kp,t=∫01kp,tdη, while aggregate labor services at the beginning of period t is (1 + n)t. In a similar manner, aggregate consumption and aggregate oil output (net of taxes) are defined as Cp,t=∫01cp,tdη and Op,t=∫01op,tdη, respectively.

14. In each period t, the households split their incomes into consumption and investment in private capital. We assume that the economy is small and open and that the households and the firm of this economy can freely engage in international trade to exchange the domestically produced consumption-capital good with the identical good produced abroad. An arbitrage condition for this good ensures that the domestic price is equal to the international price. We normalize the international price of this good to one, which implies that the domestic price is always equal to one in equilibrium (see below). We also assume that households receive transfers net of any lump-sum taxes from the government, denoted by xt. Households face a budget constraint that sets their total spending less than or equal to their income:

cp,t+ip,t≤wt⋅lt+rt⋅kp,t+xt+op,t⋅qt,(A6)

where ip,t is the investment in private capital. Note that households do not have any savings technology other than accumulating physical capital. Private capital follows the following law of motion:

kp,t+1=ip,t+(1+δ)kp,t,(A7)

where δp is the depreciation rate and satisfies 0 < δp < 1.

15. The state of the economy that affects households’ decision making is summarized by a triple (kp,t, Kp,t, Kg,t). This reflects the fact that the households’ incomes are affected by rental and wage rates that are functions of aggregate private capital, Kp,t, and public capital, Kg,t. Moreover, the households maximize their lifetime utility (A4), and thus the future path of aggregate states matter to the households’ decision making. By denoting aggregate private and public investments by Ip,t and Ig,t, respectively, the future path of aggregate states is represented by the following laws of motion of the aggregate private capital and public capital:

Kp,t+1=(1+δp)Kp,t+Ip,t(A8)

and

Kg,t+1=(1−δg)Kg,t+Ig,t,(A9)

respectively, where δg is the depreciation rate of public capital and satisfies 0 < δg < 1.

16. A representative household’s maximization problem can be written as follows:

max{cp,t,ip,tkp,t+1}∑t=0∞βtu(cp,t),(A10)

subject to the budget constraint (A6), the law of motion of private capital (A7), and the laws of motion of aggregate capitals (A8 and A9).

The government

17. One of the government’s sources of revenue is the tax on export receipts of the endowment (oil) sector, og, t ·qt, where og,t is the volume of oil tax (in per capita terms) and qt is the export price. Aggregate volume of oil tax is defined as Og,t=∫01og,tdη. The government also collects a lump-sum tax from the households. We assume that the government can save tax revenues to increase its holding of the world’s financial assets (e.g., U.S. treasury bonds) which yields a net return of re t. Alternatively, the government can use tax revenues to purchase a consumption-investment good to augment the stock of public capital or consume the good. Lastly, the government can simply transfer tax revenues to the households. The lump-sum transfer, net of lump-sum taxes, is denoted by xt in per capita terms. The aggregate lump-sum net transfer is defined as Xt=∫01xtdη.

18. The government faces a sequence of budget constraints (a period-by-period budget constraint). Any fiscal deficit or surplus in period t, including interest payments on the government’s holdings of the world’s financial assets at the beginning of period t, is reflected in a change in the government’s holdings of the world’s financial assets at the beginning of period t+1. The government’s budget constraint can be written as

Ig,t+Cg,t+Xt+Bt+1=qt⋅Og,t+(1+re,t)Bt,(A11)

where Ig,t is the public spending to augment public capital stock, Cg,t is the public spending on the consumption good, Xt is the aggregate lump-sum net transfer, Bt is the government’s beginning-of-the-period holdings of the world’s financial assets, and re,t is the return on world’s financial assets.

Equilibrium

19. A sequence of market equilibrium is a set of prices and quantities, {rt, wt} and {cp,t,cg,t,lt,ltf,ip,t,ig,t,kp,t,kp,tf} and transfers {xt} such that

  • given transfers and prices, the quantities solve the household problem;

  • given the prices, the quantities solve the firm problem;

  • factor market-clearing conditions
    Lt=ltfand Kp,t=kp,tf

    are satisfied;

  • given the prices and quantities, the government’s budget constraint is satisfied; and

  • the resource constraint

Cp,t+Cg,t+Ip,t+Ig,t≤yt+qt⋅(Op,t+Og,t)+(1+re,t)⋅Bt−Bt+1(A12)

is satisfied.

