Trends and determinants of domestic private investment in Ethiopia: Time series analysis

Abstract The Ethiopian Investment Commission was established to formulate suitable investment policy with good investment climate that enables investors to operate in a more macroeconomic friendly environment. Despite all efforts, domestic private investment in Ethiopia is still in its early stages. Therefore, this study aimed to investigate the trends and determinants of domestic private investment in Ethiopia using a time series data over the period 1992–2022. To meet its goal, the study employed an Autoregressive Distributed Lag (ARDL) bounds testing approach through E-views version 12. The findings show that domestic private investment was negatively and significantly affected by the inflation rate, public investment, and real effective exchange rate over a period of both the short- and long-run. While domestic credit to the private sector, foreign direct investment, real GDP and trade openness were found to have positive and significant effect on domestic private investment in long run. Unemployment rate was found to have positive and significant effect in short run but insignificant in long run. Annual interest rate was found negative significant effect in short run but insignificant in long run. While total government expenditure was insignificant in both short and long run. Inflation has a negative relation with domestic private investment in both short and long run, therefore the study suggested that policymakers should recognize the cause for fluctuations in inflation and keep in a stable manner.


PUBLIC INTEREST STATEMENT
Private investment in general and domestic private investment in particular plays a vital role in expanding the productive capacity of the economy.Ethiopian investment commission has doing on enhancing the growth of the private investment by creating an enabling environment both in the domestic and foreign markets.Even though the situation of investment has been improved from the previous period by lifting various challenges and boosting public infrastructure with the country's favorable investment climate, the participation of domestic private investment is still at its infancy stage.This indicates that the government is so much concerned about policies to boost private investment without much knowledge on the factors that could influence it.The main objective of this study is therefore to analyse the trends and determinants of macroeconomic factors that influence the development of domestic private investment in Ethiopia from the period 1992 to 2022.

