SUSTAINABLE ECONOMIC DEVELOPMENT IN NIGERIA. DOES WAGNER LAW MATTERS? Isiaq Olasunkanmi, OSENI (PhD) and Ibrahim Ayoade, ADEKUNLE Department of Economics, Faculty of Social and Management Science Olabisi Onabanjo University
[email protected] Abstract Fiscal intervention, through expansionary budgeting, has made a little impact considering the policy blockages that have restricted foreign investment inflows, made importation of machinery and intermediate products impossible, and exacerbated inflationary pressures. So far, the government has been needlessly fixated on narrow-minded policies such as import substitution and maintaining a currency policy that imposes risks on businesses while promoting cronyism. All these policies have only exacerbated the country’s economic crisis and prompted the debate on whether increased government spending is sufficient for sustainable economic growth in Nigeria. In evaluating this, the paper examines the validity or otherwise of Wagner’s theory in Nigeria from 1980 to 2016, using time series data on RGDP, TGEX, DINV and MS. The paper uses econometric techniques to validate or otherwise the Wagner’s law with the given variables. It is therefore expected that increased government spending in Nigeria should be sufficient to sustain growth. Hence, this would enable the policymakers to be proactive in their decisions towards achieving sustainable growth in Nigeria. Keywords: Government Expenditure, Economic Growth, Wagner law and Granger causality. JEL Codes: E62, O11 1. Introduction Public expenditure and economic growth have been the focus of public finance since the magnitude of public expenditure has been increasing over time in almost all the countries of the world. It is, therefore, necessary for governments to know the causal relationship between the two. Theoretically, there are two competing schools of thought defining this causal relationship. First, Wagner (1883) postulated that public expenditure is an endogenous variable and that there exist long-run tendencies for public expenditure to grow relatively to some national income aggregates such as the gross domestic product (GDP). Moreover, public expenditure is a consequence rather than the cause of national income. In other words, the causality between public expenditure and national income runs from national income to public expenditure. Therefore, Wagner’s law viewed that public expenditure plays no role in generating national income. However, Keynes (1936) argued that public expenditure is an exogenous variable and can be used to generate national income. For this reason, public expenditure is a cause rather than the effect of national income which is in contrast with Wagner’s law. He raised the idea that during 1|Page
economic depression government expenditure can be used to heighten economic activities. Therefore, the causal relationship should run from public expenditure to national income (Tang 2009). The term fiscal policy has conventionally been associated with the use of taxation and public expenditure to influence the level of economic activities. It has to do with two major activities; taxation on one side; then government expenditure on the other side. This study will only concentrate on the government expenditure side. In Nigeria, government expenditure has been on the rise owing to the huge receipts from production and sales of crude oil, and the increased demand for public goods like roads, power, education, communication, and health. Moreover, there is increasing need to provide both internal and external security for the people and the nation. Unfortunately, this rising government expenditure has not translated into meaningful growth and development, as Nigeria ranks among the least developed countries in the world (UNDP report, 2016). The result of a government role in economic activities and the achievements in economic performance have been mixed. For instance, the economy will experience growth in real output in some years and declines in others. Meanwhile, the economy is mostly dominated by the public sector except recently that the government is trying to adopt privatization policy. But the overall picture is low scoring for the country’s developmental efforts. The objectives of monetary and fiscal policies in Nigeria are wide-ranging, involving Gross Domestic Product growth rate, reduction in the rates of inflation and unemployment, improvement in the balance of payments, accumulation of financial savings and external reserves as well as stability in Naira exchange rate. The guiding principle as well as instruments applied to attain these objectives, however, have until recently been far from adequate. Perhaps, this could be attributed to inconsistency in the formulation and implementation of vibrant policies. Various empirical studies on the relationship between government expenditure and economic growth arrived at different and even conflicting results. Some studies suggest that increase in government expenditure on socio-economic and physical infrastructures impact on long-run growth rate. For instance, government expenditure on health and education raises the productivity of labour and increase the growth of national output. Equally, expenditure on infrastructure such as road, power etc. reduces production costs, increase private sector investment and profitability of firms, thus ensuring an increase in economic activities and economic growth (Barro, 1990; Okojie, 1995). On the other hand, observations that growth in government spending, mainly based on non-productive spending is accompanied by a reduction in income growth has given rise to the hypothesis that the greater the size of government intervention