Public and private investments and economic growth in Ghana and Kenya

Mose, Naftaly; Fumey, Michael Provide Article Public and private investments and economic growth in Ghana and Kenya Financial Internet Quarterly Provided in Cooperation with: University of Information Technology and Management, Rzeszów Suggested Citation: Mose, Naftaly; Fumey, Michael Provide (2024) : Public and private investments and economic growth in Ghana and Kenya, Financial Internet Quarterly, ISSN 2719-3454, Sciendo, Warsaw, Vol. 20, Iss. 3, pp. 29-41, https://doi.org/10.2478/fiqf-2024-0017 This Version is available at: https://hdl.handle.net/10419/329878 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/3.0/ 10.2478/fiqf-2024-0017 Abstract A general conception is that investment induces economic growth, but there is still debate over which type of investment contributes more to economic growth. The disaggregation of investment into public and private components allows estimation of the impact of the two types of investments on economic growth. This research, therefore, empirically estimates the relationship between each investment component against economic growth by constructing panel data for Ghana and Kenya from 1991 to 2022. The empirical strategy adopted in this study can be divided into three major stages. First, the LLC unit root test in the panel series is undertaken. Second, if integrated in the same order, a Kao co-integration test is conducted. Finally, if the series is co-integrated, the vector of cointegration in the long run is estimated using the dynamic ordinary least squares (DOLS) method. Our estimation results, based on the panel cointegration approach confirm a long-run relationship between the study variables. Further analysis shows that public investment can promote economic growth in the long run. In contrast, the results indicate that private investment can obstruct growth. The study has shown that private investment did not always increase economic growth in Ghana and Kenya. The study findings indicate that public investment is more efficiently allocated in Ghana and Kenya than private investment, suggesting the best economic strategy is for private investment to be complementary and promote higher public investment to improve public sector productivity. Therefore, policymakers should focus on creating a favourable investment climate, providing fiscal stimulus and promoting public-private partnerships to enhance infrastructure development and stimulate private -sector investment, which can sustain long-term economic growth. JEL classification: E22, O47, R42 Keywords: Economic Growth, Investment Project Financing, Private Investment, Public Investment Received: 29.03.2024 Accepted: 05.06.2024 Cite this: Mose, N. & Fumey, M.D. (2024). Public and private investments and economic growth in Ghana and Kenya. Financial Internet Quarterly 20(3), pp. 29-41. © 2024 Naftaly Mose & Michael Provide Fumey, published by Sciendo. This work is licensed under the Creative Commons AttributionNonCommercial-NoDerivatives 3.0 License. 1 University of Eldoret, Kenya, e-mail: [email protected], ORCID: https://orcid.org/0000-0003-0467-235X. 2 Northwestern Polytechnical University, China, e-mail: [email protected], ORCID: https://orcid.org/0009-0000-32132233. growth in Africa, it needs to be accompanied by institutional improvements and stability to maximize its effectiveness. It is important to note that public investment can have a crowding-in effect on private investment in the long run, stimulating economic growth indirectly. Supplementing private and public investments with foreign direct investment (FDI), technology, human capital, and public consumption is essential for economic growth. Gross capital formation, labour growth, FDI and government expenditure explain economic growth in most developing countries (Yaremenko, 2022). FDI brings in external capital and expertise, stimulating economic activity and creating jobs (Popescu, 2014; Sirbulescu et al., 2022). It is essential for developing countries, with foreign direct investment (FDI) being a significant capital inflow. FDI has been found to positively impact economic growth and development, improving macroeconomic indicators and the country's image on the world stage (Sirbulescu et al., 2022). High institutional quality, including robust governance, transparent regulatory frameworks, effective legal systems, and reduced corruption levels, is crucial in attracting FDI (Lan et al., 2011). At the same time, public expenditure has mixed effects on economic growth. In most African countries, government