Abstract
This study investigates the effect of disaggregated capital expenditure on economic growth in Ethiopia over the period from 1981 to 2021. An autoregressive distributed lag (ARDL) model, combined with a Granger causality test, is used for the econometric analysis. The empirical results reveal that capital expenditure on economic development (CAEE), recurrent expenditure (RCE), and the inflation rate (CPI) have significant positive impacts on economic growth in both the long run and short run. However, capital expenditure on administrative and general development (CEAG) has a significant negative effect on economic growth in the long run. The Granger causality test further indicates unidirectional causality from economic growth (GNI) to capital expenditure on economic development (CAEE). Based on these findings, the study recommends that the Ministry of Finance should increase its budget allocations for capital expenditures aimed at economic development in order to further stimulate economic growth. Additionally, the government should implement stricter followup and monitoring mechanisms to ensure the proper management of budget allocations, particularly regarding capital expenditures on social development and administrative and general development. These measures are crucial for sustaining longterm economic growth in Ethiopia.
Keywords
Capital Expenditure, GNI, CAEE, CES, CEAG, ARDL
1. Introduction
Economic growth is one of the most important macroeconomic variables, reflecting the overall performance of a society. It results from the production of more goods and services, which requires improvements in productivity and growth in the labor supply. Productive growth involves a combination of a more educated and efficient workforce, increased private physical capital (such as plants and equipment), greater use of modern technology, expanded public infrastructure (such as roads and utilities), efficient markets for setting prices, and the rule of law to enforce contracts. To ensure wellfunctioning markets and stimulate economic growth, the government must expend resources to enforce contracts, maintain national security, protect against criminal activity, and provide valuable public goods
[12]  Nurudeen, A. and Usman, A., 2010. Government expenditure and economic growth in Nigeria, 19702008: A disaggregated analysis. Business and economics journal, 4(1), pp. 111. 
[12]
.
A nation's economic goals typically include achieving high levels of output, full employment, price stability, fair income distribution, and strong international relations. A government uses different policy instruments to achieve the goals of a nation. Fiscal and monetary policies are the two major tools available to policymakers to change the level of economic variables
[4]  Bargicho, S. E. I. D. A. (2016). The effects of government expenditure and tax on economic growth in Ethiopia. 
[4]
. This study focuses on fiscal policies, specifically government expenditure, which is the expense incurred by a government for the maintenance of itself as well as the provision of goods and services to promote economic growth and improve the welfare of the people in society. Through the provision of social amenities, the government reaches out to its citizens, enabling them to make a living and, thus, enhancing the growth of the economy
[1]  Aigheyisi, O. S. (2013). The relative impacts of federal capital and recurrent expenditures on Nigeria’s economy (19802011). American Journal of Economics, 3(5), 210221. 
[1]
.
Economists have different views about the role of government in economic activities. Endogenous growth models sates that only productive government expenditures will positively affect the longrun growth rate
[5]  Barro, R. J., 1990. Government spending in a simple model of endogeneous growth. Journal of political economy, 98(5, Part 2), pp. S103S125. 
[5]
. Neoclassical economists believe that reducing the role of the private sector through a crowding out effect is important because it lowers inflation; an increase in public debt raises interest rates, which lowers inflation and output. The new Keynesians present the multiplier effect in response and argue that the increase in government expenditure will increase demand and thus increase economic growth.
1.1. Statement of the Problem
The goal of ensuring sustainable development and reducing mass poverty is central to the development strategies of most developing countries. Abandoning growth as a key objective would be a tragic mistake, as it would condemn a substantial proportion of the population in developing countries to a life of poverty, even if accompanied by full employment, stable prices, and equitable income distribution. Only growth can create, if not certainty, at least the option of a more comfortable life for the masses
[8]  Ketema, T. (2006). The impact of government spending on economic growth: The case of Ethiopia. Addis Ababa University School of Graduate Studies, (Unpusblished Master Thesis). 
[8]
.
The relationship between government expenditure and economic growth has continued to generate a series of controversies. Some researchers argue that the effect of government expenditure on economic growth is insignificant
[2]  Akpan, N. I., 2005. Government expenditure and economic growth in Nigeria: a disaggregated approach. 
[14]  Romer, P. M., 1990. Endogenous technological change. Journal of political Economy, 98(5, Part 2), pp. S71S102. 
[2, 14]
, while others indicate that the effect is positive and significant