the more negative is its impact on economic growth (Glomm and Ravikumar, 1997; Abu and Abdullah, 2010). Government expenditure is considered an important variable which may determine changes in national income in developing countries like Nigeria. In other words, fiscal policy is a major economic stabilisation weapon that involves measure taken to regulate and control the volume, cost and availability as well as the direction of money in an economy to achieve some specified macroeconomic policy objective and to counteract undesirable trends in the Nigerian economy (Gbosi, 1998). To stimulate the economic growth by means of fiscal policy, the country must adopt more instruments. These according to Ebimobowei (2010) include; the financing of direct investments which the private sector would not provide an adequate quantities; the efficient supply of certain public services which are necessary to ensure the basic conditions to display the 2|Page
economic activity and long-term investments; and the financing of public activities so as to minimize the distortions to come up with the decisions to spend and invest proper in the private sector. These instruments can be gotten through the nature and level of government spending in the economy. Though it can also be achieved either by an increase or a decrease in taxes, government expenditures constitute the bedrock of fiscal policy but in reality, government policy requires a mixture of both fiscal and monetary policy instruments to stabilize an economy because none of these single instruments can cure all the problems in an economy (Ndiyo and Udah, 2003). Despite several fiscal measures introduced since 1986, and given the prominence of fiscal policy in macroeconomic management in Nigeria, growth has not accelerated as expected and as such poverty remains widespread and pervasive, particularly in the rural areas. One could ask, what is the role of fiscal policy in inducing sustainable economic growth in an economy, redistributing income and reducing poverty in Nigeria? Could fiscal policy be designed so as to ensure economic growth and reduce poverty while maintaining macroeconomic stability? Furthermore, does government spending in Nigeria contribute to economic growth and development? These are crucial questions to ask given the renewed interest of the current democratic structure in poverty alleviation and given that fiscal policy is the arrowhead of the policy package of the current policy framework in Nigeria. This study intends to focus specifically on one side (government expenditure) in achieving the following objectives; 1. To determine the nature and direction of causality between government spending and economic growth in Nigeria, by testing for the Wagner’s hypothesis and its reverse (Keynesian approach). 2. Determining the relationship between governments spending and economic growth with other control variables like money supply, domestic investment measured with the gross capital formation and population growth as a control variable. This will help to decide if the current pace of public spending in our economy is productive and should be encouraged or not. Having introduced the study, the subsequent sections are section two which contains the literature review, section three contains the methodology, section four gives the empirical results and discussion while section five is conclusion and policy recommendations. 2. Literature Review Many studies show that government expenditure is positively related to economic growth and poverty reduction but due to high expenditure most of the developing countries are facing the problem of fiscal deficit. Fiscal deficit leads to inflation in the economy. Mehmood and Sadiq (2010), argues that in many developing countries, high fiscal deficit crowding out the private investment in the long run and decreases the employment and output which adversely affects the poverty. As economic growth may increase through government spending. Jamshaid et al (2010) examined the relationship between economic growth and government expenditure, both at bivariate (aggregate) and multivariate (disaggregate) systems and concluded that economic growth causes government expenditure at bivariate level and also supported that increase in GDP causes growth in government expenditure (Wagner’s hypothesis). Jamshaid et al. (2010) examined the nature and the direction of causality in Pakistan between public expenditure and national income. Applying the Toda-Yamamoto causality test for annual data, they concluded that there was a unidirectional causality running from GDP to government expenditure, which supports Wagner’s Law. Interest for the Wagner hypothesis attracted the 3|Page
attention of many economists after the translation of the original work of Wagner by Cooke (1958), however, the interest had declined at the end of the 1970s. Although, the increased public spending in most countries, new development of econometric techniques, and the last translation of Wagner’s work by Biehl (1998) attracted again the interest of many policymakers and economists. Omoke (2009) investigated the direction of causality between Government expenditure (GE) and National Income (NI) in Nigeria using co-integration and Granger Causality tests for annual time series data. In this result, he discovered that there was no long-run relationship existed between government expenditure and national income in Nigeria between 1970 and 2005. Also, the Granger causality test revealed that causality ran from government expenditure to national income thus concluding that government expenditure plays a significant role in promoting economic growth in Nigeria. Similarly, Olugbenga and Owoye (2007) studied the relationships between government expenditure and economic growth for a group of 30 OECD countries, using annual data during the period 1970-2005. The variables of interest were total government expenditure (TGE) and gross domestic product (GDP) with the use of co-integration and Granger causality tests. The results showed the existence of a long-run relationship between government expenditure and economic growth. More so, the authors observed a unidirectional causality from government expenditure to growth for only 16 countries, hence supporting the Keynesian hypothesis. Nevertheless, causality runs from economic growth to government expenditure in 10 among the 30 countries, confirming Wagner’s law as quoted in Sevitenyi (2012), while a bi-causal relationship between government expenditure and economic growth, for four countries, was discovered. Ergun and Tuck (2006) in studying the direction of causality between national income and government expenditures for Indonesia, Malaysia, Philippines, Singapore, and Thailand using Granger causality test, discovered no Support for the hypothesis that causality runs from government expenditures to national income. This was found only in the case of Philippines. There was no evidence for this hypothesis and its reverse for the other countries. In Muhlis and Hakan (2003) work, an investigation of the long-run relationship between public expenditure and GDP for the Turkish economy was studied using time series annual data. They employed co-integration and Granger Causality tests and discovered that neither Wagner’s Law nor Keynes’ hypothesis was valid in Turkey. Singh and Sahni (1984) investigated the relationship between national income and government expenditures in India and discovered no causal relationship among the variables indicating the failure of both Wagner’s law and Keynes hypothesis. Theoretical Background The role and the size of a public sector in many developed countries increased especially after the World War II considerably. This increase led to the development of a large number of explanations for changes in the size of the public sector. Beside economic explanations, there are some studies considering fiscal, political, institutional and international dimensions of expansion as well, which are out of scope for this study. In the context of economic explanations, Wagner’s 4|Page
model, which is originally called as “the Law of Increasing State Activity” is one of the earliest attempts in this field and attracted great attention worldwide. In 1893, the German political economist Adolph Wagner put forward his well-known proposition that regards public expenditure growth as a natural consequence of economic growth. He did not express his ideas in the form of a law and avoided making definitive formulations. His views were later formulated as a law and came to be known as “Wagner’s law” or “Wagner’s hypothesis” Henrekson (1993); Halicioglu (2003). The Law suggests that an expansion of a country’s level of economic development leads to an increase in its relative size of the public sector. This statement includes a comparison of development between private and public sectors. According to Wagner’s law, as the national economy grows, the public sector will grow at a faster rate than the private sector. There are several underlying reasons causing this result. First, with economic growth, industrialization and urbanization would generate an increase in government expenditures. Development of the economies makes legal relationships between the economic agents more complex, which triggers the administrative, regulatory and protective functions of the government. Second, real income growth would lead to a higher level of demand for basic infrastructure. In such a case, there would be a need for increased provision of social and cultural goods and services. As a result, as the economy develops, expenditures on social welfare of society such as education and health expand. Third, the government has to interfere with the market to ensure the functioning of natural monopolies and to enhance economic efficiency (Bird, 1971). Following the explanations and debates on a theoretical level, Wagner’s law has been empirically tested by various researchers. The empirical evidence concerning the relationship between public income and expenditure is based on the assessment of the elasticity of expenditure to income. Only if such elasticity is superior to the unit and the coefficient sign is positive, it can be affirmed that the link between the two variables exists and it is consistent with Wagner’s hypothesis. As pointed out by Dutt and Ghosh (1997), Wagner did not present his Law in a mathematical form and he was not explicit in the formulation of his hypothesis. Therefore, over the years, different mathematical forms have been applied by the authors. There are at least six different versions of the Law. Several empirical specifications have been introduced to test the Wagner’s law in different versions. We will critically analyse six (6) different versions of Wagner’s law: Peacock and Wiseman (1961), Gupta (1967), Goffman (1968) , Pryor (1969), Musgrave (1969), Goffman and Mahar (1971) and Mann (1980) as indicated by Richter and Dimitrios (2012). These are listed below; 1.Peacock-Wiseman version 𝑙𝐺𝑡 = 𝛽0 + 𝛽1 𝑙𝑌𝑡 + 𝜀𝑡
𝛽1 > 1
……(1)
Where 𝑙𝐺 is the log of real government expenditures, 𝑙𝐺𝐶 is the log of real government consumption expenditure, 𝑙𝑃 is log of population, 𝑙 𝐺⁄𝑌 is the log of the share of government spending in total output, 𝑙 𝑌⁄𝑃 is the log of the per capita real output, 𝑙 𝐺⁄𝑃 is the log of the per capita real government expenditures , 𝑙𝑌 is the log of real GDP. 5|Page
2.