Introduction
Private investment in general and domestic private investment in particular play a vital role in expanding the productive capacity of the economy and promoting the long term economic growth (World Bank Group, 2013).Bakare (2011) defines investment to be an operation involving the purchase of items that will be used right away rather to consume immediately.It is an act of current spending for expected future return.According to Sachs (2006) investment is the accumulation of newly produced physical entities, such as factories, machinery, houses, goods, and inventories.
Empirically, countries that were able to accumulate high levels of investment achieved faster rates of economic growth and development (Akçay & Karasoy, 2020).One of the key parameters that divides industrialized countries from a developing country is their degree of investment.Investment is the major foundation of enhancement in the level of literacy, improvement in technology and increase in the capital stock (Nwankwo & Allison, 2021).Investment activities can be done by two main sectors, public and private.Majority of public investment commonly focused to finance physical and non-physical development infrastructure which could not be conducted by society (Resmini, 2019).However, private investment, both foreign and domestic, increases the productive capacity, creates employment opportunities, promotes technical advancement, raising growth rates, increasing export and introducing innovations, and reduces poverty in the country (World Bank, 2019).The private investment is a crucial pre-requisite for economic growth because it allows entrepreneurs to set economic activity in action by making resources together to produce goods and services (Legass et al., 2022).
The government of Ethiopia follows an integrated 5 year development plan, preparing the GTP II five year program (2015/16-2019/2020) as well as achieve the Sustainable Development Goals and attain middle-class income status by 2025 in which private investment has play great role (Ethiopian Investment Commission (EIC, 2017).Ethiopia is making a concerted effort towards structural transformation where manufacturing (especially investment) is expected to play a noticeable role in the economy (Debebe & Bessie, 2022).The country is taking steps to make the private sector to by targeting to reduce its interest rate, provides investment incentive and expanding access to medium and long-term finance (EIC, 2021).Ethiopian Investment Commission has also doing on enhancing the growth of the private business by creating an enabling environment both in the domestic and foreign markets (role Ministry of Finance and Economic Cooperation ([MOFEC], 2021).According to the Ethiopian Investment Commission (EIC, 2021)), a total of 113,127 private sector investment projects were registered across all regional states and city administrations between 1991/2002 and mid-2021.Among the projects, the majority of the investment is owned by domestic private investors (94.75%) (107,189 projects), and the remaining 5.25% (5,938 projects) is owned by foreigners.Of the total investment projects registered, 7.87% or 8,901 projects are in the implementation stage 43,363 projects, or 38.33%, have launched operations, while the remaining 59,400 projects, or 52.5%, are pre-implementation (licensed investment) projects bymid-2021.This means that less than half (46%) of the total registered private investment projects are converted into actual investments, indicating the slow pace of implementation of private sector investment projects.
In Ethiopia, foreign private projects are capital intensive while domestic private projects have higher share in employment creation (Gebremariam & Ying, 2022).Investment in the country shows progressive trends with speedy starting from announcement of liberal policy in 1992.Even if the situation of investment has improved from the previous period, the participation of domestic private sector is not satisfactory (Ago & Dollo, 2020).Even though the country has favorable investment climate, the growth of domestic private investment remains very low (World Bank, 2020).Ethiopia's public investment rate is the third highest in the world, but private investment rate is the sixth lowest (World Bank, 2018).After the introduction of the current government; Prosperity Part, through lifting various challenges, Ethiopia tries to give support for the domestic private investment.Though growing, as compared to the public infrastructure boosting, the expansion of domestic private investment is still at its infancy (Debebe & Bessie, 2022;EIC, 2021).This indicates that the government is so much concerned about policies to boost private investment without much knowledge on the factors that could influence domestic private investment.
Various studies have been done related to this topic in Ethiopia.Most of them (Debebe & Bessie, 2022;Jobir, 2023;Kibret, 2018;Legass et al., 2022;Waktola, 2020) are studied about macroeconomic determinants of private investment using secondary data and VAR Model.However, they draw contradicting conclusions on some variables such as inflation, and interest rate are positive effects on private investment.Other studies focus on the determinants of public investment (Adugna, 2013;Tilahun & Tajani, 2021) and they state private investment as one of the explanatory variable.Shiferaw (2016) examined both macro and microeconomic determinants of private investment in Ethiopia at regional level using primary data.However, it is important to study the determinants of private investment at country wise because all policy review, evaluation and corrections are made at national level.Some of them (Abinet, 2022;Aklilu, 2021;Saxena, 2021;Waktole & Bogale, 2018) are even based on primary data that are collated from specific town and city of Jimma City, Dire Dawa City, Debre Tabor Town and Finoteselam and Bure Towns, respectively, which cannot conclude the determinants and the effects of private investment as country wise.Others studies like Teklay (2017) only focus on financial determinants of private investment using time series data.However, he missed other influential variables like government expenditure, unemployment rate and, domestic credit to the private sector and public investment.
In knowledge of researcher's as reviewing the previous literatures, there is only one published study Abate (2016) to date that address macroeconomic determinants of domestic private investment in Ethiopia by using VAR model from the time period of 1971 to 2014.Though his study was related to this study than the others, he is missing important variables that might extremely determine domestic private investment in Ethiopia like external debt service, government expenditure, unemployment rate and Foreign Direct Investment (FDI).He also used VAR model to analyze the data, however, it is advised if ARDL model is used for analyzing of determinant variables as most of previous studies are used it.For the reason that ARDL model uses a combination of endogenous and exogenous variables, unlike a VAR model that's strictly for endogenous variables.Accordingly, the existence of contradicting conclusions, and adding irrelevant or omission of relevant variables in the previous studies initiates the researcher to another new investigation.Hence, in depth investigation of these missed variables incorporating with others is required by using the ARDL model from the time period of 1992 to 2022.Since the effect of such variables varies depending on the time period covered (due to the nature of the variables has short and long-run effects) and methodology followed.This study is thus an effort to fill up this gap in knowledge and provide recent information regarding the trends and determinants of domestic private investment in Ethiopia by adopting relevant analytical methodology (ARDL model) through using macro-economic variable such as real GDP, real effective exchange rate, lending interest rate, external debt service, trade openness, average annual inflation rate, domestic credit to the private sector, unemployment rate, total government expenditure and foreign direct investment.The main objective of this study is therefore to analyse the trends and determinants of macroeconomic factors that influence development of domestic private investment in Ethiopia from the period 1992 to 2022 based on secondary data.

