HOW EFFECTIVE IS THE NIGERIAN OIL-PRICE-BASED FISCAL RULE?

Transcription

1 HOW EFFECTIVE IS THE NIGERIAN OIL-PRICE-BASED FISCAL RULE? Adesola Bamidele Ibironke PhD Researcher, Economics Division, Newcastle University Business School, United Kingdom. Abstract This paper investigates the level of the effectiveness of the Nigerian oil-price-based fiscal rule that was introduced in 2004 to limit the oil-driven volatility of the country. The analysis of the study is quantitative and it has two main parts. The first part examines the degree of volatility in Nigeria before and after the fiscal rule was introduced. The findings indicate that the fiscal rule has been highly effective in reducing the volatility. The other part of the analysis involves the comparison of the level of the effectiveness of the fiscal rule with that of the Norwegian 2001 stabilization reform, because the two policies have similar objectives. The result of the comparison confirms the effectiveness of the Nigerian fiscal rule and reveals that the stabilization reform has not significantly lowered volatility in the Norwegian economy.

2 1. Introduction Nigeria is an oil-producing country that began to depend greatly on the resource in the period of the oil boom of the 1970s (see for example, Odularu, 2007; and Akinlo, 2012). The dependence made the economy to become susceptible to the volatility transmission emanating from the international oil market (Baunsgaard, 2003). The volatility enters the economy primarily through the oil revenue of the country. The country is a member of the Organization of the Petroleum Exporting Countries (OPEC). Therefore, the oil revenue of the nation comes from its sales as an OPEC player in the world oil market. On the supply side, the main players in the oil market can be categorized into OPEC producers and non-opec producers. Like all the other OPEC players, the country s supply in the oil market is based on the production quota set for it by the organization. After the volatility is transferred into the oil revenue, the usual track through which it spreads into the rest of the economy is as follows: it gets to the government expenditure, when the revenue is spent; and thereafter the rest of the economy becomes volatile, when the expenditure affects the incomes and activities of the economic agents. World Bank (2003) compared the standard deviations of vital economic indicators of several countries and found that the Nigerian economy was among the most volatile in the world between 1961 and 2000.The nature of the country s volatility shows that the following two factors have significant contributions to it: (i) The price volatility attribute of oil. Generally, price volatility and non-renewability are the two main characteristics of oil. (ii) The tie between public expenditure and oil revenue in the country caused by the oil-driven nature of the economy (see figure 1 in the appendix). To deal with the volatility challenge, the oil-price-based fiscal rule was introduced in 2004 and later integrated into the law of the country in 2007 when the Fiscal Responsibility Act (FRA) was enacted. The aim of the rule is to restrain government expenditure through oil revenue smoothening, which involves setting a volatilityabsorbing reference oil price through which the revenues from the resource will be channelled into the budget. It is expected that this process will de-link government expenditure from oil revenue and limit the domestic volatility caused by the international transfer. Basically, the FRA is an Act that was put in place to promote the efficiency of fiscal policy procedure and implementation in the nation (Nigerian National Assembly, 2007). Therefore, it captures various aspects of the country s fiscal policy. For example, apart from the oil-price-based fiscal rule, the FRA requires the preparation of the Medium- Term Expenditure Framework (MTEF). A vital feature of the MTEF is that it is a multiyear budgeting approach that involves the allocation of public resources on a rolling basis over a medium-term period that consists of three financial years (Nigerian Federal Ministry of Finance, 2010). 2