Peacock-Wiseman share version (Mann version) 𝐺⁄ = 𝛽 + 𝛽 𝑙𝑌 + 𝜀 0 1 𝑡 𝑡 𝑌
𝛽1 > 0
……(2)
𝛼1 > 0
……(3)
𝛿1 > 1
……(4)
𝛾1 > 1
…….(5)
𝜋1 > 1
……(6)
3. Musgrave version 𝐺⁄ = 𝛼 + 𝛼 𝑙𝑌 + 𝜀 0 1 𝑡 𝑡 𝑌𝑡 4.Gupta version 𝐺⁄ = 𝛿 + 𝛿 𝑙𝑌 + 𝜀 0 1 𝑡 𝑡 𝑃𝑡 5.Goffman version 𝑙𝐺𝑡 = 𝛾0 + 𝛾1𝑙𝑌𝑡 + 𝜀𝑡 6.Pryor version 𝑙𝐺𝐶𝑡 = 𝜋0 + 𝜋1 𝑙𝑌𝑡 + 𝜀𝑡
Derimbas (1999) stated that Public finance studies, following Wagner, have considered public expenditure as a behavioural variable, similar to private consumption expenditure. By contrast, macroeconometric models, essentially following Keynes, have treated public expenditure as an exogenous policy instrument designed to correct short-term cyclical fluctuations in aggregate expenditures (Demirbas 1999). 3.0 Methodology 3.1 Model Specification In validating the applicability of Wagners law or otherwise in Nigeria, this study is a prototype of Kesavarajah (2012) who specify a Solow growth model that emphasized the significance of investment (i.e. capital) and labour effectiveness in promoting growth. The Solow growth model is symbolically represented below: 𝑄 = 𝑓(𝐾, 𝐿)
(7)
Where 𝑄 is the national output, 𝐾 represents capital resources employed and 𝐿 for unit of labour employed in the production process. However, since our focus is on the public sector influence, the model includes Government expenditure as one of the factors that explain growth. The output (growth) model specified for the purpose of this study is presented thus: 𝑅𝐺𝐷𝑃𝑡 = 𝑓(𝐺𝑂𝑉𝐸𝑋𝑃 𝑡 , 𝐺𝐶𝐹𝑡 , 𝑃𝑂𝑃𝑡 , 𝑀2 𝑡 )
(8)
Where: 𝑅𝐺𝐷𝑃𝑡 = Real GDP at time t, 𝐺𝑂𝑉𝐸𝑋𝑃 𝑡 = Government expenditure at time t, 𝐺𝐶𝐹𝑡 = Rate of Investment (proxied by Gross Fixed Capital formation)𝑃𝑂𝑃𝑡 = population growth at time t, 𝑀2 𝑡 = Money supply at time t Restating the model in an econometric form: 𝑅𝐺𝐷𝑃𝑡 = 𝛽0 + 𝛽1 𝐺𝑂𝑉𝐸𝑋𝑃 𝑡 + 𝛽2 𝐺𝐶𝐹𝑡 + 𝛽3 𝑃𝑂𝑃𝑡 + 𝛽4 𝑀2 𝑡 + 𝜀𝑡 (9) Where 𝜀𝑡 represents error term and 𝛽0 , 𝛽1 , 𝛽2 , 𝛽3 , 𝛽4 𝑎𝑛𝑑 𝛽5 𝑎𝑟𝑒 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑒𝑠 6|Page
These variables are log-linearised to adjust for heteroskedasticity and variance in dimension in units and measurements 𝑙𝑛𝑅𝐺𝐷𝑃𝑡 = 𝛽0 + 𝛽1 𝑙𝑛𝐺𝑂𝑉𝐸𝑋𝑃 𝑡 + 𝛽2 𝑙𝑛𝐺𝐶𝐹𝑡 + 𝛽3 𝑙𝑛𝑃𝑂𝑃𝑡 + 𝛽4 𝑙𝑛𝑀2 𝑡 + 𝜀𝑡 (10) 3.2 Data Sources and Measurements Our study used time series data for sustainable economic growth and development (measured with real GDP) and indicators for Wagner's proposition in Nigeria (government expenditure, gross capital formation, money supply, and population growth) from 1980 through 2016. The data are mainly obtained from the CBN statistical bulletin various issues up until 2015 and World Bank Database (World Development Indicator, 2016). 3.3 Estimation Technique The study employs three-prong procedural steps. The first phase consists of pre-estimation evaluation, These are the preliminary evaluation of the data using the descriptive statistics method in order to help show, describe and summarize the data in a meaningful way and also to know if the data are normally distributed through their averages and Jarque-Bera values (Gujarati & Dawn, 2009). The next step is the determination of the stability of the variables. For the purpose of this research, the Phillips Perron and Augmented Dickey-fuller (ADF) unit root tests were deployed. This test of the time series data is required because a non-stationary regressor invalidates many standard empirical results. The presence of a stochastic trend is determined by testing the presence of unit roots in time series data. Thereafter, the Johansen co-integration test is applied to establish whether there is a long-run relationship between the variables. The primary step in the Johansen cointegration test is to obtain the optimal lag length because the Johansen cointegration test is sensitive to lag length. If the lag length is too much, the test will give a misleading result. The optimal lag length was determined by the Akaike Information Criterion (AIC) and Schwarz Information Criterion (SC). However, in a situation where any of the criteria (AIC or SC) picks an optimal lag length different from the other, the Schwarz Information Criterion is a better criterion to be used to determine the optimal lag length (Koehler &Murphree, 1988). The Error Correction Model (ECM), a test for a short run and long-run dynamics between variables is then conducted. The error correction model is specified as in equation (11): 𝛥𝑅𝐺𝐷𝑃𝑡 = 𝛼0 + ∑𝑛𝑖=1 𝛼1𝑖 𝛥 𝑅𝐺𝐷𝑃𝑡−1 + ∑𝑛𝑖=1 𝛼2𝑖 𝐺𝑂𝑉𝐸𝑋𝑃 𝑡−1 + ∑𝑛𝑖−1 𝛼3𝑖 𝛥𝐺𝐶𝐹𝑡−1 + ∑𝑛𝑖=1 𝛼4𝑖 𝛥𝑃𝑂𝑃𝑡−1 + ∑𝑛𝑖=1 𝛼5𝑖 𝛥 𝑀2 𝑡−1 + 𝛿𝐸𝐶𝑇𝑡−1 + 𝜀𝑡 (11) The third phase is the post estimation. In order to confirm the robustness and validity of regression model, some post-estimation tests were conducted. These are the Breusch-Godfrey Serial Correlation to test for autocorrelation, Breusch-Pagan Heteroscedasticity test to test for the violation of homoscedasticity and Granger Causality test to determine the nature of causal relationship that exist between the variables. Also, the Cusum structural stability test was conducted to examine the structural stability of the model.