Solution of the model

20. The firm’s optimization implies marginal returns of inputs are equal to marginal products:

rt(Kp,t,Kg,t)=∂yt∂kp,tf=αytkp,tf(A13)

and

wt(Kp,t,Kg,t)=∂yt∂ltf=(1−α)ytltf.(A14)

Since kp,tf=Kp,t and ltf=Lt in equilibrium, these conditions can be rewritten as

rt(Kp,t,Kg,t)=αytKp,t(A15)

and

wt(Kp,t,Kg,t)=(1−α)ytLt.(A16)

21. The first-order condition for the household’s problem is

ucp,t=βucp,t,t+1(rt+1+1−δp).(A17)

Given the functional forms, the firm’s first-order condition, and the market-clearing condition, this Euler equation reduces to

(cp,t+1cp,t)σ=β[αyt+1Kp,t+1+1−δp].(A18)

22. The value of the domestically produced consumption-capital good is equal to the factor payments by the firm. As a result, the households’ factor income is equal to the value of inputs.

23. (A18) is a variant of the familiar second-order Euler difference equation in {kp,t, kp,t+1, kp,t+2} for the optimal growth model. However, the state space method cannot be used to solve this difference equation since TFP in the production function is not assumed to follow a stochastic process with Markov properties. The problem can be expressed using value functions, and thus the value function iterations method can be used to obtain an equilibrium.23

Appendix II: The Detrended Model

1. The model economy exhibits a balanced growth if investment in public capital grows at the rate (1 + γ)(1 + n) −1 and, as a result, Kg t grows at the same rate (1 + γ)(1 + n) −1. A heuristic explanation for the existence of a balanced growth path is provided later in this appendix. On a balanced growth path under this assumption, lt grows at the rate n and ct, ip,t,kp,t and yt grow at the rate (1 + n)(1 + γ) −1. By assumption, At grows at a constant crate γ.

2. Detrended variables are defined by dividing the original variables by their respective growth rates. Specifically, detrended variables, represented by notations with hats, are defined as follows:

l^t=lt(1+n)t,l^tf=ltf(1+n)t,A^t=At(1+γ)t,o^p,t=op,t[(1+n)(1+γ)]t,o^g,t=og,t[(1+n)(1+γ)]t,c^p,t=cp,t[(1+n)(1+γ)]t,c^g,t=cg,t[(1+n)(1+γ)]t,i^p,t=ip,t[(1+n)(1+γ)]t,i^g,t=ig,t[(1+n)(1+γ)]t,k^p,t=kp,t[(1+n)(1+γ)]t,k^p,tf=kp,tf[(1+n)(1+γ)]t,  and x^t=xt[(1+n)(1+γ)]t.

The detrended aggregate economy-wide variables, L^t,C^p,t, C^g,t,I^p,t,X^t,I^t,K^p,t,O^p,t,O^g,t⁢ andK^g,t can be defined in a similar manner. These definitions imply that public capital per worker, kg,t=Kg,t/Lt, grows at the rate γ, and the detrended variable is defined as ̂kg,t = kg,t/(1+γ)t.

3. The original economy can be recast in a detrended version. To define such an economy, both the firm’s problem (A2) and the households’ problem (A10) need to be reformulated as problems based on detrended variables, as discussed below. Detrended output is defined as

y^t≡(A^t+θk^g,t)1−α(l^tf)1−α(k^p,tf)α.