3 The objective of this paper is to examine the level of the effectiveness of the country s oil-price-based fiscal rule through a quantitative analysis of the extent to which the rule has achieved its aim. Since the international volatility is transferred into the domestic economy primarily through the tie between the country s oil revenue and government expenditure, the first part of the analysis focuses on the behaviour of the two variables before and after the rule was introduced in This involves estimations relating to the two periods ( and ) on: (i) The standard deviation, correlation, and covariance, which measure the volatility, linear relationship, and co-movement of the variables respectively. (ii) The Autoregressive Conditional Heteroscedasticity (GARCH) model, which measures volatility clustering. (iii) The Chow test to test whether there is a structural change in the year (2004) the fiscal rule was introduced. (iv) The dummy-based structural break test to further test whether there is a structural change and to explore the factor(s) causing the change. (v) The error-correction model to examine the short-term and long-term relationships between the two variables. The other part of the analysis involves the use of GARCH-in-mean (GARCH-M) models on Nigeria and Norway to compare the levels of effectiveness of the oil-price-based fiscal rule of the former and the 2001 stabilization reform of the latter. The rationale for the comparison is that the Norwegian economy also had the predicament of oil-driven volatility (see for example Mehrara, 2012). The 2001 stabilization reform of the country was introduced to address the challenge. Furthermore, the framework of the fiscal policy of the Norwegian stabilization reform is widely accepted in the literature to be a sound one from which other oil- rich economies could draw lessons ( for example, see Schmidt- Hebbel, 2012). The remaining parts of the paper are organized as follows: a review of the factors that have caused the volatility of the Nigerian economy is presented in section 2; the nature of the oil-price-based fiscal rule of the country is discussed in section 3; section 4 presents data and empirical results; while concluding remarks are made in section A Review of the Factors that are Responsible for the Nigerian Volatility Challenge Below we discuss some of the main features of the Nigerian economy, as identified in the literature, that have contributed to the volatility problem of the country: (i) The Oil-Driven Nature of the Economy: How this this factor contributed to the volatility predicament has been discussed above. Table 1 in the appendix gives a picture of the dominant role of oil causing the oil-driven feature of the economy. Apart from the oil-driven nature, the literature also indicates other factors which made the country to be 3

4 vulnerable to the international transfer of the oil price volatility. Some of these factors are discussed below. (ii) The Small Size of the Economy: Unlike large open economies like the USA, the features of the Nigerian economy indicate that it is a small open economy. Salient three of such features are: (i) It is a price-taker in the world markets. (ii) Its domestic interest rate adjusts to the world interest rate. (iii) It has a small income. Based on the country s Gross National Income (GNI) per capita, the World Bank currently classifies it as a lower middle income economy. Using the National Income (GNP) per capita as a criterion, the Bank classifies the countries of the world annually into the following categories: low income, middle income (subdivided into lower middle and upper middle), and high income economies (World Bank, 2013). Basically, small economies like Nigeria usually have difficulties on managing external shocks. (iii) The High Degree of Openness: There are two sides of openness, namely Trade Openness (TO) and Financial Openness (FO). The TO can be measured by the ratio of foreign trade (exports plus imports) to the GDP. The level of Nigeria s TO is relatively high (see for example, Obadan, 2008). On the other hand, the level of FO may be estimated through the ratio of equity-based foreign liabilities to GDP (Calderon et al, 2005). The financial liberalization policy pursued by the country has also made the level of FO to be relatively high in the economy. (iv) The High Degree of Global Integration: The degree at which the economy integrates into the global economy is high. That is, it has a great level of interconnections with the other economies of the world in the globalization process. For example, based on an evaluation of 207 countries, the nation is currently ranked 67th in the overall globalization rankings of the world (KOF Swiss Economic Institute, 2013). These rankings are based on the globalization indices of the evaluated countries, which were calculated by using the economic, social, and political factors of global integration. (v) The Emerging Market Feature: The economy is one of the emerging markets of the world. The International Monetary Fund (IMF) classifies the country as an emerging market in the World Economic Outlook (WEO). The IMF divides the countries of the world into the following two major groups in the World Economic Outlook (WEO): (i) advanced economies; (ii) emerging market and developing economies (IMF, 2013). Emerging markets are usually very volatile economies due to certain factors, one of which is the effect of external price shock (Mody, 2004). 3. The Nature of the Nigerian Oil-Price-Based Fiscal Rule Generally, fiscal rules consist of numerical targets or procedures that governments must adhere to in dealing with the fiscal policies of their countries (see Kopits et al, 1998; and Eden et al, 2012). Numerical targets set certain values as limits on specified budgetary aggregates like expenditure and revenue. On the other hand, procedures are compulsory non-numerical steps that must be followed to achieve vital objectives like increased 4