7|Page
4.0
Empirical Result
4.1
Descriptive Statistic, Normality Test and Correlation Matrix
Table 1: Descriptive Statistics of the Data Set
𝒍𝒏𝑅𝐺𝐷𝑃
𝒍𝒏𝐺𝑂𝑉𝐸𝑋𝑃
𝒍𝒏𝐺𝐶𝐹
Mean
8.149449
5.565807
22.81514
18.58724
6.349659
Median
8.395581
6.370539
22.17989
18.58490
6.317335
Maximum
11.45259
8.032117
25.22128
19.04120
9.846986
Minimum
4.546746
2.265558
21.42738
18.13941
2.672158
Std. Dev.
2.317413
1.894469
1.283214
0.269842
2.471091
Skewness
-0.165288
-0.580534
0.771602
0.017800
-0.048259
Kurtosis
1.723156
1.816919
2.120550
1.820879
1.598291
Jarque-Bera Probability
2.609417 0.271252
4.121639 0.127350
4.732367 0.093838
2.087389 0.352151
2.961154 0.227506
Description
ln 𝑃𝑂𝑃
𝒍𝒏𝑀2
Source: Authors computation (E-views), 2017 Table 1 shows the mean and median of all the observations in the data set lie within the maximum and minimum values indicating the high tendency of the normal distribution. Gross capital formation and the population are positively skewed while money supply, government expenditure and real GDP are negatively skewed. The kurtosis statistics showed that 𝑅𝐺𝐷𝑃𝑡 𝐺𝑂𝑉𝐸𝑋𝑃 𝑡 , 𝐺𝐶𝐹𝑡 , 𝑃𝑂𝑃𝑡 , 𝑀2 𝑡 were platykurtic, suggesting that their distributions were flat relative to normal distribution. The Jarque-Bera statistics shows that the series are normally distributed since the p-values of all the series are not statistically significance at 5% level. Thus informing the acceptance of null hypothesis that says each variable is normally distributed. Table 2: Correlation Matrix of the Data Set REAL_GDP REAL_GDP GOV_EXP GCF POP M_2
GOV_EXP
GCF
POP
M_2
1 -0.523671 -0.274067 -0.075093
1 0.3076 -0.031851
1 0.280163
1
0.486981
0.489765
0.654883
0.493106
1
Source: Authors computation, 2017 Furthermore, studies have argued that testing of the correlation among the variables of estimates would make the researchers detect whether the variables have high multicollinearity among themselves. As a result, the parameter estimates may contradict what the theory says due to the unexpected effect of multicollinearity among the independent variables (Agung, 2009; Hamsal, 2006 as cited in Oseni, 2016). However, Iyoha (2004) argued that multicollinearity among 8|Page
variables occur when the result of the correlation coefficient is above 0.95. In line with this explanation, the study presents the results of the correlation analysis of the set of variables employed in Table 2 above. The table shows that the correlation coefficients among the variables 𝑅𝐺𝐷𝑃𝑡 𝐺𝑂𝑉𝐸𝑋𝑃 𝑡 , 𝐺𝐶𝐹𝑡 , 𝑃𝑂𝑃𝑡 , 𝑀2 𝑡 are below 0.95 indicating that there is no tendency for multicollinearity to occur among the independent variables. Time Series Properties of the Variables The ADF test is used to test for stationarity of the data. The ADF test consists of estimating the following regression. ∆𝑌𝑡 = 𝛼 + 𝛽t + 𝛿𝑌𝑡−1 + ∑𝑚 𝑖=1 𝜑𝑖∆𝑌𝑡−𝑖 + 𝜀𝑡
(12)
Where 𝛼 represents the drift, t represents deterministic trend and m is an optimal lag length ample enough to ensure that 𝜀𝑡 is a white noise error term. Table 4 Variables
Unit Root Test: Augmented Dickey-Fuller Test (ADF) Level T-Stat
Critical Value @ 5%
First Difference T-Stat
Critical Value @ 5%
Order of Integration