By substituting detrended variables and defining prices for the detrended economies as r^t≡rt⁢ (K^p,t,K^g,t,) and w^t≡wt⁢ (K^p,t,K^g,t), the first-order conditions for the firm’s optimization problem (A13) and (A14) become

r^t=αy^tk^p,tf(A19)

and

w^t=(1−α)y^tl^tf.(A20)

Similarly, by substituting detrended variables into the household’s utility function, the budget constraint, and the law of motion of capital in (A7), (A8), and (A9), the household’s transformed maximization problem becomes

max{c^t,ip,t}∑t=0∞{β[(1+γ)(1+n)]1−σ}tc^t1−σ1−σs.t.     c^p,t+i^p,t≤r^t⋅k^p,t+w^t⋅l^t⋅+x^+qt⋅o^p,t,(A21)(1+γ)(1+n)k^p,t+1=(1−δp)k^p,t+i^p,t,(1+γ)(1+n)K^p,t+1=(1−δp)K^p,t+I^p,t,and(1+γ)(1+n)K^g,t+1=(1−δg)K^g,t+I^g,t,

given r^t, andw^t.

The Euler condition for the household’s optimal intertemporal substitution, equation (A17), is rewritten as

u′(c^p,t+1)u′(c^p,t)=(1+n)σ(1+γ)σβ(r^t+1+1−δp).(A22)

The government’s budget constraint can be redefined as

I^g,t+C^g,t+X^t+(1+γ)(1+n)B^t+1=qt⋅O^g,t+(1+re,t)B^t.(A23)

4. Given the reformulated problem of the firm and households, as well as the government’s budget constraint, the competitive equilibrium of the detrended model can be defined in a similar way to the original model.

5. With the reformulated detrended model, it is straightforward to confirm that the model exhibits balanced growth. l^t is constant by definition. Under fiscal rules that ensure constant per capita expenditures after the depletion of oil and gas, Îg,t, Ĉg,t, and K^g,t remain constant after the depletion of gas and oil. Suppose that K^p,t is constant. Then all detrended variables that appear in optimization problems for households and firms are constant. From (A19) and (A20), the rental rate, ̂rt and the wage rate, ŵt are constant since they are functions of constant aggregate variables K^g,t, K^p,t, and L^t in equilibrium. A constant K^p,t implies a constant i^p,t by virtue of the law of motion of K^p,t. By virtue of the household’s budget constraint, i^p,t is also constant, which also satisfies the Euler equation (A22).

6. Now that the detrended economy has been defined, it is possible to solve analytically for the balanced growth path of a competitive economy of the detrended model. From the Euler equation (A22),

r^=(1+γ)σ(1+n)σβ−1+δp.(A24)

For any value of the government’s financial asset on a balanced growth path, B^,I^g⁢ andX^ are determined to satisfy the government’s detrended budget constraint (A23). From the law of motion of K^g,t,K^g=1(1+γ)(1+n)−(1+δg)I^g.Given r^⁢ andK^g,t, the firm’s optimization condition (A19) implies that

k^p=(αr^)11−α(A^L^+θK^s).(A25)

Given k^p⁢ andK^g, consumption, ĉp, and the wage rate, ŵ, follow from the household’s budget constraint in (A21) and the first-order condition for the firm’s profit maximization (A20), respectively:

c^p=w^l^+r^k^p+x^+qt⋅o^p,t−i^p=y^+x^+qt⋅o^p,t−[(1+γ)(1+n)k^p−(1−δp)k^p](A26)=(A^+θK^gL^)1−αl^1−αk^pα+x^+qt⋅o^p,t−[(1+γ)(1+n)k^p−(1−δp)k^p],

and

w^=(1−α)(A^L^+θK^g)1−αK^paL^.(A27)

Appendix III: Calibration and Simulation

1. This appendix provides a detailed discussion on the calibration of the model and the simulations that are summarized in the main text. It first discusses the exogenous variables that are necessary for conducting simulations, then how parameter values underlying household preference and production technology, as well as initial conditions, are chosen.

Exogenous variables

2. In order to conduct a simulation exercise based on the model, several exogenous variables need to be set in advance. Paths of the following exogenous variables, along with fiscal policy variables that are discussed below, are predetermined for the entire simulation period:

  • Labor force. As a proxy for the labor force, employment is used as a factor input. Historical data on employment are available from official sources. The long-term projection of Russia’s labor force is approximated by a fraction of the projected working-age population (available in the UN population database that covers the period up to 2050). The average growth rate of working-age population for 2045–50 (-0.55 percent) is applied to extend the projection of labor force from 2050 to 2150.