consumption, and public investment do not positively influence economic growth in most cases (Ndiaye, 2018). According to Aydin and Demiröz (2023), human capital has been associated with long-term productivity gains, with a more significant effect from quality improvements. The two countries under consideration, Ghana and Kenya, albeit being in Sub-Saharan Africa, have somewhat diverse economic systems and policies. Uppermiddle-class Kenya is a growing market with a more diverse economy and a robust manufacturing and tourism industry; Ghana has a large agricultural sector and a burgeoning services sector. Concerning trade policies, Ghana has often maintained a more protectionist stance with slightly higher tariff barriers. In contrast, Kenya has adopted a more open trade policy, mainly due to its membership in regional economic bodies like the East African Community (EAC) (Wasike, 2022). Notwithstanding these differences, both countries have recently implemented fiscal stimulus programs and public investment plans to promote economic growth and development. While Ghana's government has focused on infrastructure development, particularly in the energy and transportation sectors, Kenya has made large expenditures in various areas, including housing, roads, and healthcare facilities (Ndiaye, 2018). These two countries were chosen for panel analysis because reliable data were available and because they had diverse economic strategies and developmental stages within the same geographic context. Investment is crucial in economic development, contributing to job creation, technological advancement, skill development, and productivity improvements (Abdulkarim, 2023). The evolution of investment project financing has led to a more holistic view that considers the role of the private sector in achieving global financial and economic goals (Reichelt et al., 2023). Measures for organizing favourable investment, financial, and economic conditions include economic stimulation, harmonization of business law, and access to markets (Sibirskaya et al., 2015). Public financial resources alone are insufficient to address sustainable change, so public actors are considering how to redirect private sector investment toward sustainable economic-related activities (Vydobora, 2022). Public and private investments contribute to economic growth, with public investment having a more substantial positive impact (Ezzahid & Rafik, 2023). The impact of public investment on economic growth may vary depending on the level of development and the degree of private-public capital substitutability (Farhadi, 2015). The empirical literature provides contradictory results, with some studies showing a positive impact of public investment on growth (Ocolișanu et al., 2022), while others find a detrimental effect (Merga, 2022). In the case of emerging European Union and Central European countries, the long-term impact of a public capital shock on Gross Domestic Product (GDP) is estimated to be negative (Merga, 2022). In advanced economies, public investment spending may have a relatively counterproductive effect on GDP growth due to high levels of crowding out (Ocolișanu et al., 2022). This is likely because of the advanced position of these countries in terms of transitional dynamics (Ari & Koc, 2020). Private investment positively influences economic growth in developing countries, as evidenced by multiple empirical studies (Merga, 2022). In Africa, public investment directly impacts output and contributes to the growth of total factor productivity (Ezzahid & Rafik, 2023). However, the effectiveness of public investment in stimulating growth can be influenced by various factors. The quality of institutions is an important determinant, as improvements in institutional quality can positively influence public investment and crowd-in private investment (Adeosun et al., 2020). Additionally, the level and stability of public investment are key factors. High and stable levels of public investment are more likely to promote economic growth, while low levels and high volatility can hinder growth (Wasike, 2022). On the other hand, private investment has been found to positively impact economic growth in both the short and long run (Ghani & Din, 2006). While public investment is essential for economic growth depends on the source of funds by the national government. If public investment is financed by higher direct tax, the net effect on economic growth may be negative despite a positive effect of marginal productivity of private capital. If public investment is financed by borrowing or debt, the