-0.811619 -2.948404 -5.049675 -2.951125 I(1) 𝒍𝒏𝑹𝑮𝑫𝑷 -0.089026 -3.544284 -3.548490 -6.226249 I(1) ln 𝑮𝑶𝑽𝑬𝑿𝑷 0.032608 -2.948404 -4.192599 -2.951125 I(1) ln 𝑮𝑪𝑭 1.345058 -2.971853 -6.523446 -2.971853 I(1) ln 𝑷𝑶𝑷 -1.377232 -2.951125 -4.215257 -3.595026 I(1) ln 𝑴𝟐 Source: Authors computation (E-views), 2017 The study used Augmented Dickey-Fuller to ascertain the order of integration of the variables. It is observed that all the variables were stationary at first difference I(1) at 5% significance level thus necessitating the conduct of Error Correlation Model to gradually adjust from the long run converging characteristics of the variables to the short run. Optimal Lag Length Selection In selecting the optimal lag length for the cointegration equation based on the hypothesis that the residuals are serially uncorrelated, the lag length which minimises the Akaike Information Criterion (AIC), Schwarz Criterion (SC) and the Hannan-Quinn Criterion and at which the model does not have autocorrelation is the optimal lag length. Table 3: Optimal Lag Length Selection Criteria Lag length LogL LR FPE AIC SC HQ -130.3963 NA 0.000123 8.023310 8.292668 0 64.80295 310.0223* -1.341350* 0.544154* 1.09e-08* 1 Source: Authors’ computation (E-views), 2017. Notes * indicates lag order selected by the criterion LR: Sequential modified LR test statistic (each test at 5% level) FPE: Final Prediction Error AIC: Akaike Information Criterion SC: Schwarz Information Criterion HQ: -Hannan- Quinn Information Criterion
8.115169 -0.698339*
9|Page
The result in Table 3 portrays different lag length criterion (LR, FPE, AIC, SC and HQ). The Schwarz information criteria depicting lag order length of (1) for the model is selected. After establishing the lag order length, the co-integration, and long-run equation results were estimated and explained in the next section. Co-Integration Test Johansen Co-Integration Test The result of the Johansen Co-integration for both the Trace Statistic and Maximum Eigen Value is reported in Table 4 and Appendix ii. With the hypothesized level of acceptance is 5 percent, Table 4: Result of Johansen Co-integration test based on Trace Statistic and Max Eigenvalue Trace Statistic Max. Eigen Value Eigen Trace 0,05 Prob. MaxCritical Prob.** No. of CE(s) value Statistic Critical Eigen Value Value Value 0.74 138.42 95.75 0.00 44.52 40.10 0.01 None * 0.69 93.91 69.82 0.00 38.61 33.88 0.01 At most 1 * 0.53 55.29 47.86 0.01 25.50 27.58 0.09 At most 2* 0.42 29.80 29.80 0.05 18.17 21.13 0.12 At most 3* 0.21 11.63 15.50 0.18 7.86 14.26 0.39 At most 4 0.11 3.76 3.84 0.05 3.76 3.84 0.05 At most 5 Source: Authors’ computation (E-views), 2017. Notes: Trace test indicates 4 cointegrating eqn(s) at the 0.05 level Max-eigenvalue test indicates 2 cointegration at the 0.05 level * denotes rejection of the hypothesis at the 0.05 level **MacKinnon-Haug-Michelis (1999) p-values The result simply means that there is a long-run relationship between real GDP, government expenditure, gross capital formation, population growth and money supply based on the rejection of the null hypothesis at 5% level of significance. The determination of the short run association is computed in the next section. 4.2.3 Error Correction Model Table 5: Short - run Estimation Variable
Coefficient
C D(ln 𝐺𝑂𝑉𝐸𝑋𝑃 ) D(ln 𝐺𝐶𝐹) D(ln 𝑃𝑂𝑃) D(ln𝑀2 ) ECM(-1) R-squared Adjusted R2 F-statistic Prob(F-statistic) Durbin-Watson
0.144023 0.119305 0.176324 -0.227719 0.191688 -0.459100 0.716278 0.753229 11.36061 0.000003* 1.977587
Std. Error 0.043916 0.077033 0.072101 0.145573 0.141091 0.129109
t-Statistic
Prob.