uA01app03fig01

Employment

Citation: IMF Staff Country Reports 2006, 430; 10.5089/9781451975307.002.A001

  • Government hydrocarbon revenues. Historical data on hydrocarbon revenues are available for 2000–05. The projection is set to keep the hydrocarbon revenues-to-projected hydrocarbon GDP ratio constant at the 2005 level. Projected hydrocarbon GDP is based on the following long-term assumptions about output and prices. Oil output growth gradually picks up over 2006–11 to reach 4 percent per annum in 2011–12, followed first by a gradual deceleration of growth and then declines in output itself. The stock of proven, probable, and possible reserves, estimated at 149.3 trillion barrels for 2004, will be depleted in 2048. Gas output growth is assumed to be 2 percent annum until 2011, 1 percent per annum for 2012-31, 0.5 percent per annum for 2032–51, and zero afterward. The stock of proven and unproven reserves, estimated at 236.1 trillion cubic meters for 2004, will not be depleted for another 300 years. However, government gas revenues are assumed to disappear in 2105. Oil and gas export prices for the period 2006–11 are based on the WEO assumptions, with some adjustments made to take into account differences between international and Russian export prices. For 2012 onwards, prices are assumed to remain constant in real terms vis-à-vis advanced economies’ long-term projected consumer price inflation of 2 percent per annum. Nominal oil and gas export prices are normalized by non-oil and gas prices deflator (both for historical and projection periods) to derive real values of oil and gas export receipts.

  • Long-run rate of technological progress. This rate has been set to ensure 1.5 percent per annum long-term real output growth rate, given the labor force discussed above. In particular, a 2.05 percent long-run rate of technological progress is implied by the 1.5 percent growth rate of long-run real non-oil output and the -0.55 percent growth rate of labor force, since balanced growth in the model is driven purely by technological progress and labor force growth. Somewhat higher rates of technological progress are assumed for 2000–30 to capture the catch-up process and to be consistent with the growth accounting exercise under an assumed effect of public capital on growth (discussed in more detail below). In particular, rates of technological progress are assumed to be 4.05 percent for 2000–11, 3.55 percent for 20012–18, 3.05 percent for 2019–24, and 2.55 percent for 2025–30.24

  • Long-run real return on foreign financial assets and government’s external debt. The long-run real return is set equal to 3 percent.25

  • Public capital stock, both for the historical part and for the projection horizon. Historical data on public capital are estimated, while public capital for the projection horizon can be generated using a perpetual inventory method, once the future fiscal policy has been identified. Official historical data on real public capital are not available, but official data on initial-period public capital stock in current prices are available. Starting from this public capital stock for the beginning of 2000, a historical path of public capital is constructed using the perpetual inventory method, given estimated government capital expenditures in constant 2000 prices and an assumed depreciation rate. Details of the derivation of the depreciation rate are discussed below. Public capital stock for the projection horizon can also be generated using the perpetual inventory method, once the future path of capital expenditures has been identified.

Calibration

3. The model is calibrated to replicate the growth experience of the Russian economy following the 1998 financial crisis. In particular, the following parameters and initial conditions need to be chosen so that the model replicates the recent growth pattern of the Russian economy:

  • the capital elasticity of the aggregate production function;

  • the time preference;

  • the rate of depreciation of private and public capital;

  • the initial level of private capital;

  • the holdings of foreign financial assets and the stock of government external debt for the end of the final year of the historical portion of the simulation (i.e., 2005);

  • a parameter that captures the contribution of public capital to aggregate production (coefficient of public capital);

  • the initial technology level; and

  • the intertemporal elasticity of substitution in households’ utility function that represents the degree to which consumption will be postponed in response to additional rewards.

4. The capital elasticity of the production function is set equal to the historical ratio of capital income (i.e., gross profits of economy and gross mixed income) to real GDP by revenue source net of net taxes.26,27 The time preference is set to 0.98, implying a subjective discount rate of 2 percent. Rates of capital depreciation are set to 7 percent per annum, so that the aggregate capital stock that is estimated using a perpetual inventory method roughly mimics the official data on the stock of fixed capital. In particular, starting from the official data on fixed capital for the beginning of 2000, a path for the capital stock can be generated using the perpetual inventory method and given the data on gross fixed capital formation for any depreciation rate. An exact depreciation rate that precisely replicates the official data on fixed capital can be calculated for each year. The average depreciation rate for 2000–04 is 7.654 percent. In the simulation, we use a slightly lower depreciation rate of 7 percent for both private and public capital stocks. The holding of foreign financial assets is set equal to the end-2005 stock of the oil stabilization fund, while the stock of government external debt is available from official sources.