5 transparency of the budgetary process. This means that a fiscal rule may be a numerical or a procedural rule. However, many countries run both together. The Nigerian oil-price-based fiscal rule demands that the annual fiscal expenditure is restrained through a reference oil price. The benchmark price is a result of some prudent analysis. It is a long-run average price that imitates a ten-year average oil price. Any surplus revenue got when actual oil price is above the benchmark price is kept in a special account called the Excess Crude Oil Account, and withdrawals are made from the account when the operating price is below the reference price. This suggests that the fiscal rule is primarily a numerical one. Furthermore, fiscal rules may also be classified based on the budgetary aggregates they are designed to constrain. This leads to the following categories: debt, budget balance, expenditure, and revenue fiscal rules. The Nigerian oil-price-based fiscal rule fits into the last two categories, because its implementation involves setting a constraint on government expenditure and creating an oil-revenue smoothening process that will consequently result in a restraint of the overall budgetary revenue. 4. Data and Empirical Results The study employs annual time series data on Nigeria spanning 1970 to 2010 on nominal government expenditure, nominal oil revenue, nominal non-oil revenue, and price deflator. 1 The price deflator was used to deflate the government expenditure, oil revenue and non-oil revenue to obtain their real values. Other data that were used in the study are Nigeria s quarterly time series data on the Gross Domestic Product (GDP) deflated by the CPI, Norway s quarterly time series data on real GDP, and the quarterly time series data on the Global Oil Price (GOIL). All the Nigerian data were sourced from the 2010 edition of the Central Bank of Nigeria Statistical Bulletin, except the CPI data which were obtained from the 2012 edition of the International Monetary Fund (IMF) International Financial Statistics. The Norwegian data and the GOIL data were sourced from the Global Vector Autoregression (GVAR) Toolbox 1.1. The results of the analysis are discussed as follows: 4.1. Unit Root Tests The Augmented Dickey-Fuller (ADF) and Philips-Perron (PP) unit root tests are carried out on all the variables of the analysis. One lag is used for the random walk regressions of Real Government Expenditure, ln(rge); Real Oil Revenue, ln(ror); and Real Non-Oil Revenue, ln(rnor), because they are annual series and any autocorrelation problems they have are expected to be corrected after one period. On the other hand, four lags are 1 Monthly or quarterly data on all the four series will produce higher numbers of observations that may result in a more robust analysis. But it was only on the price deflator on which quarterly data could be obtained at the time of this study. 5

6 used for the regressions of Nigerian GDP, ln(nigoutput); Norwegian GDP, ln(noroutput); and Global Oil Price, ln(goil), because they are quarterly series and their own autocorrelation problems are expected to be corrected after four periods. The results are presented in table 1 below: Table 1: Unit Root Tests Augmented Dickey-Fuller Philips-Perron Variable Level First Remarks Level First Remarks Without drift Difference Without Without drift Difference Without drift drift ln(rge) * 1(1) * 1(1) ln(ror) * 1(1) * 1(1) ln(rnor) * 1(1) * 1(1) ln(nigoutput) * 1(1) * 1(1) ln(noroutput) ** 1(1) * 1(1) ln(goil) * 1(1) * 1(1) Note: * indicates stationarity at 1%, 5%, and 10% levels of significance; while ** points to stationarity at 5% and 10% levels of significance. As shown by this table, all the series are Difference Stationary Processes at the 1%, 5% and 10% levels of significance, except ln(noroutput) which is stationary only at the 5% and 10% levels. The values of the calculated t of all the series are positive at level. This common feature indicates that they are all explosive at level. This explosiveness may be caused by the volatility of the considered economies Standard Deviation, Correlation and Covariance The Standard Deviation of the Real Government Expenditure and the Real Oil Revenue for the and the periods are estimated to determine the degree of the volatility of the variables before and after 2004, while the Correlation and Covariance between the variables for the same periods are estimated to determine the level of the linear association and co-movement between the two variables before and after The results are shown below: 6