Theoretical literature reviews
In explaining investment, a few theoretical approaches are at work in different pieces of literature and underpinning studies on investment decisions.These models fit the argument with the realization of theoretical baselines about investment.Among those approaches, the accelerator theory, flexible accelerator theory, neoclassical theory, and the Expected Profits also known as the Cash Flow or Tobin's Q theory/models of investment behavior have wider coverage in investment literature.

Accelerator theory of investment
After Keynes, the accelerator principle was the dominant theory of investment behavior especially during the 1950s and early 1960s (Ghura & Goodwin, 2000).The accelerator theory of investment postulated the linear relationship between investment and output.The theory stipulates that capital investment outlay is a function of output.According to this theory, given the incremental capital/output ratio, it is easy to compute the investment requirements associated with a given target for output growth which means that an increase in the rate of output of a firm will require a proportionate increase in its capital stock.

Flexible accelerator model
Flexible accelerator model is a modified version of the accelerator theory/model based on the optimal accumulation of capital with an assumption that investment is a function of the level of output and the user cost of capital.The basic idea of the flexible accelerator model is to take account of lagged effects (Khan and Kumar 19,977).This model assumes that investment is not only determined by the current change in output but also by its earlier changes.However, this theory has limitations with regard to its assumption of perfect competition and exogenously determined output.

Tobin's Q-Theory of investment
James Tobin (Wildasin, 1984) has proposed the Q-theory of investment which links a firm's investment decisions to fluctuations in the stock market.Tobin postulated that investment decisions are functions of the ratio of the addition to the value of the firm due to an extra unit of capital installed to its replacement cost.When a firm finances its capital for investment by issuing shares in the stock market, its share prices reflect the investment decisions of the firm.Stock prices show the incentive to invest.The advantage of Tobin's q as a measure of the incentive to invest in that it reflects the expected future profitability of capital as well as the current profitability.For example, if corporate tax is reduced next year, the value of the stock today, Tobin-Q, will rise, increase investment.

Neoclassical theory of investment
The neoclassical approach or the user cost model is a version of the flexible accelerator model.It was formulated and developed by Jorgenson and others in 1971 based on the determination of the optimal capital stock.This theory argues that investment depends on the rate of interest and the level of income (Jorgenson, 1971).According to this theory, the investment that depends on the rate of interest focuses heavily on the cost of finance as a key variable; all other costs (including the availability of finance, economic infrastructure, source of capital, policy, etc.) being assumed given.The neoclassical model, in contrast with the accelerator model, assumes that the desired stock depends not only on planned output but also on the ratio of output price to the implicit rental price of the services of capital goods (Bischoff, 1971).

Summary of theoretical literatures reviews
In developing countries, the above theoretical models of private investment have been applied with a fair degree of success (Rivas & Villarroya, 2017).Nevertheless, empirical studies have not yet clarified which of those models is more accurate to represent the way in which capital formation occurs in developing countries.Because, in developing countries like Ethiopia, there are no wellfunctioning financial markets; there is no perfect market, and there is no accurate data (Martin & Rogers, 2000).Hence, in the investigation of determinants of private investment, one should consider the political and institutional factors (governance issues) besides macroeconomic factors because it is not only the macroeconomic factors that matter for the performance of private investment but also governance or political and institutional factors (Acosta & Loza, 2005;Chen & Feng, 1996)

Empirical reviews
Related empirical literatures from Ethiopia and the rest of the World were summarized in the below Table 1 and 2.

Research gap
Today in Ethiopia, domestic private investment has got important place as engine of economic growth and development.In Ethiopia, the presence of little empirical analysis in this context makes this study vital to show the role of the private investment in the economy and to help the policy formulation incentive provision to the sector.This research can contributes to the existing literature on why the performance of domestic private investment is low despite the macroeconomic reforms and incentive package for investment.This research also focused on studying the key determinants that are critically influence domestic private investment in Ethiopia based on the data gather from 1992 to 2022.The study period is selected based on the: major economic reforms encompassing currency devaluation, trade liberalization, deregulation of markets, removal of restrictions on private sector participation, and modest privatization and reform of State-Owned Enterprises.Hence, it is assumed that major economic variables are somehow "normal", can contribute positively to economic growth.Furthermore, only a few explanatory variables were addressed under the previous studies that focus on determinants of domestic private investment, and there is no the same sign agreement on the significance of explanatory variables on the dependent variable.Accordingly, the existence of contradicting conclusions and omission of relevant variables in the previous studies initiates the researcher to another new investigation.Hence, in-depth investigation of incorporating new explanatory variables that are not yet studied is important.Since the effect of such variables varies depending on the time period covered and methodology followed.