economic agents will save the extra income due to non-taxation of today, to pay future taxes, and thus depress consumption (Gupta, 2020). Thus, increased government expenditure is offset by the low level of consumption and the impact of fiscal policy reduction (Yovo, 2017). According to Gupta (2020), an increase in income tax revenue can be utilized to finance investment in public capital, positively affecting the endogenous growth rate and the rate of savings. It also emphasizes the role of institutions, government policies, and incentives in promoting and supporting these endogenous factors. The Investment-led growth theory, as proposed by Harrod (1956), emphasizes the importance of investment in stimulating growth. It suggests that higher levels of investment, particularly in innovation, can contribute to both short-term recovery and long-term structural transformation. The main themes and assumptions of the investment-led growth theory can be summarized as follows. First, capital growth theory focuses on studying optimal financial investments for long-term growth (Evstigneev et al., 2015). Second, infrastructural investment is considered a necessary but insufficient condition for economic recovery and industrial growth. Third, the model emphasizes the conditions for capital growth and its relationship to other long-run investment models (Li & Hoi, 2014). Fourth, the impact of gross and foreign investments on economic growth is analyzed, with a positive correlation between investment and GDP growth (Bjelić, 2021). Table 1 presents the summary of empirical studies on the impact of investment components on economic growth across the globe. The first hypothesis builds on the Keynesian view, it argues that the regulation of economic activities by the government passes through countercyclical processes. This leads the government to be involved in the market by supporting the agent activities when the demand is depressed and to slow activities when inflation fears set in. In the short run, government expenditure can be employed to stimulate aggregate demand and stimulate growth (Yovo, 2017). The argument in favour of public investment is that some expenditures, especially public investment, such as electricity, roads, sewers, street lights, water systems, education and health generate externalities that enhance the productivity of private factors and thus boost economic growth (Blejer & Khan, 1984; Aschauer, 1989; Calderón & Servén, 2010; Yovo, 2017). The theory suggests that countries can foster sustainable and long-term economic growth by investing in education, health, and research and development (Alam et al., 2021). The model argues that productivity decline had been caused by a decline in public expenditure on infrastructure (Aschauer, 1989). The second hypothesis based on the neoclassical view, argues that accelerated public expenditure are harmful to economic growth in the long run. The recovery policies by public expenditure may even have a depressive impact on the economy mainly because of the crowding-out effects they exert on investment and private consumption. As a result, these negative effects influence economic agents' anticipation of future consequences of fiscal policy and lead them to adjust their behaviour accordingly to consumption and savings thus making fiscal policy tools impotent (Barro, 1990; Yovo, 2017). Ricardian equivalence theory says that financing government spending out of current or future taxes and deficits will have equivalent effects on the overall economy. The impact of public expenditure on growth Table 1: Review of Investment-Growth literature References Country (Period) Methods Results Khan & Reinhart (1990) Developing countries (1970-1979) OLS Public investment has no impact Private investment is beneficial Ghani & Din (2006) Pakistan (1973-2004) VAR Public investment has no impact Private investment is beneficial Makuyana & Odhiambo (2018) South Africa (1970-2017) ARDL Public investment is harmful Private investment is beneficial Nguyen & Trinh (2018) Vietnam (1990-2016) ARDL Public investment has no impact Private investment is beneficial Doménech & Sicilia (2021) OECD countries (1960-2017) Descriptive Public and private investments are beneficial Ahamed (2022) Developing countries (1990-2019) ARDL Public and private investments are beneficial vestment challenges after the financial meltdown of 2007-2008 and the COVID-19 pandemic period of 20202023 (Mose, 2021). Furthermore, Kenya and Ghana have experienced low economic growth rates and insufficient investment project financing to stimulate