7 Table 2: Standard Deviation, Correlation and Covariance Pre-Fiscal Rule Period ( ) Post-Fiscal Rule Period ( ) Standard Deviation Correlation between Dln(RGE) and Dln(ROR) Covariance of Dln(RGE) and Dln(ROR) Dln(RGE) Dln(ROR) Dln(RGE) Dln(ROR) Remark Reduction of volatility after the introduction of fiscal rule Reduced and negative association after the introduction of fiscal rule Reduced and negative comovement after the introduction of fiscal rule The post-2004 conditions of volatility, association and co-movement shown in this table suggest the effectiveness of the fiscal rule Structural Break The Chow Breakpoint test is employed to test whether there is a structural change in 2004 in a cointegration regression where Real Government Expenditure is the dependent variable and Real Oil revenue and Real Non-Oil Revenue are the regressors. Cointegration between the variables is determined by subjecting the error term of the regression into unit root analysis. 2 The unit root analysis shows that the error term is stationary at level. A dummy-based structural change test is also conducted using the cointegrating regression between Real Government Expenditure and Real Oil Revenue (see table 8 for evidence of cointegration between the two variables). The dummy, denoted by DU, takes value 0 between 1970 and 2003 and value 1 between 2004 and The results are presented in tables 3-6 below: 2 None of the other techniques of multivariate cointegration analysis is employed, because the main objective is to determine whether a structural change exits in a predetermined date in the three series. Furthermore, the literature shows that it is possible to extend the Engle-Granger cointegration approach to a multivariate case as done in this study (see for example, Gujarati et al, 2009). 7

8 Table 3: Cointegration Regression of ln(rge) on ln(ror) and ln(rnor) Dependent Variable: ln(rge) Method: Least Squares Sample: Include Observations: 41 Coefficient Std. Error t-statistic Prob. C ln(ror) ln(rnor) R-squared Adjusted R-squared F-statistic Prob(F-statistic) Durbin-Watson stat Table 4: Chow Breakpoint Test of Cointegration Regression Breakpoint Year: 2004 Null Hypothesis: No Break at Specified Breakpoint Varying Regressors: Intercept and ln(ror) Equation Sample: F-statistic Prob. F(3,35) Table 5: Unit Root Test on the Residuals from the Regression of Table 3 Null Hypothesis: U has a Unit Root Exogenous: None Lag Length: 0 (Automatic based on SIC, MAXLAG= 9) t-statistic Prob* Augmented Dickey-Fuller Test Statistic Test Critical Values: 1% Level % Level % Level

9 Table 6: Dummy-Based Structural Change Test Dependent Variable: ln(rge) Method: Least Squares Sample: Include Observations: 41 Coefficient Std. Error t-statistic Prob. C ln(ror) DU*ln(ROR) DU R-squared Adjusted R-squared F-statistic Prob(F-statistic) Durbin-Watson stat According to table 4, there is the evidence of a structural break at the 10% level of significance. Table 6 confirms the structural break and it also indicates that the movement along the slope and the intercept shift both contribute significantly to the structural change. These findings indicate that the fiscal rule caused a structural change in the considered series in Error Correction Mechanism The error correction analysis is done to examine the changes the introduction of the oilprice-based fiscal rule has caused in the short- and long-run relationships between Real Government Expenditure and Real Oil Revenue. The results are presented in tables 7-8 below: 9

10 Table 7: Error Correction Model for the Period Dependent Variable: Dln(RGE) Method: Least Squares Sample (adjusted): Include Observations: 33 after adjustments Variable Coefficient Std. Error t-statistic Prob. C Dln(ROR) ECM(-1) R-squared Adjusted R-squared F-statistic Prob(F-statistic) Durbin-Watson stat Table 8: Error Correction Model for the Period Dependent Variable: Dln(RGE) Method: Least Squares Sample (adjusted): Include Observations: 40 after adjustments Variable Coefficient Std. Error t-statistic Prob. C Dln(ROR) ECM(-1) R-squared Adjusted R-squared F-statistic Prob(F-statistic) Durbin-Watson stat According to tables 7-8, the explanatory variables and the error correction terms in the two sample periods are statistically significant. The coefficient of the error correction term changes from about in the period to about in the period. This indicates a reduction in the percentage of the discrepancy between shortterm and long-term Real Government Expenditure that is corrected within a year from about 76% to about 58%. The slower rate of adjustment to equilibrium suggests that the oil-price-based fiscal rule has been effective in limiting the dependence of Oil Revenue on Government Expenditure. 10