5. The initial-period technology level is derived from a growth-accounting exercise, given an assumed coefficient of public capital. A simple growth-accounting exercise based on the Cobb-Douglas function implies a Solow residual, given private capital stock, employment, and non-oil output for 2000–05.28,29 The Solow residual in our model reflects two components: technological progress and public capital stock. Only one of these components can be identified from the implied Solow residual, given an exogenous assumption about the other component. We assume that the coefficient of public capital is 0.01. Implied technological progress is roughly 4 percent for 2000–05, close to previous estimates of TFP growth from other studies. The assumed coefficient of public capital, together with the Solow residual and historical data on public capital stock and employment for 2000, implies an initial-period technology level of 8.4767.

6. The intertemporal elasticity of substitution remains as a free parameter that is used to match the annualized growth rate of simulated non-oil GDP under the unchanged policy to the actual annualized growth rate for the period 2000–05.

Parameters and initial conditions

article image

Solution method

7. Given the exogenous variables, fiscal policy variables, and parameter values, a full equilibrium path can be calculated using a technique designed to solve a standard optimal growth model.30 The paths of world financial assets and government external debt are simulated for 2006 onward only, given the data on end-2005 stocks.31

References

  • Balassone, Fabrizio, 2006, “Measuring Fiscal Performance in Oil Producing Countries”, paper presented at the 8th Banca d’Italia Workshop on Public Finance — “Fiscal Indicators”, Rome.

    • Search Google Scholar
    • Export Citation

  • Barnett, Steven, and Rolando J. Ossowski, 2003, “Operational Aspects of Fiscal Policy in Oil-Producing Countries,” in Fiscal Policy Formulation and Implementation in Oil-Producing Countries, edited by Jeffrey M. Davis, Rolando J. Ossowski, and Annalisa Fedelino (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Barnett, Steven, and Alvaro Vivanco, 2003, “Statistical Properties of Oil Prices: Implications for Calculating Government Wealth”, in Fiscal Policy Formulation and Implementation in Oil-Producing Countries, edited by Jeffrey M. Davis, Rolando J. Ossowski, and Annalisa Fedelino (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Barro, Robert, J., and Xavier Sala-i-Martin, 1992, “Regional Finance in Models of Economic Growth,” Review of Economic Studies, Vol. 59 (4), pp. 645 –661.

    • Search Google Scholar
    • Export Citation

  • Barro, Robert, J., and Xavier Sala-i-Martin, 1995, Economic Growth (New York: McGraw-Hill, Inc.).

  • Blanchard, Olivier J., 1990, “Suggestions for a New Set of Fiscal Indicators”, OECD Working Paper No. 79 (Paris: OECD).

  • Blejer, Mario I., and Adrienne Cheasty, eds., 1993, How to Measure the Fiscal Deficit: Analytical and Methodological Issues (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Davis, Jeffrey and others, 2003, “Stabilization and Savings Funds for Nonrenewable Resources”, in Fiscal Policy Formulation and Implementation in Oil-Producing Countries, edited by Jeffrey M. Davis, Rolando J. Ossowski, and Annalisa Fedelino (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Kaldor, Nicholas, 1963, “Capital Accumulation and Economic Growth,” in Freidrich A. Lutz and Douglas C. Hague, eds., Proceedings of a Conference Held by the International Economics Association (London: Macmillan).

    • Search Google Scholar
    • Export Citation

  • Oomes, 2006, “The Utilization-Adjusted Output Gap: Is the Russian Economy Overheating?” IMF Working Paper No. 06/68 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Skancke, Martin (2003), “Fiscal Policy and Petroleum Fund Management in Norway,” in Fiscal Policy Formulation and Implementation in Oil-Producing Countries, edited by Jeffrey M. Davis, Rolando J. Ossowski, and Annalisa Fedelino (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Spilimbergo, 2005, “Measuring the Performance of Fiscal Policy in Russia,” IMF Working Paper No. 05/241 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Stokey, Nancy L., and Robert Lucas, 1989, Recursive Methods in Economic Dynamics (Cambridge, Massachusetts: Harvard University Press).

