Government Expenditure and Economic Growth in Sub-Saharan Africa

The paper sought to investigate the effect government expenditure on economic growth in Sub-Saharan Africa using a panel data for 35 Sub-Saharan African countries for the period 2006-2018. The paper adopted dynamic panel data and estimates were achieved by using two-step system GMM while taking into account the problem of instrument proliferation. The paper provided evidence that education and health expenditure are key determinants of income growth for SSA. The impact of education spending on cross-country income variation is more effective in low income SSA countries than the middle income SSA countries. However, military expenditure on output growth is more effective in improving income level of middle income SSA countries than low income SSA countries. SSA countries should allocate more funding towards education sector and should also avail compulsory and free primary and secondary education. SSA should carry out health reforms which improve primary health and universal health insurance coverage.

sector thus leading to shrinking of the private sector borrowing. Inadequate funds for the private sector reduce private investment resulting in low aggregate demand. Reduced private investment negatively effect on overall economic performance. Secondly, Crowding out can occurs through government spending which crowds out private spending. State intervention in the provision of public goods like education can crowd out private spending in some of these sectors. Efficient provision of public services by the government can create low demand for similar goods in the private sector hence discouraging private sector investment. Lastly, indirect crowding out occurs when government expenditure is financed through increased taxation thereby reducing private savings. A decrease in private savings is associated with low investment undertakings and this could result in low aggregate demand subsequently leading to slow economic growth.
Fourth, Neo-classical advanced by Solow-Swan (1956) is premised on the idea that increasing physical capital results in diminishing returns thus capital has a transitional effect on economy income level. The theory therefore suggests that it is necessary to increase labour productivity to spur economic growth.
Accordingly, steady state economic growth can be achieved through accumulation of capital, labour, and advances in technology. The theory states that an equilibrium state can be achieved by varying the right quantities of capital and labour in the production function. Technological changes significantly augment output therefore output growth cannot be realized in the absence of advanced technology.
Neo-classical growth theory thus suggests an investments in modern technology to augment the existing labour force to enable steady economic growth rate.
Finally, endogenous growth theory postulates that steady economic growth is achieved through technological change that is endogenously determined (Frankel, 1962;Romer, 1990;Mankiw et al., 1992;Barro et al., 1992;Karras, 1996). Government investment in R&D and human capacity building is associated with increase in economic growth. Endogenous growth model predicts that an increase in the proportion of people working in the research and development and the knowledge sectors will increase economic growth of a country. Therefore countries can stimulate growth by investing in capital, education and R&D. This theory emphasizes that the key to economic growth is investment in education. Barro (1990) predicts that public spending has both temporary and permanent effect on income growth. Barro (1991) conducted a cross sectional study on economic growth for 98 countries. The study used OLS estimation technique to arrive at the estimates. Finding showed that government consumption expenditure negatively impacts economic growth. This was attributed to distortions emanating from tax rates that discourages investment hence inhibits growth. For 96 non-communist countries between 1960non-communist countries between and 1970non-communist countries between , Landau (1983 analysed spending and output growth. OLS inference approach was used by the author. Findings provided evidence that public spending by government negatively impacts income growth. Devarajan (1996) (1985) analysed statistical relationship of spending and income growth. The paper used OLS estimation technique. The paper predicted that government expenditure in general impedes income per capita products while expenditure on transfers has positive link with growth. Ram (1986)  The government is assumed to finance its spending by assuming a balanced budget and imposing a flat tax-rate ( ).
The proportion, Ø (0≤ Ø≤ 1), of tax revenue goes towards productive expenditure ( ). Therefore and is given as: The problem of a representative agent taking the government decision on and Ø is to optimize his satisfaction by choosing consumption, c, and capital, k.
Where ρ denotes time preference. A utility function with a constant elasticity of marginal utility is specified as: Hamiltonian function is set up as follows to get the growth rate of consumption Obtaining Hamiltonian first order conditions, the growth rate of consumption can be rewritten as; Assume the steady-state growth rate of consumption is represented by Ϙ such that equation (8) is From equation (9) the proportion of government expenditure devoted for productive expenditure ( ).
Productive expenditure can now be defined as that component of public expenditure whose increase in proportion will raise the steady-state growth rate of the economy. From equation (10), component is

Empirical Model
In estimating the effect of government expenditure on economic growth, the baseline model is specified as: Published by SCHOLINK INC.
include: savings, inflation and lagged GDP growth rate. Savings is expected to positively predict income level as suggested by Modigliani (1970). Inflation captures the effect of macroeconomic instability on growth and expected to negatively predict economic outcomes. There exists dynamic interaction between a country's current economic performance with that of the previous income level, i.e., the economic activities in the preceding year have a bearing on current economic activities.
Therefore lagged values of GDP growth rate was included in the model. The study therefore adopted dynamic panel model. The parameter estimates for infrastructure expenditure is expected to be positive.
Where is the disturbance and is composed of fixed effect and time-specific effects.
Equation (12) is further modified to include dummy variable for middle-income SSA countries. Two income categories of SSA was included; the middle income and lower income group. Therefore instead of estimating two different equations for each group, dummy variable was included in the empirical model.