infrastructure development. The study used the dynamic ordinary least squares (DOLS) estimation technique during analysis. The secondary data was obtained from the World Bank online database (World Development Indicators). The dependent variable, economic growth, was measured as real Gross Domestic Product (GDP) per capita, whereas the independent variables were private and public investments. Table 2 shows the description of the variables used in the current study. The empirical literature review in Table 1 has confirmed that the relationship between investment components and economic growth is varied and inconclusive. Furthermore, very few studies have been carried out in Kenya and Ghana using the dynamic OLS (DOLS) estimation approach and the above studies have used different estimation methods and sample sizes. The current study will adopt a quantitative research design to analyze the trend and role of different types of investments on economic growth in Ghana and Kenya for the period 1991-2022. The period has been chosen since the two countries faced several inReferences Country (Period) Methods Results Jahan et al. (2022) Asian countries (2020) ANOVA Public investment has no impact Private investment is harmful Abdulkarim (2023) Nigeria (1981-2020) ARDL Public investment is harmful Private investment is beneficial Notes: Autoregressive Distributed Lag (ARDL); Ordinary Least Squares regression (OLS); Vector autoregression (VAR) Source: Author’s own work. Table 2: Description of study variables Variable Measure Proxy Data source Expected sign Dependent variable Economic growth (GDP) Constant US Dollar (2015) GDP per capita WDI Not predicted Abdulkarim (2023) Independent variables Public investment (PUB) Percent Gross capital formation (% GDP) WDI Positive (Ahamed, 2022) Private investment (PRV) Percent Domestic credit to private sector (% GDP) WDI Positive (Ahamed, 2022) Foreign direct investment (FDI) Percent Foreign direct investment, net inflows (% GDP) WDI Positive (Nguyen & Trinh, 2018) Public consumption (PC) Percent Government consumption expenditure (% GDP) WDI Negative (Ghani & Din, 2006) Human capital (HC) Number Labor force (total) WDI Positive (Makuyana & Odhiambo, 2018) Source: Author’s own work. (1) The econometric baseline model measuring the effect of selected variables on economic growth is rewritten as follows. (2) Where: GDPi,t - Economic growth, measured by GDP per capita growth, PUBi,t - Public investment, measured by gross capital formation, PRVi,t - Private investment, measured by domestic credit to the private sector, According to the neoclassical growth model of Solow output depends on capital and labour inputs (Solow, 1956). As hypothesized by Solow, an extension developed by Aschauer (1989); Barro (1990) as well as Yovo (2017) in the empirical literature is adopted. In the current study, it is hypothesized that economic growth depends on the investment used for the capital stock and on the labour force. The study adopts the empirical work of Yovo (2017) as well as Makuyana and Odhiambo (2018) to build a simple theoretical baseline estimation growth model as presented below. ( , , , , )GDP f PUB PRV FDI PC HC= , 0 1 , 2 , 3 , 4 , 5 , , i t i t i t i t i t i t i t GDP PUB PRV FDI PC HC        = + + + + + + test is Kao (1999), which is based on the Engel-Granger approach and controls for homogeneity on units in the panel set (Alam et al., 2021). If the cointegration is confirmed, this will suggest that the selected variables share the same stochastic trend, and thus, they can be combined in the long run. If the series is co-integrated, the panel vector error correction model is conducted to confirm the long-run causality or convergence of the target variables. In the analysis, the error correction term (ECT) value is expected to be negative and significant at 1%. Once the panel series is tested for causality using the vector of cointegration, the model is estimated using DOLS to check for long-run relationships between coefficients. In analysis, several estimation options are available when using panel data or between group data, for instance, panel ordinary least squares (POLS), pooled mean group (PMG), fully-modified OLS (FMOLS), and dynamic OLS (DOLS) estimation technique (Pedroni, 2001). The panel ordinary least squares (POLS) estimation technique is the most common analysis approach when combined with fixed or random effects. One advantage of POLS is that it can confirm the presence of convergence in panel data; however, sometimes POLS leads to biased estimates. Thus, to confirm the estimates that are not biased, the