Research design
A research design is a master plan that specifies the methods and procedures for measurement and analyzing the needed information (Abbott & McKinney, 2013).The study used quantitative research design to investigate the trends and determinants of domestic private investment in Ethiopia.The study uses secondary data which is readily available for convenience, in terms of time available.31 years data (i.e. from 1992 to 2022) were gathered for some important variables.The data set was restricted to this period due to the availability of consistent information especially about the private sector.The collected secondary data was summarized using tables and graphs.Then, the data were analyzed using ARDL model.

Data types and sources
The annual time series data were collected ranging from 1992 to 2022 from World Bank Database (World Development Indicators), National Bank of Ethiopia (NBE) and Ethiopian Investment Commission (EIC) for macroeconomic variables and trends of domestic private investment in Ethiopia.

Method of data analysis
The study used ARDL model in order to assess the short-run and long-run effects of independent variables on dependent variable.Econometric time series estimation technic was used and analyzed using EVIEWS version 12 statistical software package.

Specification and justification for ARDL bounds test
The theoretical models such as Keynesian, neoclassical and neo-liberal alone cannot identify the determinants of domestic private investment.Hence, the study used the autoregressive distributed lag (ARDL)-bounds testing approach developed by Pesaran et al. (2001) to examine the determinants of domestic private investment in Ethiopia.The ARDL model proves to be a suitable method to estimate the influence of time lag effects of explanatory variables in multivariate time series analysis.Unlike the competing VAR models, the ARDL model treats the parsimonious lags of the right-hand variables as an independent variable (Pesaran et al., 1999).The approach has advantages over the other cointegration techniques such as the Engle and Granger (1987) approach and the Johansen and Juselius (1990) approach.First, the variables included in the model do not have to be integrated of the same order, it can be a mixture of I(0) an I(1) (Pesaran et al., 2001).Second, it reduces serial correlation and endogeneity problems.Third, it renders robust results for small samples.Fourth, ARDL bound test approach estimate the long and short-run parameters of the model simultaneously.Fifthly, ARDL approach to co-integration helps in identifying the cointegrating vector(s) and if one co-integrating vector is identified the ARDL model of the cointegrating vector is re-parameterized into ECM.The re-parameterized result gives short-run dynamics and long-run relationship of the variables of a single model.Hence, the empirical model used to investigate the determinants of domestic private investment follows that of Karagöz (2010, Ngoma andBerke 2020).Maluleke et al. (2023) among others.In the ARDL model the variables are expressed as follows: The model explains that domestic private investment (PINV) is subject to changes in real GDP, Public Investment(PUINV), Annual Inflation Rate(AIR), Real Effective Exchange Rate (REER), Trade Openness (OPEN), Annual lending interest rate (IR), External Debt service (DEBT), domestic credit to the private sector (CREDIT), Unemployment Rate(UNEMP), Total Government expenditure(EXPEN) and Foreign Direct Investment (FDI).
Therefore, the ARDL model specified for this study as follows: Where: lnPINV is log (natural logarithm) of Domestic Private Investment, lnRGDP is log of Real Gross Domestic Product, lnPUINV is log of Public Investment, lnAIR is log of Annual Inflation Rate, lnREER is log of Real Effective Exchange Rate, lnOPEN is log of Trade Openness, lnIR is log of Annual lending interest rate, lnDEBT is log of External Debt service, lnCREDIT is log of Domestic Credit to the private sector, lnUNEMP is log of Unemployment Rate, lnEXPEN is log of Total Government expenditure and lnFDI is log of Foreign Direct Investment, Δ is the first difference of a variable, μ 1t is error term, α 0 is constant term, α & β is the respective coefficients and n is lag length.