3.279479** 1.548755** 2.445524** -1.564291** 1.358615* -3.555899
0.0029* 0.0331** 0.0213** 0.0294** 0.0255** 0.0014*
10 | P a g e
stat Source: Authors’ computation (E-views), 2017. *(1%)**(5%) indicates significance levels Table 5 represents the result of short-run estimates by using Error Correction Model(ECM). The estimated coefficient of the error correction vector is 0.4591. This means ECM(-1) is the speed of adjustment correcting back at the rate of 45.91 percent annually. The negative sign and the significant probability signify the existence of co-integration among the variables. This shows that approximately 46% of the previous year's disequilibrium in the economy is corrected in the long run which implies that adjustment of the deviation of the explanatory variable back to normality is very high. The result of the short run in table 5 indicates that government expenditure, gross capital formation and money supply have a positive and significant relationship with RGDP in the short run, while population has a negative and significant relationship with RGDP in the short run. The value of the adjusted R2 of 0.75 indicates that 75.32% of variations in RGDP are explained by GOVEXP, GCF, MS and POP while the remaining 24.68 are captured outside the model. The value of Durbin Watson is 1.98 for the model. This implies that our model is free from problems of serial correlation. The F-statistics of 11.36061 is statistically significant at 1 percent level, indicating that the explanatory variables are jointly significant suggesting that the model has a very good fit. 𝑙𝑛𝑅𝐺𝐷𝑃𝑡 = 𝛽0 + 𝛽1 𝑙𝑛𝐺𝑂𝑉𝐸𝑋𝑃 𝑡 + 𝛽2 𝑙𝑛𝐺𝐶𝐹𝑡 + 𝛽3 𝑙𝑛𝑃𝑂𝑃𝑡 + 𝛽4 𝑙𝑛𝑀2 𝑡 + 𝜀𝑡 4.3 Granger Causality Test To model the direction of causality that exists between government expenditure and sustainable economic growth in Nigeria, the functional relationship is specified below; m
n
i 1
j 1
RGDPt i RGDPt i j GOV _ EXPt j 1t m
n
i 1
j 1
(13)
GOV _ EXPt i RGDPt i j GOV _ EXPt j 2t
(14)
Where GOV_EXP (government recurrent expenditure and government capital expenditure) is government expenditure and RGDP is real GDP as an indicator for sustainable development. 𝜀1𝑡 and 𝜀2𝑡 are the disturbances which are assumed to be orthogonal. In this framework, there are four possible hypotheses. Case 1: Unidirectional causality from RGDP to GOV_EXP. This is indicated if i 0 and j
=0
Case 2: Unidirectional causality from GOV_EXP to RGDP This is indicated if i =0 and j 0. j Case 3: Bilateral causality. This is indicated if i 0 and 0.
j Case 4: No causality. This is indicated if i =0 and =0.
11 | P a g e
Table 7: Granger Causality Test Result Null Hypothesis
F-Statistic
LNGOV_EXP does not Granger Cause LNRGDP
2.15914
LNRGDP does not Granger Cause LNGOV_EXP
0.00341
Prob.
Granger Causality Unidirectional causality 0.0039 RGDP→GOV_EXP 0.3433
Source: Authors’ computation (E-views), 2017 The result from the granger causality test was shown in Table 7. It reveals that unidirectional causality from real GDP to government expenditure in Nigeria. This assertion is in tandem with the proposition of Wagner (1883). Table 10: Serial Correlation Test Breusch-Godfrey Serial Correlation LM Test: F-statistic Obs*R-squared
0.845890 2.236642
Prob. F(3,25) Prob. Chi-Square(3)
0.4433 0.3268
Source: Authors’ computation (E-views), 2017 Given the probability value of 32.68 percent, we fail to reject the null hypothesis and conclude that our short-run model is free from problems of serial correlation. Table 4.3.1.2 Heteroscedasticity Test Result Breusch-Pagan-Godfrey Heteroscedasticity Test: F-statistic 0.565126 Prob. F(13,7) Obs*R-squared 10.75370 Prob. Chi-Square(13) Scaled explained SS 0.746089 Prob. Chi-Square(13) Source: Authors’ computation (E-views), 2017.
0.8225 0.6314 1.0000
Given the probability value of 63.14 percent, we fail to reject the null hypothesis and conclude that our short-run model is free from problems of heteroskedasticity.
12 | P a g e
15
10
5
0
-5
-10
-15 94
96
98
00
02 CUSUM
04
06
08
10
12
14
5% Signific anc e
Fig. 3: CUSUM Stability Test Source: Authors’ computation (E-views), 2017. The above figure shows that the CUSUM line is within the critical bounds of 5 percent level of significance which indicates that the model has structural stability. 5.0 Conclusion and Recommendation The paper test the efficacy of Wagner’s law for sustainable development in Nigeria using times series data on key macroeconomic indicators from 1981 to 2016. In evaluating its objectives, the paper adopts error correction modelling to account for the short-run dynamics of the model. The empirical result reveals that government expenditure, domestic investment and money supply induce sustainable economic growth in Nigeria. As such sensitive for the performance of major sectors of the economy. The findings of this study are in tandem with the findings of Sevitenyi (2012); Olugbenga and Owoye (2007); Omotoye (2007) and in stark contrast to Ergun and Turk (2006); Muhlis and Hakan (2003); Singh and Sahni (1984) who found negative association between government expenditure and growth. The findings of the study validate the applicability of Wagner’s law in Nigeria for the period observed. As long as government expenditure grows, economic growth and subsequently sustainable economic growth will be achieved. It is therefore recommended that short-run policies should be tailored towards augmenting total government spending to trigger a reversal growth in GDP which can then be sustained over time.