11 4.5. Volatility Break We investigate whether there is volatility clustering through a standard GARCH model in which the Real Government Expenditure is regressed on the Real Oil Revenue in the conditional mean equation. After the existence of volatility clustering (ARCH effect) is determined, we estimate another GARCH model in which a dummy variable is included as part of the regressors of the conditional variance equation to determine whether there is volatility break in the year (2004) in which the oil-price based fiscal rule was introduced. The dummy variable, represented by DU, takes value 0 for each of the years of the non-fiscal rule ( ) period and value 1 for each of the years of the fiscal rule ( ) period. Tables 9-10 present the results of the GARCH estimations. Table 9: GARCH Model Dependent Variable: Dln(RGE) Method: ML-ARCH (Marquardt) Normal distribution Sample (adjusted): Include Observations: 40 after adjustments Failure to improve Likelihood after 20 iterations Presample variance: backcast (parameter = 0.7) GARCH = C(3) + C(4)*RESID(-1)^2 + C(5)*GARCH(-1) Variable Coefficient Std. Error z-statistic Prob. C Dln(ROR) Variance Equation C RESID(-1)^ GARCH(-1) Adjusted R-squared Durbin-Watson stat

12 Table 10: GARCH Model with Dummy Variable Dependent Variable: Dln(RGE) Method: ML-ARCH (Marquardt) Normal distribution Sample (adjusted): Include Observations: 40 after adjustments Convergence achieved after 8 iterations Presample variance: backcast (parameter = 0.7) GARCH = C(3) + C(4)*RESID(-1)^2 + C(5)*GARCH(-1) + C(6)*DU Variable Coefficient Std. Error z-statistic Prob. C Dln(ROR) Variance Equation C RESID(-1)^ GARCH(-1) DU Adjusted R-squared Durbin-Watson stat As shown in table 9, RESID(-1)^2 is not statistically significant, suggesting it does not have statistically significant impact on the regressand. However, the GARCH(-1) term is positive and statistically significant, indicating that there is positive correlation between the conditional variance and its lagged term. The correlation indicates the existence of some degree of volatility clustering. These findings confirm that the tie between the Government Expenditure and the Oil Revenue makes the volatility of the latter to be transferred to the former, with correlation between the swings. However, as shown in table 10, the introduction of the oil-price-based fiscal rule in 2004 led to a break in the volatility, because the dummy variable that quantify volatility break is statistically significant. The coefficient of the dummy is negative as expected, because the measure of volatility (i.e. the regressand of the conditional variance equation) and the measure of volatility break are supposed to be inversely related Comparison of the Nigerian Oil-Price-Based Fiscal Rule with the Norwegian Stabilization Reform Using GARCH-M Models A further analysis of the level of effectiveness of the Nigerian oil-price-based fiscal rule is done by comparing it with the 2001 Norwegian stabilization reform through GARCH- M models. Two sets of GARCH-M models were estimated for each country, using data spanning 1979Q2-2009Q4. The conditional mean equation of each country model is based on the cointegration between output and global oil price in each economy. The cointegration is determined by subjecting the error terms of the output-oil-price regressions into unit root analysis. The analysis shows that the Nigerian error term is stationary at level, while the Norwegian error term is stationary after the first differences. The results of the unit root analysis are presented in tables