The error correction model (ECM) of the ARDL model is expressed as follows:
Where ECM t −1 is the error correction term is lagged once and θ is the coefficient.
The existence of long-run relationship among the variables in co-integrating equation was determined by the value of F-statistic.The decision rule is that if the computed F-statistic is greater than the upper critical bound at 5% significant level, it indicates there is long-run relationship among the variables, suggesting co-integration.On the other hand, if the computed F-statistic is less than the lower critical bound, it shows that there is no co-integration.When the F-statistic lies between the lower and the upper bound critical values, the result is inconclusive (Pesaran et al., 1999).

Diagnostic tests
The study has ensured the fitness of model, by using pre-estimation test including Unit root test, Co-integration test and maximum lag length, and post estimation test which include stability test, normality test, auto-correlation test, heteroscedasticity test, multicollinearity test and model misspecification test to ensure that the data fits the basic assumptions of the ARDL model.

Description of variables and their expectation symbols
Based on the checkup of the existing related literatures (by understanding the effect of variables') and the researcher's knowledge (observing and predicting the necessity of variables in situation of Ethiopia), the major independent variables are selected such as RGDP, public investment, annual inflation rate, real effective exchange rate, trade openness, annual inflation rate, external debt service, domestic credit to the private sector, unemployment rate, total government expenditure and foreign direct investment.

Descriptive statistics
Table 3 shows that the mean value of domestic private investment is 2.48806 while public investment is 2.736766 in Ethiopia, meaning that the average value of domestic private investment (2.48%) less growing than public investment (2.745).However, the variation is very less.Real GDP is one of the most primary economic healths of the country.If GDP grows, the likelihood of selling manufacturing products also grows and domestic private investments are likely to benefit from that apprise of higher profits.The mean value of GDP across the study periods has 8.964516 and the standard deviation has 2.514829 which indicate that there is a good variation in the value of GDP rate across the study years.The maximum and minimum values are 12.6000 and 1.60000, respectively (Table 3).The mean value of exchange rate for domestic private investment during the study period is about 15.51160 and the value of standard deviation is 12.85085 which imply that 15.51 percent of average of Ethiopian birr per USD and the presence of good variation among the USD across the study period included for this study.Exchange rate is easily monitored and provides good collaterals.The maximum and minimum values were 52.07480 and 2.07000 respectively (Table 3).The mean value of external debt service for domestic private investment during the study period is about 8.047744 and the value of standard deviation is 1.427786 which indicates that there is a good variation in the value of external debt service across the study years.The maximum and minimum values are 11.46570 and 6.234782, respectively (Table 3).

Trends of domestic private investment in Ethiopia
As shown in Figure 1, domestic private investment in Ethiopia increases throughout the year due to suitable investment climate for the investor and appropriate investment policy.Nevertheless, in 1997(6.5009%), the trend shows that there is a downfall of domestic private investment due to high inflationary rate and then starts to rise in 2000 (11.847%).The domestic private investment as a percentage share of GDP shows different trends in Ethiopia from 8.4308% of GDP in 1992 to the end of period specified in this study 20.0245% of GDP in 2022.In the period the maximum amount 67.99% registered in 2008, while the minimum 6.50% in 1997.In the study period on average the percentage share of private investment to GDP is 17.34%, the Ethio-Eritrean war period (1998-1999) registered a smallest while the Ethiopian millennium year (2008) largest share of domestic private investment shown (Figure 1).The percentage share of domestic private investment to GDP in Ethiopian was declined between 2019 (21.4325%) to 2021 (12.5434%) for the reasons of COVID-19 pandemic disease, political tensions, and a devastating conflict in the Tigray region (Figure 1).The economic impact of COVID-19 includes the increased price of basic foods, rising unemployment, a slowdown in growth, and an increase in poverty.An underdeveloped private sector investment would limit the country's trade competitiveness and resilience to shocks (World Bank, 2021).According to the Primary Manufacturing Survey (2020), 84.64% of the interviewed domestic private sector investors reported that COVID-19 adversely affected their investment (EIC, 2021).In general, the last three years record (2019-2021) shows a minimum development of domestic private investment in Ethiopia relative to the 2018 (21.4326%).