[9]  Korman, J. and Bratimasrene, T. (2007), “The relationship between Government Expenditure and Economic Growth in Thailand”, Journal of Economic Education, Vol 14: pp 234246. 
[16]  Ghosh, S. and Gregoriou, A., 2006. On the composition of government spending, optimal fiscal policy, and endogenous growth: Theory and evidence. 
[9, 16]
. Others also concluded that government spending has no significant impact on economic growth
[17]  Schaltegger, C. A. and Torgler, B., 2006. Growth effects of public expenditure on the state and local level: evidence from a sample of rich governments. Applied Economics, 38(10), pp. 11811192. 
[18]  Hasnul, A. G., 2015. The Effects of Government Expenditure on Economic Growth: the Case of Malaysia. 
[17, 18]
.
Previous research in this study area in Ethiopia was very general, which means that the researchers did not examine disaggregated capital expenditures, but government expenditure in aggregate. This study is designed to address the impact of capital expenditure (disaggregated) on economic growth in Ethiopia by examining the composition of capital expenditure on economic development, social development, and administrative and general development within the study period from 1981 to 2021.
1.2. Objectives of the Study
The main objective of the study is to investigate the impacts of capital expenditure on economic growth in Ethiopia. More specifically, the study has the following objectives to:
1. Analyze the long run impact of disaggregated capital expenditure on the Ethiopian economic growth.
2. Investigate the short run effect of disaggregated capital expenditure on the Ethiopian economic growth.
3. Investigate whether there is causality between disaggregated capital expenditure and Ethiopian economic growth.
1.3. Research Questions
To launch a valid argument and corroborate necessary facts concerning the objectives of the study, the following research questions were proposed:
1. Which category of capital expenditure significantly affects economic growth in the long run?
2. Which category of capital expenditure significantly affects economic growth in the short run?
3. Is there a causal relationship between disaggregated capital spending and economic growth?
1.4. Significance of the Study
The benefits of this study are that it incorporates the most recent data and employs more advanced econometric techniques such as autoregressive distributed lag (ARDL), bounds testing approach to cointegration, and the Granger causality test to evaluate causation with data spanning from 1981 to 2021, so it can be used as a literature review by students and academic researchers for future related area studies. This study helps the Ministry of Finance to focus on budgets that contribute to economic enhancement and to revise policies regarding government spending. It also assists financiers in determining where to invest by providing information on the government's spending priorities.
1.5. Scope of the Study
The conceptual scope of this study is to analyze the effects of disaggregated capital expenditure on the Ethiopian economy. Economic growth was treated as a dependent variable, and capital expenditure on economic development, capital expenditure on social development, capital expenditure on administrative & general development, recurrent expenditure, and inflation rate are independent variables. The period covered in this study is from 1981 to 2021. The researcher used models such as the autoregressive distributive lag (ARDL) model and the bound cointegration test to investigate longrun & shortrun impact of explanatory variable on dependent variable.
1.6. Organization of the Study
The first chapter is an introduction, in which the background of the study, the statement of the problem, the objective of the study, the research questions, the significance of the study, and the scope of the study are explained. In the second chapter, related theoretical and empirical literatures are summarized. The third chapter contains data types and sources, the methodology used in the study, and the estimation techniques used, which are described in detail. In the fourth chapter, econometric analysis and diagnostic tests are conducted. In the last chapter, conclusions and policy implications are presented.
2. Method of Data Analysis
After collecting data from various sources, the researcher employed descriptive statistics and econometric analysis to present the findings. The statistical software Eviews 9 was used to analyze the data. Autoregressive distributed lag (ARDL), the bounds testing approach to cointegration, and the Granger causality test to evaluate the causality of the dependent and independent variables were applied. An autoregressive distributed lag (ARDL) testing approach was introduced to investigate the existence of cointegration relationships among variables
[13]  Pesaran, M. H., Shin, Y. and Smith, R. J., 2001. Bounds testing approaches to the analysis of level relationships. Journal of applied econometrics, 16(3), pp. 289326. 
[13]
. It has several advantages:
1. In small samples of data, the ARDL model is the more statistically significant approach to find the cointegration relationship
[13]  Pesaran, M. H., Shin, Y. and Smith, R. J., 2001. Bounds testing approaches to the analysis of level relationships. Journal of applied econometrics, 16(3), pp. 289326. 
[13]
.
2. It can be applied whether the regression is purely order zero I (0), purely order one I (1), or a mixture of both. The ARDL approach avoids the pretesting problems associated with standard cointegration, which requires that the variables be already classified into I (0) or (1), or mixture of both.