For example
The variable D is the dummy variable. The coefficient of D measures the difference in the two intercept terms. Therefore equation (13) can further be modified to include a dummy variable such that; = 0 + 1 −1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 + + Where = ( 1 − 0 ) Equation (14) is modified to include the interactions involving the middle income and dummy variables. This is to test whether the effect of spending on output growth varies with income level of countries.
Equation (14) thus becomes; = 0 + 1 −1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 + + 10 * +   (1996) and Obialor (2017)  Inflation positively and significantly influences GDP growth in the short run (p<0.05). Therefore 1 percent increase in inflation is associated with 0.0001 percent decrease in income level. Thus the magnitude of the effect of inflation on output in SSA is marginal. Productive labour force significantly improves GDP growth. One percent increase in labour force productivity will enlarge output by 0.0512 percent in the short run. The coefficient of the dummy variable is statistically significant (p<0.1). The output growth of middle income economies is 2.26 percent greater than those of low income category in the short run. Model 2 provides the beta estimates for the interaction term between education expenditure and the dummy variable. The beta estimates of the interaction term is negative and significant (p<0.001). This illustrates that the effect of education expenditure on output growth is less in middle income SSA countries than in low income SSA countries. In particular, income level of middle income SSA countries will grow by less than 2.031 percent as compared to low income SSA countries for every 1 percent additional expenditure on education.

Government Expenditure and Growth in SSA Countries
The interaction term in model 3 tests if the effect of military expenditure on income growth differ by income level of SSA countries. The interaction term for military expenditure is positive and statistically significant at 1 percent level. This suggests that the impact of military expenditure on income growth for middle income SSA countries is more than for low income SSA countries. Output growth in middle income SSA economies will be 4.08 percent higher than low income SSA economies for 1 percent increase in military expenditure. Model four captures the interaction term on the return to health expenditure. The beta estimate of the interaction term on health expenditure is positive but non-significant.
www.scholink.org/ojs/index.php/jepf Journal of Economics and Public Finance Vol. 7, No. 4, 2021 Published by SCHOLINK INC.  Table 3 provides long-run estimates of the relation between spending and output growth. A percentage change in education expenditure is associated with 0.25 percent increase in income level in the long run.
The result demonstrates that education expenditure has a significant larger effect on output growth in the long run (0.25) than in the short run (0.199). Health expenditure significantly explains output variation in SSA in the long run. Output will grow by 0.042 percent in the long run for 1 percent rise in health spending. Health expenditure has a larger positive effect on output growth in the long run (0.042) than the short run effect (0.0340). However, in the long run, military expenditure negatively impacts on output growth albeit non-significant.

Conclusion
The study sought to analyse the effect of government expenditure on economic growth in Sub-Saharan Africa. The analysis reveals that both education and health expenditure significantly predicts an improvement in income level of SSA countries. Military spending does not predict income level of SSA countries both in the long run and in the short run. Education and health spending effectively predict income level in the long period than in the short period. The study provided evidence that education spending is more effective in improving output level of low income SSA countries than in middle income countries. Military expenditure significantly improves income level of middle income economies of SSA countries than low income SSA countries. In contrast, health expenditure does not significantly predict income variation for SSA countries. Productive labour force is associated with significant and positive effect on income level. Inflation meets the priori expectation of negative influence on income level of a country.
The study underscores the need for the governments to increase budgetary allocation for education and military expenditure. Funding in education will enhance provision of quality education infrastructures and better remuneration of teachers which will enhance literacy. Low income countries should also consider free and compulsory primary and secondary education. This will work towards improved literacy level and consequently impact on growth. Policy makers in the education sector need to introduce skill based courses, technical institutions and the government should invest locally to create Published by SCHOLINK INC. employment opportunities. Creation of employment opportunities locally will discourage brain drain.
Military expenditure for middle income countries should be expanded but with a caution. The government in middle income countries should consider external borrowing rather than domestic borrowing in funding military expenditure. Domestic borrowing has the crowding out effect which might retard growth. Low income countries should cut on budgetary allocation for military expenditure.
Some of the money spent on military should be diverted to productive expenditure like building roads and schools which have multiplier effects on the economy. SSA needs to address health reforms through increased funding towards social policies which involve improved primary health and universal health insurance coverage as wells as R&D to eliminate tropical diseases such as malaria.