  • Wakeman-Linn, and others, 2004, Managing Oil Wealth: The Case of Azerbaijan, (Washington: International Monetary Fund).

1

Prepared by Fabrizio Balassone, Hajime Takizawa, and Harm Zebregs.

2

In this Chapter, the non-oil sector includes all sectors in the economy excluding oil and gas.

3

See Appendix I of Barnett and Ossowski (2003) for a simple algebraic exposition of the maximization problem.

4

The discounted value of oil wealth also depends on the assumed depletion path. This path is typically treated as exogenous and we follow this practice in this chapter. However, the depletion rate of oil resources is, in fact, a policy variable because the fiscal authorities can influence it both directly (if they are shareholders of oil companies) and indirectly (through the tax treatment of oil).

5

See, for instance, Barnett and Ossowski (2003).

6

See Balassone (2006) for an application to a sample of OECs.

7

See Skancke (2003) for a discussion.

8

A discussion of stabilization policy typically involves reference to a trend around which output fluctuates over a business cycle. However, trend output may be difficult to estimate in OECs. Many of these are emerging economies, often are embarking on major reforms that can change the structural characteristics and performance of the economy. Export concentration in the oil sector exposes OECs to sustained shocks in terms of trade.

9

It is also possible that certain expenditure categories fluctuate with swings in the cycle; for example, unemployment benefits.

10

Inter alia, they point out that futures price data imply about 60 percent of any given price shock is expected to be reversed within the following year.

11

The cutoff price was raised to $27 per barrel on January 1, 2006.

12

The Russian authorities are considering the introduction of a three-year medium-term budget framework.

13

A proposal to introduce the indicator in official budget documents was recently put forward by the Ministry of Finance.

14

The rate of extraction of oil and gas reserves is fixed exogenously.

15

Indeed, the method that is used to solve the model for an equilibrium path entails calculating the sum of a discounted utility stream for all possible private capital stock levels for the entire simulation period.

16

At the same time, the non-oil primary deficit would be about 5 percent of GDP higher than under unchanged policies, which assumes the expenditure-to-GDP ratio to be fixed at the 2005 level.

17

For computational simplicity, the natural resources sector is assumed to employ no domestic production factors. While this will not be entirely realistic, it is a good approximation for oil-rich countries, many of which rely on foreign capital and labor for exploration, development, and extraction activities, with the government collecting part of the rents.

18

The assumption of a single firm is made to simplify presentation, but it does not alter the results.

19

The government consumption adds nothing to the model since neither the utility nor the production of the nonnatural resources sector is affected by the government consumption.

20

This congestion effect in the use of public goods is also analyzed by Barro and Sala-i-Martin (1992 and 1995, pp. 158–59).

21

That the production function is consistent with balanced growth can be confirmed by multiplying At by (1 + γ), lt and Lt by (1 + n), and kp,t and Kg,t by (1 + γ)(1 + n) in the production function. It is straightforward to confirm that output grows at the rate (1 + γ)(1 + n)−1.

22

A typical model designed to study the consumption-savings decision of the government endowed with natural resources tends to assume that social welfare is a function of government spending. This assumption is often given an interpretation that government spending is akin to consumption. While this assumption can be easily accommodated in our model, we do not consider such an extension for simplicity.

23

See Stokey and Lucas (1989).

24

Using a standard Cobb-Douglas production function, Oomes (2006) estimates TFP growth for 1999–2004 is 4.1 percent.

25

Alternatively, the rate can be set equal to the steady state marginal product of private capital net of a depreciation rate implied by the calibrated model. The calibration that is discussed below implies a 9.5 percent marginal product of capital. While this rate is high compared with historical real returns on safe financial assets, such as U.S. treasury bonds, simulation results under this alternative assumption are likely to result in less bias in welfare rankings across different fiscal rules than the results discussed in the text.