3. In the ARDL approach, it is possible that different variables have different optimal numbers of lags.
4. The bound testing approach also has the advantage of deciding the long run and short run parameters of the model in equations at the same time. Lastly, by applying the ARDL technique, it can obtain unbiased and efficient estimators of the model.
Model Specification
Economist Paul Douglas and mathematician Charles Cobb developed the CobbDouglas production function which is commonly used in both macroeconomic and microeconomic models because it has several convenient and realistic properties. The CobbDouglas production function with constant returns to scale was used as a starting point to build model.
$Y=f\left(A{K}^{\alpha},{L}^{\beta}\right),\alpha +\beta =1$(1)
where Y = total output, K = capital used in production, L = total number of workers involved in production, and A = a positive constant number used to show the change in output that is not caused by the main production factors, $\mathit{\alpha}\mathrm{and}\mathit{\beta}$ are the output elasticity’s of capital and labor respectively.
Capital expenditure is divided into three categories to account for the effects of capital expenditure on economic growth: capital expenditure on economic development, capital expenditure on social development, and capital expenditure on administrative & general development. In general, the variables included in the model are gross national income per capita, capital expenditure on economic development, capital expenditure on social development, capital expenditure on administration & general development, recurrent expenditure, and consumer price index. Therefore, the models to be estimated in the study are specified as follows:
$\mathit{GNI}=f\left(\mathit{Capital}\mathrm{}\mathit{expenditure}\mathrm{},\mathit{Recurrent}\mathrm{}\mathit{expenditure},\mathrm{}\mathit{Consumer}\mathrm{}\mathit{price}\mathrm{}\mathit{index}\right)$(2)
$\mathit{GNI}=f\left(\mathit{CAEE},\mathit{CES},\mathit{CEAG},\mathit{RCE},\mathit{CPI}\right)\mathrm{}$(3)
The above equation is converted into regression form as follows:
${\mathit{GNI}}_{t}=\mathrm{}{\beta}_{0}+{\beta}_{1}{\mathit{CAEE}}_{t}+{\beta}_{2}{\mathit{CES}}_{t}+{\beta}_{3}{\mathit{CEAG}}_{t}+{\beta}_{4}{\mathit{RCE}}_{t}+\mathrm{}{\beta}_{5}{\mathit{CPI}}_{t}+{\epsilon}_{t}$(4)
GNI stands for Gross income per capita
CAEE stands for Capital Expenditure on Economic Development
CES stands for Capital Expenditure on Social Development
CEAG stands for Capital Expenditure on Administrative & General Development
RCE stands for recurrent expenditure
CPI stands for Consumer price index
β_{1}, β_{2}, β_{3}, β_{4}, and β_{5} are the slopes of the regression line concerning each variable when other variables are kept constant and ${\epsilon}_{t}$ is the white noise residual.
The general ARDL model is written as follows: 
${\u2206\mathit{LNGNI}}_{t}=\mathrm{}{\alpha}_{0}$+$\mathrm{}\sum _{i=1}^{n}{{\alpha}_{1}\u2206\mathit{LNGNI}}_{t1}+\mathrm{}\sum _{i=1}^{n}{\alpha}_{2}\u2206\mathit{LN}{\mathit{CAEE}}_{t1}+\sum _{i=1}^{n}{\alpha}_{3}\u2206\mathit{LN}{\mathit{CES}}_{t1}+\sum _{i=1}^{n}{\alpha}_{4}\u2206\mathit{LN}{\mathit{CEAG}}_{t1}$
${+\sum _{i=1}^{n}{\alpha}_{5}\u2206\mathit{LN}{\mathit{RCE}}_{t1}+\sum _{i=1}^{n}{\alpha}_{6}\u2206{\mathit{CPI}}_{t1}+\mathrm{}{\beta}_{1}\mathit{LN}{\mathit{GNI}}_{t1}+\mathrm{}\beta}_{2}{\mathit{LNCAEE}}_{t1}$
+${\beta}_{3}$LN${\mathit{CES}}_{t1}+{\beta}_{4}\mathit{LN}{\mathit{CEAG}}_{t1}+{\beta}_{5}\mathit{LN}{\mathit{RCE}}_{t1}+{\beta}_{6}{\mathit{CPI}}_{t1}$+${\epsilon}_{t}$(5)
Where, ∆ denotes the first difference operator for short run, ${\epsilon}_{t}$ is the white noise residual, t is a linear trend or time. The lefthand side of the equation is the gross national income per capita and the first six coefficients (${\alpha}_{1}{\alpha}_{6}$) on the righthand side correspond to coefficients of variables on the short run. The remaining coefficients (${\beta}_{1}{\beta}_{6}$) corresponds to coefficients of variables in the long run.
All variables are in natural logarithm form except consumer price index (CPI). Log transformation can help with heteroscedasticity by compressing the scale on which they are measured, reducing a tenfold difference in two values to a twofold difference
[7]  Gujarati, D. and Porter, D. C., 2010. Functional forms of regression models. Essentials of econometrics, 6, pp. 132177. 
[7]
. It is worth noting that the model is multiplicative, with all parameter coefficients corresponding to constant elasticity.
The ECM is used to estimate both shortrun and longrun effects of one time series on other variables and to know at what percentage the disequilibrium in the short run will be corrected in the long run. The equation for ECM is as follows:
${\u2206\mathit{LNGNI}}_{t}={\alpha}_{0}$+$\sum _{i=1}^{n}{{\alpha}_{1}\u2206\mathit{LNGNI}}_{t1}+\sum _{i=1}^{n}{\alpha}_{2}\u2206\mathit{LN}{\mathit{CAEE}}_{t1}$
$$+$\sum _{i=1}^{n}{\alpha}_{3}\u2206\mathit{LN}{\mathit{CES}}_{t1}+\sum _{i=1}^{n}{\alpha}_{4}\u2206\mathit{LN}{\mathit{CEAG}}_{t1}+\sum _{i=1}^{n}{\alpha}_{5}\u2206\mathit{LN}{\mathit{RCE}}_{t1}$
$+\sum _{i=1}^{n}{\alpha}_{6}\u2206{\mathit{CPI}}_{t1}$+${\mathrm{}\mathit{\theta ECM}}_{t1}+{\epsilon}_{t\mathrm{}}\mathrm{}$(6)
Where; $\u2206$ is the first difference operator, ${\epsilon}_{\mathit{t}}$is the white noise residual, and $\theta $ is the coefficients of error correction model (ECM); which measure the speed of adjustment to the longrun equilibrium after any convulsion in the short run.