Unit root test results
The primary step before starting the empirical analysis is to test for stationary properties of determinants of domestic private investment to check whether a series is stationary or not.Because using the classical estimation methods to estimate relationships with non-stationary variables results in spurious regression (Gujarati et al., 2004).If the variables are all stationary in level, we apply an estimate based on a linear regression.On the other hand, if the variables are all stationary into the first difference, our estimates are based on an estimate of the ARDL model (Pesaran et al., 2001).To test for stationarity, the study utilizes the Augmented Dickey-Fuller (ADF) and Phillips-Perron (PP) unit root test.The results of the stationarity tests in levels and at first difference for all the variables are presented in  5).It means that the t-statistics is greater than the t-critical value at 5% significant level.However, some variables that include domestic credit to the private sector, unemployment and external debt stock were nonstationarity at level ( smaller than the critical value in absolute terms and vice versa.This means some variables have to be differenced to make them stationary.The results show that the ADF and PP t-statistics exceeded the t-critical value at 5% significant level for all the variables at first difference (Table 5).The implication is that all the variables are found to be stationary at first difference.
The null hypothesis stated that data are non-stationary, or contains a unit root is rejected at 5% significance level.Therefore, the analysis can be performed Co-integration test using the ARDL bounds testing approach.

Maximum lag selection
Cointegration test is usually preceded by a test of optimal lag length selection because of the test is affected by the number of lags included in the ARDL model (Smith, 2001).The maximum lag length for this study was determined by using the Schwarz information criteria (SIC) as this method has been confirmed in most empirical studies to be superior to other tests.The values indicated by astrix (*) show that lag order that are selected by criterion at 5% level of significance.
As indicated in Table 4 the optimal lag length for this study is the SC one (1).Because the lag order that has many number of astrix is more optimal than few astrix (Pesaran et al., 2001).

Bound tests for cointegration
There is a long-run relationship among all the variables at the time of their F-statistic values are greater than the upper-bound critical value at the 5% level (Pesaran et al., 2001).Accordingly, the results show that the computed F-statistic of 12.38438 are greater than the upper critical bound value of 3.04 at 5% significant level (Table 6).This indicates that the variables are cointegrated.Therefore, the study rejects the null hypothesis of no cointegration and decides that there exists a long-run equilibrium relationship between domestic private investment and the explanatory variables.

Results of post-estimation diagnostic test
It is mandatory to test the econometric assumptions in order to know the proposed model statistical good fit.To accept this model as a good one, the study conducted post estimation tests such as normality, model misspecification, stability, auto-correlation, heteroscedasticity, and multicollinearity.

Normality test
The most commonly applied test for normality is the Jarque-Bera (JB) test.The JB uses the property of normally distributed random variable that the entire distribution is characterized by the first two moments, those are mean and variance.As shown in the graph, the result of Prob Chi = 0.411251 > α = 5% or 0.05.Since, the histogram is bell-shaped and JB statistic is not significant, which means P-value given at the bottom of the normality was >5%, so it is bigger than it and normal (Figure 2).Thus, the null hypothesis (Ho= Error terms are normally distributed) cannot be rejected rather it is accepted.

Model misspecification, auto-correlation and heteroskedasticity tests
To check the Misspecification of the model, the Ramsey RESET test was used.Ramsey test is used to check whether there is an omitted variable or not in the model besides, the specification error of incorrect functional form and correlation between explanatory variables and error term will be tested (Shrestha & Chowdhury, 2005).The test reports that the p-value of F-statistic and likelihood ratio were greater than 5% significance level (0.3291 and 0.0513 > α = 5% or 0.05) (Table 7).Therefore, the model is free from misspecification.To check the existence of autocorrelation the study used Breusch-Godfrey Serial Correlation LM test.According to Brooks (2008), the disturbance term is the covariance between the error terms over time is zero.In other words, it is assumed that the errors are uncorrelated with one another.The test reports that the p-value of F-statistic and R-squared were greater than 5% significance level (0.6472 and 0.3932 > α = 5% or 0.05), respectively (Table 7).Both F-statistic and R-squared results were insignificant.The conclusion from both versions of the test in this case is that the null hypothesis (H0= No Autocorrelation) is accepted.
This study used both Breusch-Pagan-Godfrey and ARCH tests to check the existence of Heteroskedasticity.The assumption of homoscedasticity says that the variance of the error term is constant, σ2 this is known as the assumption of homoscedasticity.If the residuals of the regression have systematically changing variability over the sample, that is a sign of Heteroskedasticity.If the errors do not have constant variance they are said to be Heteroskedasticity (Brooks, 2008).In this study; the test result presented that both F-statistics and observed R-squared shows that there is no evidence that display heteroskedasticity because the result in the P value is more than 0.05 (Table 7).In this case, the study fails to reject the null hypothesis of constant variance (homoscedasticity) and concluding that Heteroskedasticity is not present in the data.