26

The first-order condition of the firm’s maximization problem implies that capital (labor) elasticity is equal to the share of capital (labor) income in output.

27

Net taxes on production and imports are excluded since they accrue to neither capital nor labor.

28

The data on private capital stock are derived starting with the data on beginning-2000 private capital stock, estimated gross fixed capital formation by private sector in 2000 prices, and the 7 percent depreciation rate just discussed.

29

In the growth-accounting exercise and the simulation, factor inputs are adjusted for respective factor utilization rates. See Oomes (2006) for a detailed discussion of this approach.

30

In order to work with a detrended version of the model elaborated in Appendix II, all fiscal aggregates are first deflated by the GDP deflator and then detrended by the long-term rate of population-cum-technological progress. Working-age population is detrended by the long-term growth rate of working-age population, while the technology level is detrended by the long-term rate of technological progress.

31

To study fiscal policies that finance at least a part of the primary deficit on a balanced growth path by the return from the holding of financial assets, it is necessary to set the real return on foreign financial assets to be greater than the output growth rate implied by exogenous technological progress and labor force growth. Otherwise, financial assets cannot be maintained over time.

Appendix I. Reestimation of the Set of Equations in Shiells, Pani, and Jafarov (2005)

Shiells, Pani, and Jafarov (2005) (henceforth SPJ) specify an econometric model that includes standard growth determinants, as well as Russian economic growth, and allows for a possible shift in the regression coefficients following the Russian crisis.

Specifically, they estimate the following equation:

(A1)yi⁢t=α+γ0⁢dt+β1⁢yR⁢t+γ1⁢dt⁢yR⁢t+b2⁢xi⁢t+γ2⁢dt⁢xi⁢t+β3⁢yi,t−1+γ3⁢dt⁢yi,t−1+μi+vi⁢t,

where γit is real GDP growth for country i in year t, γRt is real GDP growth for Russia in year t, xit is a vector of exogenous determinants of growth in country i, dt is a dummy variable equal to 0 prior to the Russian crisis and 1 thereafter, and µt and vit are disturbance terms.61

Shiells, Pani, and Jafarov (2005) run different specifications of the above equation to check for the robustness of their results. First, they try different exogenous determinants of growth (xit). Second, they repeat their estimations assuming break points in both 1998 and 1999 because of some uncertainty regarding the precise timing of the shift in output correlations.62 Third, they calculate Arellano-Bond estimators, given that the inclusion of a lagged dependent variable gives rise to a bias in standard estimators of either the fixed- or random effects model. However, Shiells, Pani, and Jafarov do not use lags of explanatory variables. Their results therefore should be interpreted with caution since lagged explanatory variables may be important.

Tables A1a and A1b present the results for the equations assuming that the structural break point was 1998.63 The specification presented in the first column includes lagged own-country growth, country dummies, the CPI, government expenditure in percent of GDP, the EBRD transition index, EU growth, the real exchange rate, Russian growth, the trade openness ratio, and interactions between a post-Russian crisis dummy variable and all of the other explanatory variables. The coefficient on Russian growth is quite substantial (1.08) and significant.64 The coefficient on the interaction of the Russian crisis dummy with Russian growth is -0.96 and highly significant. These results imply that, on average, a 1 percentage point increase in Russian growth was associated with a similar size increase in another country’s growth rate, holding other factors constant, before the Russian crisis. After the crisis, this effect dropped to 0.12 percentage points and was not significantly different from zero.

Table A1a.

CIS and Baltic Countries: Coefficient Estimates in Real GDP Growth Regressions with Structural Break in 1998, 1993–2004 1/

article image
Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level. t -statistics are in parentheses beneath the coefficient estimates.

Regression (4) includes the CIS countries only (i.e., the Baltics are excluded).

Regression (5) includes the Baltic countries only (i.e., the CIS countries are excluded).

Table A1b.

CIS and Baltic Countries: Coefficient Estimates in Real GDP Growth Regressions with Structural Break in 1998, 1993–2004 1/

article image
Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Regression (4) includes the CIS countries only (i.e., the Baltics are excluded).

Regression (5) includes the Baltic countries only (i.e., the CIS countries are excluded).