3. Empirical Results and Interpretation
3.1. Unit Root
A stationary series can be defined as one with a constant mean, constant variance, and constant auto covariance for each given lag, and the value of the covariance between the two time periods depends only on the distance or lag between the two time periods, not on the time at which the covariance is calculated
[7]  Gujarati, D. and Porter, D. C., 2010. Functional forms of regression models. Essentials of econometrics, 6, pp. 132177. 
[7]
. Otherwise, it is a nonstationary time series variable. An integrated of order d time series is a nonstationary time series that must be differed d times to become stationary. The order of integration refers to the number of unit roots in the series or the number of differencing operations it takes to make a variable stationary.
If the dependent variable is a function of nonstationary variables, the regression will produce spurious results (a nonsense regression)
[6]  Brooks, C., 2008. RATS Handbook to accompany introductory econometrics for finance. Cambridge Books. 
[6]
. Even though the trending variables are completely unrelated, it is likely that significant tratios were obtained. Thus, to avoid the problem of spurious regression, it is necessary to test for stationary time series variables before running any sort of regression analysis. There are several tests for stationarity, including the DickeyFuller (DF) and PhillipsPerron (PPT) tests, as well as the augmented DickeyFuller (ADF) used in this paper.
The hypothesis used to test the unit root was stated as follows:
H0 = There is a unit root (nonstationary)
H1 = No unit root exists (stationary)
If the calculated statistic (in absolute terms) is less than the MacKinnon (1991, 1996) critical values, the null hypothesis is accepted and there is a unit root in the series. In other words, it means the time series is not stationary. The opposite is true when the calculated statistic is greater than the MacKinnon critical value.
The unit root test performed under ADF tests shows that the variables are stationary at different levels of stationarity. For example, variable LNGNI, LNCAEE, LNCES LNRCE, and CPI are stationary at 1^{st} difference with 5 percent significance level, while LNCEAG is stationary at level with 5 percent significance level.
3.2. Lag Length Selection Criteria
Prior to the cointegration test, the maximum lag length k was chosen by the unrestricted VAR model to decide the optimal lag length that should be included in the model. The optimal lag length is found with the sequential modified Likelihood Ratio test statistics (LR), the Akaiki information criteria (AIC), Schwarz information criteria (SIC), Bayesian information criteria (BIC), the final prediction error (FPE), and the HannanQuinn information criteria (HQ). Based on the test result, Akaiki Information Criteria (AIC) at a 5% significance level was chosen, with the model having a maximum lag length of two.
3.3. CoIntegration Test
Time series variables that are not stationary may have some linear combination of them that is stationary. In such a case, the variables are said to be cointegrated. This implies that there is a longrun relationship among the nonstationary variables. If the tests for stationarity reveal that some of the variables are not stationary, there is a need to conduct a cointegration test. The presence of longrun correlations among the variables is investigated using the Pesaran approach
[13]  Pesaran, M. H., Shin, Y. and Smith, R. J., 2001. Bounds testing approaches to the analysis of level relationships. Journal of applied econometrics, 16(3), pp. 289326. 
[13]
. The Ftest was used to do bound test. The Ftest compares the existence or absence of cointegration among the variables to the hypothesis of no cointegration among the variables. The Waldtest (Fstatistic) is the foundation of the ARDL bound test. The lower bound critical values assumed the explanatory variables were integrated of order zero, or I (0), while the upper bound critical values assumed they were integrated of order one, or I (1)
[13]  Pesaran, M. H., Shin, Y. and Smith, R. J., 2001. Bounds testing approaches to the analysis of level relationships. Journal of applied econometrics, 16(3), pp. 289326. 
[13]
. Therefore, if the computed Fstatistic is greater than the upper critical bounds, the null hypothesis should be rejected, shows that there is a longrun relationship between the selected variables. If the calculated Fstatistic is below the lower critical bounds, the null hypothesis cannot be rejected, and there is no longrun relationship between them. On the other hand, if the computed Fstatistic falls inside the critical value bounds, the test is inconclusive unless it is unknown the order of integration of the underlying variables. Subsequently performing the bound test, if the existence of a longrun relationship among the variables in the model is confirmed, then, with the error terms obtained from the longrun equation, the shortrun equation among the variables is estimated through the ARDL method developed by
[13]  Pesaran, M. H., Shin, Y. and Smith, R. J., 2001. Bounds testing approaches to the analysis of level relationships. Journal of applied econometrics, 16(3), pp. 289326. 
[13]
. The hypothesis of the bound test is stated as follows.
HO: No long run relationship
H1: Ho is not true
Decision rule: Reject HO, if Fstatistic > I (1)