13 The GARCH-M approach is employed for the international comparison, because the risk factor of each country model, measured by the conditional variance, will capture the effect of volatility on growth in the economy. The existence of volatility in each economy is first determined through a set of estimations of pure GARCH-M models, after which another set of models that have dummy variables in their conditional variance equations are estimated. The dummy variable for Nigeria, denoted by NIGDU, takes value 0 from 1979Q2-2003Q4 and value 1 from 2004Q1-2009Q4. On the other hand, the dummy variable for Norway, denoted by NORDU, takes value 0 from 1979Q2-2000Q4 and value 1 from 2001Q1-2009Q4. The results of the GARCH-M estimations are presented in tables below: Table 11: Nigerian GARCH-M Model Dependent Variable: ln(nigoutput) Method: ML-ARCH (Marquardt) Normal distribution Sample : 1972Q2 2009Q4 Include Observations: 123 Convergence achieved after 8 iterations Presample variance: backcast (parameter = 0.7) GARCH = C(4) + C(5)*RESID(-1)^2 + C(6)*GARCH(-1) Variable Coefficient Std. Error z-statistic Prob. Log(GARCH) C ln(goil) Variance Equation C RESID(-1)^ GARCH(-1) Adjusted R-squared Durbin-Watson stat

14 Table 12: Nigerian GARCH-M Model with Dummy Dependent Variable: ln(nigoutput) Method: ML-ARCH (Marquardt) Normal distribution Sample: 1972Q2 2009Q4 Include Observations: 123 Convergence achieved after 17 iterations Presample variance: backcast (parameter = 0.7) GARCH = C(4) + C(5)*RESID(-1)^2 + C(6)*GARCH(-1) + C(7)*NIGDU Variable Coefficient Std. Error z-statistic Prob. Log(GARCH) C ln(goil) Variance Equation C RESID(-1)^ GARCH(-1) NIGDU Adjusted R-squared Durbin-Watson stat Table 13: Norwegian GARCH-M Model Dependent Variable: ln(norutput) Method: ML-ARCH (Marquardt) Normal distribution Sample: 1972Q2 2009Q4 Include Observations: 123 Convergence achieved after 7 iterations Presample variance: backcast (parameter = 0.7) GARCH = C(3) + C(4)*RESID(-1)^2 + C(5)*GARCH(-1) Variable Coefficient Std. Error z-statistic Prob. Log(GARCH) E ln(goil) E Variance Equation C E RESID(-1)^ E GARCH(-1) E Adjusted R-squared Durbin-Watson stat

15 Table 14: Norwegian GARCH-M Model with Dummy Dependent Variable: ln(noroutput) Method: ML-ARCH (Marquardt) Normal distribution Sample: 1972Q2 2009Q4 Include Observations: 123 Convergence achieved after 41 iterations Presample variance: backcast (parameter = 0.7) GARCH = C(3) + C(4)*RESID(-1)^2 + C(5)*GARCH(-1) + C(6)*NORDU Variable Coefficient Std. Error z-statistic Prob. Log(GARCH) ln(goil) Variance Equation C RESID(-1)^ GARCH(-1) NORDU Adjusted R-squared Durbin-Watson stat Tables 11 and 13 indicate the existence of strong volatility clustering in the two economies. The conditional variance terms of the conditional mean equations are negative and statistically significant in all the four regressions, reflecting that volatility, measured by the conditional variance, has statistically significant impacts on growth in the two economies. These findings agree with the empirical literature on oil-driven volatility in the economies. (Note: The estimations on the Norwegian economy are done without intercepts in the conditional mean equations, because error messages are shown by the software when the intercepts are included. This may be due to the statistical irrelevance of the intercepts of the conditional mean equations in the GARCH-M processes of the country). Table 12 again confirms that the oil-price-based fiscal rule led to volatility break in the Nigerian economy, because the Nigerian dummy, NIGDU is statistically significant. Furthermore, only RESID(-1)^2 is statistically significant after the dummy is included in the model of the country. The statistical insignificance of the GARCH(-1) term in the model with the dummy is further evidence that the fiscal rule of the country has limited the volatility. However, according to table 14, the Norwegian dummy, NORDU is not statistically significant, although volatility clustering still exists in the economy as shown by the statistically significant RESID(-1)^2 and GARCH(-1) terms of the table. This suggests that the Norwegian 2001 stabilization reform has not been effective in reducing volatility in the economy. This finding agrees with that of Boye (2011) who did a wide quantitative analysis on the Norwegian oil revenues and volatility. 15