study can conduct analysis using FMOLS or DOLS estimation that favours small sample data. The DOLS model is robust to various departures from standard regression assumptions in terms of residual correlation, heteroscedasticity, misspecification of functional form and non-normality of residuals (Stock & Watson, 1993). Therefore, this study utilized the DOLS estimator pioneered by Stock and Watson (1993) to estimate the long-run relationship. Post-estimation panel diagnostic tests (heteroscedasticity, autocorrelation, cross-sectional dependence, and normality) were conducted during the regression analysis to avoid misleading inferences. Vector autoregression (VAR) lag selection criteria were applied to determine the lag length and best model estimator. The result of the VAR selection criteria is presented in Table 3. FDIi,t - Foreign direct investment, measured by foreign direct investment net inflow, PCi,t - Public consumption, measured by government consumption expenditure, HCi,t - Human capital, measured by Labour growth rate, β - the regression coefficient, εi,t - is the error term and the subscripts i and t represent country and time dimensions respectively. The econometric baseline model was rewritten in logarithm form (ln), to increase variance stability and remove outliers, as shown below. (3) To analyze the long-run effect of investment variables on economic growth using DOLS equation 3, above, was utilized during regression analysis. Vector autoregression (VAR) lag selection criteria were applied to confirm the lag length and best model estimator. The selection criteria are essential in reducing residual correlation problems. Countries like Ghana and Kenya can interact through imports, exports, human capital transfer, and economic integration, thus creating cross-sectional dependence by experiencing standard shocks. Therefore, this study's cross-sectional dependence test is essential to check if the data exhibit cross-sectional dependency (CD). The study adopted the Pesaran CD developed by Pesaran (2004) to check for cross-sectional dependency. Panel data series are usually characterized by stochastic trends that are easily removed through differencing. The selection of the unit root test depends upon cross-sectional dependence (Phillips & Sul, 2003). Various panel unit root tests, such as IPS, ADF, Phillips Perron, and Levin-LinChu (LLC) tests, are available. LLC unit root test is commonly used in cross-sectional dependence (Levin-LinChu 2002). If the unit root of the panel series is integrated in the same order, a cointegration test is performed. Hausman (1987) test was conducted to support the long-run estimation method used in the study based on OLS when combined with Fixed or Random effect. In this study, two cointegration tests have been suggested for use in investigating the long-run relationship. The first test is Pedroni (2001), and the second , 0 1 , 2 , 3 , 4 , 5 , , i t i t i t i t i t i t i t lnGDP lnPUB lnPRV lnFDI lnPC lnHC        = + + + + + + Table 3: Lag length selection criteria Lag LogL LR FPE AIC SC HQ 0 249.3659 NA 6.77e-12 -8.691639 -8.474637 -8.607508 1 443.7425 340.15910 2.38e-14 -14.347950 -12.828930 -13.759030 2 775.4439 509.39850 6.46e-19 -24.908710 -22.087680* -23.815000* 3 816.4799 54.2262* 6.06e-19* -25.088570* -20.965530 -23.490080 4 850.4569 37.61733 8.26e-19 -25.016320 -19.591270 -22.913040 Note: * = lag order selected by the criterion; SIC = Schwarz information criterion; AIC = Akaike information criterion; HQ: Hannan-Quinn information criterion; FPE: Final prediction error; LR: sequential modified LR test statistic (each test at 5% level) Source: Author’s own work. To check if the data exhibit cross-sectional dependence (CD), the Pesaran CD test developed by Pesaran (2004) was adopted by the study. The Pesaran CD result is reported in Table 4 below. According to the various lag length selection criteria, such as AIC and SIC, a lag of 4 is chosen for the VAR as it reported the minimum value. A lag of four was chosen to reduce the serial correlation problem. In addition, the AIC estimation model was preferred since it had the lowest value (-25.01632) compared to other estimation criteria. Table 4: Pesaran CD result Test Statistic Df Prob Breusch-Pagan LM -2.797984 1 0.0944 Pesaran scaled LM -1.271367 - 0.2036 Bias-corrected scaled LM -1.239109 - 0.2153 Pesaran CD -1.672718 - 0.0944 Source: Author’s own work. A unit root test was conducted to ascertain the order of integration and avoid spurious correlations in DOLS estimation. The study investigated the stationarity of study variables using the