According to the bounds test result, the calculated Fstatistic which is 10.242955 above the upper critical bound values (higher than at 90%, 95%, 97.5% and 99% upper bounds), which means that; the model rejects the null hypothesis of “No Long run relationship” and this implies that there is cointegration between the variables.
3.4. Longrun Estimation Coefficients
The long run estimation output reveals that capital expenditure on economic development (CAEE) has a positive significant effect on economic growth (GNI) at the 5 percent significance level. This complies with the prior expectation; a positive relationship was expected between the two variables. It proves that increasing capital expenditure for economic development by one percent; increases economic growth by 47 percent. Capital expenditure on administrative & general development (CEAG) has a negative and significant contribution to economic growth as excepted. It implies that when capital expenditure on administrative & general development increases by one percent; economic growth decreases by 15 percent, at ceteris paribus.
Recurrent expenditure (RCE) also has a positive & statistically substantial impact on economic growth; thus, when recurrent expenditure increases by 1 percent, economic growth increases by 27.7 percent. This is in line with Kryeziu et al. result
[11]  Modebe, N. J., Okafor, R. G., Onwumere, J. U. J. and Ibe, I. G., 2012. Impact of recurrent and capital expenditure on Nigeria’s economic growth. European Journal of Business and Management, 4(19), pp. 6674. 
[11]
. They examined the impact of government expenditure on Nigeria’s and Ghana’s economic growth respectively, and they both confirmed that recurrent expenditure had a positive impact on their respective country’s economic growth. The result also shows that inflation rate has a positive and significant contribution to economic growth as expected and which is supported by previous researchers
[10]  Kryeziu, N. and Durguti, E. A., 2019. The impact of inflation on economic growth: The case of Eurozone. International Journal of Finance & Banking Studies (21474486), 8(1), pp. 0109. 
[15]  Umaru, A. and Zubairu, A. A., 2012. Effect of inflation on the growth and development of the Nigerian economy (An empirical analysis). International Journal of Business and Social Science, 3(10). 
[10, 15]
.
Adjusted Rsquared is 0.998593, meaning that a total of 99% of economic growth is explained by the explanatory variables, while the remaining 1% is explained by variables outside the model. Fstatistics test the overall significance of the model under study. Fcalculated is compared with Ftabulated; when FCal is greater than Ftab, reject the null hypothesis (H_{0}) and conclude that the variable is statistically significant in explaining the dependent variable. According to the table, Fstatistics is 3954.858 and Prob (Fstatistic) is 0.000000. In this case reject the null hypothesis and accept the alternative hypothesis. This is because it is greater than the critical value or the tabulated value. Thus, it implies that the model is statistically significant. In other words, the explanatory variables considered together play a significant role in explaining variation in the dependent variable.
Durbin–Watson statistic is used to test the presence of autocorrelation (a relationship between values separated from each other by a given time lag), which shows whether there is serial correlation in the model. If there is serial correlation in the model, it therefore implies that the model has lost its predictive power. The value of DW always lies between 0 and 4. If d = 2 indicates no autocorrelation, d < 2 indicates positive serial correlation, d > 2 indicates negative serial correlation in general, if d is less than 1.5 or greater than 2.5 then there is potentially a serious autocorrelation problem. Otherwise, if d is between 1.5 and 2.5 then autocorrelation is likely not a cause for concern. The DW statistic is given as 2.083371 and shows that the model is free from autocorrelation.
Table 1. Longrun estimation coefficients.
Variable  Coefficient  Std. Error  tStatistic  Prob. 
LNCAEE  0.4692  0.0882  5.3189  0.0000** 
LNCES  0.0497  0.0893  0.5563  0.5821 
LNCEAG  0.1512  0.0710  2.1282  0.0416* 
LNRCE  0.2775  0.1211  2.2915  0.0291* 
CPI  0.6884  0.1155  5.9560  0.0000** 
Constant  7.6110  1.6237  4.6872  0.0001** 
Rsquared  0.9988  Mean dependent var  7.5543 
Adjusted Rsquared  0.9985  S.D. dependent var  1.4093 
S.E. of regression  0.0528  Akaike info criterion  2.8651 
Sum squared resid  0.0894  Schwarz criterion  2.5273 
Log likelihood  65.303  HannanQuinn criter.  2.7430 
Fstatistic  3954.8  DurbinWatson stat  2.0833 
Prob(Fstatistic)  0.0000   
Source: Author’s Computation using EViews 9 software
3.5. Shortrun Estimation Coefficients
The goal of running the error correction model is to estimate both shortterm and longterm effects of one time series variable on other variables and to know at what percentage the disequilibrium in the short run will be corrected in the long run. The coefficient of the short run error correction model, denoted by CointEq (1) in the short run estimation result, is 0.467260, which is negative, less than one, and significant at the selected level of significance, as expected. If ECM is negative and statistically significant, the variables are cointegrated in the long run
[3]  BahmaniOskooee, M. and Brooks, T. J., 1999. Bilateral Jcurve between US and her trading partners. Weltwirtschaftliches Archiv, (H. 1), pp. 156165. 
[3]
. This indicates that 46.7 percent of the previous year’s disequilibrium is adjusted in the current year (
Table 2). Shortrun disequilibrium and inconsistencies are adjusted and corrected in the long run at a rate of 46.7 percent.