16 Table 15: Unit Root Test on the Residuals from the Nigerian Output-Oil-Price Regression Null Hypothesis: ε has a Unit Root Exogenous: None Lag Length: 4 (Fixed) t-statistic Prob* Augmented Dickey-Fuller Test Statistic Test Critical Values: 1% Level % Level % Level Table 16: Unit Root Test on the Residuals from the Norwegian Output-Oil-Price Regression Null Hypothesis: Dε has a Unit Root Exogenous: None Lag Length: 4 (Fixed) t-statistic Prob* Augmented Dickey-Fuller Test Statistic Test Critical Values: 1% Level % Level % Level Conclusion This study quantitatively assesses the effectiveness of the Nigerian oil-price-based fiscal. The findings show that the fiscal rule has been very effective. The analysis also compares the fiscal rule with the Norwegian stabilization reform of The comparison confirms the effectiveness of the fiscal rule and reveals the ineffectiveness of the Norwegian reform in limiting domestic volatility. These findings suggest that although the framework of the Norwegian stabilization policy is widely celebrated in the literature, its contribution to the reduction of oil-driven domestic volatility is limited. Based on the effectiveness of the Nigerian fiscal rule, this study recommends the continued adoption of the rule in the Nigerian economy. 16

17 References Akinlo, A. E How Important is Oil in Nigeria s Economic Growth? Journal of Sustainable Development, Vol.5, No.4, April. Boye, F Oil Revenues and Macroeconomic Volatility in Norway. OPEC Energy Review. Baunsgaard, T Fiscal Policy in Nigeria: Any Role for Rules? IMF Working Papers, WP/03/155. Calderon, C, Loayza, N, and Schmidt-Hebbel, K Does Openness Imply Greater Exposure? World Bank Policy Research Working Paper No Central Bank of Nigeria Statistical Bulletin Eden, V. H, Emery, D, and Khemani, P Developing Legal Frameworks to Promote Fiscal Responsibility Design Matters, Gujarati, N. G, and Porter D. C Basic Econometrics. New York: McGraw-Hill Inc. International Edition. IMF World Economic Outlook. IMF International Financial Statistics. KOF Swiss Economic Institute Globalization Index. Zurich, Switzerland, Kopits, G. and Symansky, S Fiscal Rules, IMF Occasional Paper 162. Mody, A What is an Emerging Market? IMF Working Paper, WP/04/177. Mehrara, M, Karsalari, A.R, and Haghiri, F Oil Fund and the Instability of Macro-Economy in Oil-Rich Countries. World Applied Sciences Journal. 16(3): Nigerian Federal Ministry of Finance Medium-Term Expenditure Framework and Fiscal Strategy Paper: Federal Budget Office of the Nigerian Federation. Nigerian National Assembly Fiscal Responsibility Act, 2007: 2007 Act No th July. 17

18 Obadan M. I Economic Globalization, Markets and National Development: How Sensibly Do the Poor Countries (Nigeria Included) Stand. 98 th Inaugural Lecture Series, University of Benin, Nigeria. Odularu, G.O Crude Oil and the Nigerian Economic Performance. Department of Economics and Development Studies, College of Business and Social Sciences, Covenant University, Ogun State, Nigeria. Schmidt-Hebbel, K Fiscal Institutions in Resource-Rich Economies: Lessons from Chile and Norway. Catholic University of Chile. Smith, L.V. and A. Galesi (2011), GVAR Toolbox 1.1, gvartoolbox/index.html. World Bank Nigeria Policy Options for Growth and Stability. Report No NGA, Washington DC: The World Bank. World Bank Country Listing

19 Appendix Figure 1: Oil Price, Oil Revenue, Total Revenue, and Total Expenditure in Nigeria from 1970 to 2000 in Billions of US Dollars Source: Thomas Baunsgaard, Note: USS$/bbl means US Dollars per Barrel Table 1: Oil Output, Revenue, Export, and Oil-GDP in Nigeria ( ) Year Production (Million Barrels) Oil Revenue (Million) Oil/Total Revenue (%) Oil/GDP (%) Oil Export (Million Oil Export/Total Export (%) Naira) Nil nil Nil nil Source: Akinlo,