Levin-Lin-Chu (LLC) test. The stationarity result is reported in Table 5. Table 4 reports our Pesaran CD result; the test indicates the absence of cross-sectional dependence in our regression variables, confirmed by the P-value 0.0944 > 0.0500. Since the p-value of 0.0944 for the Z-statistic is more than 0.05, the null hypothesis should not be rejected. This confirms that cross-sectional dependence is not a problem. As a result, the study applied the Levin-Lin-Chu (LLC) test in unit root analysis. Table 5: Stationarity Test Variables LLC at Level Order LLC at first difference Order Statistic Prob. Statistic Prob. GDP 4.88497 1.0000 I(1) -4.09049*** 0.0000 I(0) PUB -0.01595 0.4936 I(1) -7.86468*** 0.0000 I(0) PRV 0.82469 0.7952 I(1) -6.33159*** 0.0000 I(0) FDI 0.62478 0.7339 I(1) -2.48678*** 0.0064 I(0) PC -0.63709 0.2620 I(1) -6.91668*** 0.0000 I(0) HC 68.57610 1.0000 I(1) -25.40740*** 0.0000 I(0) Note: *** Denotes significance at a 1% level of significance, H0: Panels contain unit roots, GDP - Economic growth; PUB - Public investment; PRV - Private investment; FDI - Foreign direct investment; PC - Public consumption; HC - Human consumption Source: Author’s own work. order and contain unit root. This validity offers the route to test for several cointegration relationships. To carry out a cointegration analysis based on OLS estimation, the Hausman test was conducted to choose between a fixed or random effect model. The Hausman test result is shown in Table 6 below. Based on the results presented in Table 5, economic growth, public investment, private investment, foreign direct investment, public consumption, and human capital were all non-stationary and integrated in the order I(1). However, they were transformed by the first difference to become stationary. The LLC test has confirmed that the variables are integrated in the same Table 6: Hausman Test output Test Summary Chi-Sq. Statistic Chi-Sq. d.f. Prob. Period random 26.729513*** 5 0.0001 Note: * * * Denotes significance at a 1% level of significance Source: Auhtor’s own work. Since all selected variables were integrated into the order I(1), a non-stationary cointegration test was performed to ascertain the long-run equilibrium and several cointegration relationships. Granger (1988) submits that if factors are integrated, the residuals will be integrated at the level, and if not, then first-order integration will be found. The study adopted the Kao cointegration test to check for the long-run relationships between selected variables. The fixed effect model is used during cointegration analysis. Table 7 shows the Kao cointegration test results. The Hausman test result supported using the fixed effect estimation method over the random method as reinforced by p-value (0.0001 < 0.0500). The fixed effect model is accepted because the chi-square statistic of the Hausman test rejected the null hypothesis of random effect. The panel data was estimated using the fixed effect model of the panel estimation technique, which is geared at controlling for time-invariant and unobservable country effects. Table 7: Kao Residual Cointegration Test t - Statistic Prob. ADF 2.369214 0.0089 Residual variance 0.001695 - HAC variance 0.000182 - Source: Author’s own work. The estimation of the vector error correction model (VECM) was to ascertain the direction of causation and adjust it towards the equilibrium. VECM is conducted to check for the long-run causality or convergence of the target variables. Table 8 shows the VECM analysis coefficients. Table 7 confirms the cointegration relationship since the ADF statistic is significant at a 1 % significance level. Hence, we reject the null hypothesis. Once the long-run relationship was confirmed, the study checked for long-run causality using vector error correction models. Table 8: Value of coefficients ECT Coefficient Std. Error t-Statistic Prob. C(1) -0.846009 0.210856 -4.012263 0.0015 C(2) -2.037543 0.491794 -4.143087 0.0012 C(3) 0.042720 0.831401 0.051383 0.9598 C(4) -0.161763 0.203908 -0.793312 0.4418 C(5) -0.214576 0.149145 -1.438708 0.1739 C(6) 1.900558 0.631102 3.011490 0.0100 C(7) 1.016450 0.494544 2.055329 0.0605 C(8) 0.388778 0.517284 0.751575 0.4657 C(9) -0.033721 0.506931 -0.066520 0.9480 C(10) 0.106619 0.079259 1.345198 0.2016 C(11) -0.088031 0.054159 -1.625418 0.1281 C(12) 0.226852 0.672490 0.337332 0.7413 C(13) -0.494261 0.743985 -0.664343 0.5181 C(14) -97.035220 44.565890 -2.177343 0.0485 C(15) -96.836900 46.206620 -2.095737 0.0562 C(16) 1.343803 0.628481 2.138177 0.0521 Source: Author’s own work. significant result also confirms long-run causality