The shotrun estimation result reveals that capital expenditure on economic development (CAEE) has a positive and notable influence on economic growth at a 5% significance level. Increasing one percent of capital expenditure on economic development increases economic growth at 21.9 percent in short run. Increasing one percent in recurrent expenditure increases economic growth at 13 percent. It also reveals that one unit increment in annual inflation rate increases economic growth at 32 percent. The other two variables (capital expenditure on social development, capital expenditure on administrative and general development) have a negative but nonsignificant effect on economic growth.
Table 2. Shortrun estimation coefficients.
Cointegrating Form 
Variable  Coefficient  Std. Error  tStatistic  Prob. 
D(LNCAEE)  0.2192  0.0566  3.8692  0.0005 
D(LNCES)  0.0232  0.0436  0.5320  0.5986 
D(LNCEAG)  0.0236  0.0200  1.1797  0.2474 
D(LNCEAG(1))  0.0306  0.0194  1.5717  0.1265 
D(LNRCE)  0.1296  0.0622  2.0847  0.0457 
D(CPI)  0.3216  0.0674  4.7715  0.0000 
CointEq(1)  0.4672  0.0767  6.0878  0.0000 
Cointeq = LNGNI  (0.4693*LNCAEE  0.0497*LNCES  0.1512*LNCEAG + 
0.2775*LNRCE + 0.6885*CPI  7.6110) 
Source: Author’s Computation using EViews 9 software
4. Conclusion and Policy Implication
4.1. Conclusion
This study investigates the impact of disaggregated capital expenditure on Ethiopian economic growth, using the auto regressive distributed lag (ARDL) model for data covered over the period 19812021. Thus, the longrun and shortrun estimation coefficients as well as the associated perand postdiagnostic tests are conducted. Specifically, the longrun empirical estimation results show that capital expenditure on economic development has a significant positive effect on economic growth. It proves that increasing by 1 percent capital expenditure on economic development increases economic growth by 47 percent. However, capital expenditure on administration & general development has a negative and significant contribution to economic growth. Regarding, the short run estimation, capital expenditure on economic development has a positive and substantial impact on economic growth. The result also reveals that recurrent expenditure and inflation rate has a positive and notable influence on economic growth. Recurrent expenditure also has a positive statistically notable influence on economic growth. In addition, capital expenditure on social development has a negative and insignificant effect on economic growth. Lastly, inflation rate has a positive and statistically notable influence on economic growth. To achieve economic growth; a nation has to create more jobs by investing new technologies or expanding their operations. They hire more workers to operate those technologies, which leads to more people having money to spend & brings inflation. Expansionary fiscal and monetary policies that government take to bring a healthy economic level, which is needed during the contractionary phase of the business cycle leads to inflation in shortrun.
Last but not least, the pairwise Granger causality test was taken, and it shows that economic growth causes capital expenditure in economic development. Capital expenditure on social development causes capital expenditure on economic development and capital expenditure on administrative & general development. Capital expenditure on economic development granger causes capital expenditure on administrative & general development.
4.2. Policy Implication
The findings of the study reveal that capital expenditure on economic development, recurrent expenditure and inflation rate has a positive and significant contribution to economic growth in both long run and short run. Capital expenditure on administrative & general development has a negative and notable influence in long run. Thus, on these bases, the following recommendations are forwarded to concerned bodies:
The minister of finance is recommended to increase the budget for capital expenditure, especially capital expenditure on economic development, which boosts economic growth. The Ethiopian government spends billions of birr on capital expenditure for social development and capital expenditure on administration and general development. However, it did not make a satisfactory contribution to economic growth. Therefore, it needs continuous monitoring to ensure that the allocated budget is used for its intended purpose.
Abbreviations
ADF  Augmented DickeyFuller 
ARDL  Auto Regressive Distributed Lag 
CAEE  Capital Expenditure on Economic Development 
CES  Capital Expenditure on Social Development 
CEAG  Capital Expenditure on Administrative & General Development 
COVID19  Corona Vires Disease Which Spread Since End of 2019 
ECM  Error Correction Model 
EPRDF  Ethiopian People's Revolutionary Democratic Front 
FDI  Foreign Direct Investment 
FDRE  Federal Democratic Republic of Ethiopia 
GDP  Gross Domestic Product 
GNI  Gross National Income 
LR  Long Run 
MOFEC  Ministry of Finance and Economic Cooperation 
NBE  National Bank of Ethiopia 
RGDP  Real Gross Domestic Product 
SR  Short Run 
VECM  Vector Error Correction Model 
Author Contributions
Wondimu Mekonnen is the sole author. The author read and approved the final manuscript.
Conflicts of Interest
The author declares no conflicts of Interest.
References
[1] 
Aigheyisi, O. S. (2013). The relative impacts of federal capital and recurrent expenditures on Nigeria’s economy (19802011). American Journal of Economics, 3(5), 210221.