References Abu, N and Abdullahi, U. (2010), Government Expenditure and Economic Growth in Nigeria, 19702008: A Disaggregated Analysis, Business and Economics Journal, Volume 2010: BEJ-4: 1-11. Barro, R. (1990). Government Spending in a Simple Model of Endogenous Growth. Journal of Political Economy, 98, S103 - S125. Barro, R. and Sala-Martin, X. (1992). Public Finance in Models of Economic Growth. Review of Economic Studies, 59, 645 - 661
13 | P a g e
16
Barro, R. J. (1991). Economic Growth in a Cross-section of countries. Quarterly Journal of Economics, 106, 407 - 444. Biehl, D. (1998), "Wagner's Law: An Introduction to and a Translation of the Last Version of Adolph Wagner's Text of 1911", Public Finance, vol. 53, pp. 102-111. Cooke, N. 1958, Three Extracts on Public Finance by Adolph Wagner Translated from German, Macmillan & Co. Ltd, London. Dickey, D. and Fuller W. (1979). Distribution of the Estimators for Autoregressive Time Series with a Unit Root. Journal of American Statistical Association, 74, 427 - 431. Ebimobowei, A. (2010). The Relationship between Fiscal Policy and Economic Growth in Nigeria (1991 - 2005). International Journal of Economic Development Research and Investment, 1(2&3) Engle, R.F., (1982), Autoregressive conditional heteroskedasticity with estimates of the variance of U.K. inflation, Econometrica 50, 987-1008. Ergun, D and Tuck, C. (2006), Government Expenditure and National Income: Causality Tests for Five South East Asian Countries, International Business and Research Journal Vol.5, No, 10: 49-58. Garba, A. G. (1999). Regime Change, External Shock and Federal Fiscal Operations in Nigeria: 1975-93. National Economic Society Annual Conference. pp 353 - 377. Gbosi, A. N. (2008). Contemporary Macroeconomic Problems and Stabilization Policies. Port Harcourt: Automatic Ventures. Gbosi, A.N. (1998). Banks, Financial Crisis and the Nigerian Economy Today, Corporate Impression Publishers, Owerri. Glomm, J.J. and Ravikumar, D.J. (1997). “The Growth of Public Expenditure in Selected Developing Nations: Six Caribbean Countries”, Public Finance/Finances Publique, 3(3): 57 – 74 Gujarati, D. N. and D. C. Poter (2009). Basic Econometrics (5th ed.). New York: McGraw-Hill Jamshaid, R, Iqbal, A. and Siddiqi, M. (2010), Cointegration-Causality Analysis between Public Expenditures and Economic Growth in Pakistan, European Journal of Social Sciences, Vol. 13, No 4: 556-565. Keynes, J. M. (1936). The General Theory of Employment, Interest and Money. New York: Harcourt Brace Publishers. Mehmood, R. and S. Sadiq, (2010), The relationship between Government Expenditure and poverty: A Cointegration Analysis, Journal of Fiscal policy vol.1 (is.1) pp. 29-37 Muhlis, B and Hakan C. (2003), Causality between Public Expenditure and Economic Growth: The Turkish Case, Journal of Economic and Social Research 6 (1): 53-72. Ndiyo, N.A. and E.B. Udah, (2003). Dynamics of monetary policy and poverty in a small open economy: The Nigerian experience, Nigerian Journal of Economics and Development Matters, 2(4), 40-68. Okojie, C.E.E, (1995). "Human Capital Formation for Productivity Growth in Nigeria”, Nigerian Economic and Financial Review, June, pp. 44-5. Olugbenga and Owoye (2007), Public Expenditure and Economic Growth: New Evidence from OECD Countries [http://iaes.confex.com/iaes/Rome_67/techprogram/S1888.HTM] Omoke P. (2009), Government Expenditure and National Income: A Causality Test for Nigeria, European Journal of Economic and Political Studies, Vo. 2: 1-11. 14 | P a g e
Ramsey, J.B. (1969) "Tests for Specification Errors in Classical Linear Least Squares Regression Analysis", Journal of the Royal Statistical Society, Series B., 31(2), 350–371. Richter, C. and Dimitrios, P.(2012), The Validity of Wagner’s Law in the United Kingdom. International Network for Economic Research working paper. Roux, A, (1994). “Defence, Human Capital and Economic Development in South Africa”, African Defence Review, No 19. Sevitenyi L. N. (2012), Government Expenditure and Economic Growth in Nigeria: An Empirical Investigation (1961-2009), Journal of Economic Analysis,vol.3 (issue 1) pp 38-51 Singh and Sahni (1984), Causality between Public Expenditure and National Income, The Review of Economics and Statistics 66, 630-644. Tang, C. F. (2009). An Examination of the Government Spending and Economic Growth Nexus for Malaysia Using the Leveraged Bootstrap Simulation Approach, Global Economic Review, 38(2), 215-227. Wagner A. (1883), Three Extracts on Public Finance, translated and reprinted in R.A. Musgrave and A.T. Peacock (Eds), Classics in the Theory of Public Finance, London: Macmillan [1958].
15 | P a g e