between investment components and economic growth. Moving forward, the study estimated the DOLS model to obtain the long-run coefficients Based on the result in Table 8, the error correction term (ECT) is negative and significant at 1 % (-0.846). This confirms the convergence; the economy can recover by 84.6% during the current year. The negative and private investment, public investment, and economic growth. Table 9 presents the long-run estimate coefficients. Once cointegration is confirmed among the selected variables, the study used the panel DOLS estimation technique to identify the long-run relationship between Table 9: DOLS Regression coefficients Variable Coefficient Standard error t-Statistics p-value PUB 0.098096 0.045578 2.152270 0.0405** PRV -0.252063 0.042296 -5.959449 0.0000*** FDI 0.041242 0.013686 3.013395 0.0056*** PC -0.282784 0.062964 -4.491214 0.0001*** HC 0.904220 0.058087 15.566760 0.0000*** Goodness of Fit R2 = 0.985673 Adjusted R2 = 0.969223 F = 44.557460 P-value(F) = 0.000000 Breusch-Pagan F = 0.861208 Prob > F = 0.523759 Breusch-Godfrey χ2 = 3.733390 Prob > χ2 = 0.058310 Pesaran CD Z = - 1.672718 Prob > Z = 0.094400 Jarque-Bera χ2 = 0.955339 P-value(χ2) = 0.620227 Durbin-Watson DW = 2.111040 - Note: * p < 0.1, ** p < 0.05, *** p < 0.01 are significance levels in which the null hypothesis is rejected. GDP - Economic growth; PUB - Public investment; PRV - Private investment; FDI - Foreign direct investment; PC - Public consumption; HC - Human consumption Source: Author’s own work. vén, 2010). The empirical literature provides contradictory results, with some studies showing a positive impact of public investment on growth (Ari & Koc, 2020; Ahamed 2022), while others find a detrimental effect (Makuyana & Odhiambo, 2018; Nguyen & Trinh, 2018). The study findings support the similar study by Ghani and Din (2006) in Pakistan and Ahamed (2022) for developing countries, where the studies argued the significance of public investment in inducing growth through positive spillover to the private sector (Ari & Koc, 2020). However, these findings contradict Makuyana and Odhiambo's (2018) empirical study in South Africa, where the study reported a negative relationship between public investment and growth. This was attributed to the crowding effect of private investment by public investment. Finally, Nguyen and Trinh (2018) found the relationship insignificant in Vietnam. The effect of private investment on growth is negative and significant at a 1 percent significance level. This means that a 1 percent increase in private investment will cause economic growth to decline by 0.25 percent. Based on the empirical result, private investment slows economic growth in Ghana and Kenya. This is against theoretical expectations, where private investment influences growth by exhibiting increasing returns to scale (Doménech & Sicilia, 2021). Sometimes, public investment can crowd out private investment, thus compromising growth in developing countries. Furthermore, negative results could be attributed Based on the result of Table 9, the coefficient of public investment is positive and statistically significant at a 5 percent significance level, as per the p-value of the coefficient. Since the p-value of 0.0405 for the tstatistic is less than 0.05, the null hypothesis should be rejected, suggesting the variable is significant. It means that a 1 percent increase in public investment will lead to about a 0.09 percent increase in economic growth, as expected in Ghana and Kenya. This implies that public investment enables economic growth in both countries through increased efficiency and productivity. Investment in core infrastructure stimulates total factor productivity growth in core sectors and private sector organizations, thus promoting growth (Aschauer, 1989). Both countries have different budgetary allocations towards government expenditure, but considering both have excess labour supply, the ability of the public sector to influence growth is higher than that of the private sector. According to macroeconomic literature, public investment can stimulate economic activities via short-term effects on aggregate demand by raising the productivity of the private sector (Ghani & Din, 2006). Most developing countries employ expansionary fiscal policy tools and fiscal stimulus packages to grow the public sector and thus encourage productivity and efficiency (Doménech & Sicilia, 2021). Public investment generates positive spillover by providing public goods such as education, health, and infrastructure development (Ghani & Din, 2006; Calderón & Ser-