[2] 
Akpan, N. I., 2005. Government expenditure and economic growth in Nigeria: a disaggregated approach.

[3] 
BahmaniOskooee, M. and Brooks, T. J., 1999. Bilateral Jcurve between US and her trading partners. Weltwirtschaftliches Archiv, (H. 1), pp. 156165.

[4] 
Bargicho, S. E. I. D. A. (2016). The effects of government expenditure and tax on economic growth in Ethiopia.

[5] 
Barro, R. J., 1990. Government spending in a simple model of endogeneous growth. Journal of political economy, 98(5, Part 2), pp. S103S125.

[6] 
Brooks, C., 2008. RATS Handbook to accompany introductory econometrics for finance. Cambridge Books.

[7] 
Gujarati, D. and Porter, D. C., 2010. Functional forms of regression models. Essentials of econometrics, 6, pp. 132177.

[8] 
Ketema, T. (2006). The impact of government spending on economic growth: The case of Ethiopia. Addis Ababa University School of Graduate Studies, (Unpusblished Master Thesis).

[9] 
Korman, J. and Bratimasrene, T. (2007), “The relationship between Government Expenditure and Economic Growth in Thailand”, Journal of Economic Education, Vol 14: pp 234246.

[10] 
Kryeziu, N. and Durguti, E. A., 2019. The impact of inflation on economic growth: The case of Eurozone. International Journal of Finance & Banking Studies (21474486), 8(1), pp. 0109.

[11] 
Modebe, N. J., Okafor, R. G., Onwumere, J. U. J. and Ibe, I. G., 2012. Impact of recurrent and capital expenditure on Nigeria’s economic growth. European Journal of Business and Management, 4(19), pp. 6674.

[12] 
Nurudeen, A. and Usman, A., 2010. Government expenditure and economic growth in Nigeria, 19702008: A disaggregated analysis. Business and economics journal, 4(1), pp. 111.

[13] 
Pesaran, M. H., Shin, Y. and Smith, R. J., 2001. Bounds testing approaches to the analysis of level relationships. Journal of applied econometrics, 16(3), pp. 289326.

[14] 
Romer, P. M., 1990. Endogenous technological change. Journal of political Economy, 98(5, Part 2), pp. S71S102.

[15] 
Umaru, A. and Zubairu, A. A., 2012. Effect of inflation on the growth and development of the Nigerian economy (An empirical analysis). International Journal of Business and Social Science, 3(10).

[16] 
Ghosh, S. and Gregoriou, A., 2006. On the composition of government spending, optimal fiscal policy, and endogenous growth: Theory and evidence.

[17] 
Schaltegger, C. A. and Torgler, B., 2006. Growth effects of public expenditure on the state and local level: evidence from a sample of rich governments. Applied Economics, 38(10), pp. 11811192.

[18] 
Hasnul, A. G., 2015. The Effects of Government Expenditure on Economic Growth: the Case of Malaysia.

Cite This Article


@article{10.11648/j.ijfbr.20241005.11,
author = {Wondimu Mekonnen},
title = {The Impact of Capital Expenditure on Ethiopian Economic Growth},
journal = {International Journal of Finance and Banking Research},
volume = {10},
number = {5},
pages = {8491},
doi = {10.11648/j.ijfbr.20241005.11},
url = {https://doi.org/10.11648/j.ijfbr.20241005.11},
eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.ijfbr.20241005.11},
abstract = {This study investigates the effect of disaggregated capital expenditure on economic growth in Ethiopia over the period from 1981 to 2021. An autoregressive distributed lag (ARDL) model, combined with a Granger causality test, is used for the econometric analysis. The empirical results reveal that capital expenditure on economic development (CAEE), recurrent expenditure (RCE), and the inflation rate (CPI) have significant positive impacts on economic growth in both the long run and short run. However, capital expenditure on administrative and general development (CEAG) has a significant negative effect on economic growth in the long run. The Granger causality test further indicates unidirectional causality from economic growth (GNI) to capital expenditure on economic development (CAEE). Based on these findings, the study recommends that the Ministry of Finance should increase its budget allocations for capital expenditures aimed at economic development in order to further stimulate economic growth. Additionally, the government should implement stricter followup and monitoring mechanisms to ensure the proper management of budget allocations, particularly regarding capital expenditures on social development and administrative and general development. These measures are crucial for sustaining longterm economic growth in Ethiopia.},
year = {2024}
}
Copy

Download

TY  JOUR
T1  The Impact of Capital Expenditure on Ethiopian Economic Growth
AU  Wondimu Mekonnen
Y1  2024/10/29
PY  2024
N1  https://doi.org/10.11648/j.ijfbr.20241005.11
DO  10.11648/j.ijfbr.20241005.11
T2  International Journal of Finance and Banking Research
JF  International Journal of Finance and Banking Research
JO  International Journal of Finance and Banking Research
SP  84
EP  91
PB  Science Publishing Group
SN  24722278
UR  https://doi.org/10.11648/j.ijfbr.20241005.11
AB  This study investigates the effect of disaggregated capital expenditure on economic growth in Ethiopia over the period from 1981 to 2021. An autoregressive distributed lag (ARDL) model, combined with a Granger causality test, is used for the econometric analysis. The empirical results reveal that capital expenditure on economic development (CAEE), recurrent expenditure (RCE), and the inflation rate (CPI) have significant positive impacts on economic growth in both the long run and short run. However, capital expenditure on administrative and general development (CEAG) has a significant negative effect on economic growth in the long run. The Granger causality test further indicates unidirectional causality from economic growth (GNI) to capital expenditure on economic development (CAEE). Based on these findings, the study recommends that the Ministry of Finance should increase its budget allocations for capital expenditures aimed at economic development in order to further stimulate economic growth. Additionally, the government should implement stricter followup and monitoring mechanisms to ensure the proper management of budget allocations, particularly regarding capital expenditures on social development and administrative and general development. These measures are crucial for sustaining longterm economic growth in Ethiopia.
VL  10
IS  5
ER 
Copy

Download