Stability test
The cumulative sum of recursive residuals (CUSUM) and cumulative sum of squares of recursive residuals (CUSUMQ) are conducted to find out whether the model parameters are stable or not.To be saying the data are stable, the blue line on the graph should be rested between two red lines.The CUSUM and CUSUMQ results presented in Figures 3 suggest that the estimated model is stable.The study fails to reject the null hypothesis (Ho= the model is stable at all conventional levels of significance) rather it is accepted.

Multicollinearity test
In this study, to check the presence of multicollinearity among explanatory variables, Variance Inflation Factors (VIF) was used.According to Schroeder et al. (1990) multicollinearity problem would be corrected when the centered VIF value should be less than uncentered VIF.In this study,   all the values of centered VIF are less than uncentered VIF which fails to reject the null hypothesis (non-existence of multicollinearity) and concluding that multicollinearity problem is not present in between variables (Table 8).negative (−1.050724) and highly significant considering the probability value (0.0000).The negative sign of the coefficient of the error correction model indicates that short-run shock was above the long-run equilibrium value.The speed of the adjustment is captured by the magnitude of the error correction coefficient (1.050724).This implies about 105% only of the disequilibrium in domestic private investment was restored in the short-run.This is a very slow speed of adjustment.Furthermore; the value of the R 2 implies that about 97.95% (0.979517) of variations in domestic private investment are explained by the variations in the independent variables considered (Table 9).Therefore, the goodness of fit of the short-run model is proved to be strong.

Short-run results (estimation of error correction model)
The short-run result indicates that all individual variables are statistically significant except external debt burden and foreign direct investment (Table 9).Explanatory variables such as inflation rate, public investment, real effective exchange rate and annual interest rate were found negative and significant effect on domestic private investment in short run (Table 9).Unemployment rate was found positive and significant effect on domestic private investment  while external debt service has positive effect but insignificant (Table 9).Foreign direct investment was insignificant and negative relationship with domestic private investment.

The long run estimation
The technique involves first estimating the model using ARDL and testing the effect of variable in long run by using "long run form and bound test".The results are stated in Table 10 .The long-run ARDL model was estimated based on the Schwarz Criterion (SC) using the maximum lag length of one.In long run, explanatory variables like inflation rate, external debt burden, public investment and real effective exchange rate were found negative and significant effect on domestic private investment (Table 10).Domestic credit to the private sector, foreign direct investment, real GDP and trade openness were found positive and significant effect on domestic private investment.
Total government expenditure and unemployment rate were insignificant and positive effect on domestic private investment while interest rate was insignificant and negative effect (Table 10).

Annual inflation rate
In this study, Annual Inflation Rate (AIR) has a negative and highly significant effect at 1% level of significance on domestic private investment in both short run and long run, which is in line with the prior expectation.This means keeping other things constant, a 1% increase in AIR results 1.10% (−1.105145) (Table 9) and 26.74% (−0.267475) (Table 10) decrease the growth of domestic private investment in the short and long run, respectively.High rates of inflation adversely affect domestic private investment activity by increasing the riskiness of longer-term investment projects, reducing the average maturity of commercial loans, and misrepresenting the information conveyed by prices in the economy (Akpalu, 2011).In other way the existence of high inflation causes for the weakening of purchasing power of money, less space to save money in the bank and limits credit available for finance investment projects and affects domestic private investment negatively.The current high inflation rate in Ethiopian seems to have affected domestic private investment in the short run by undercutting the saving capacity of citizens by generating diversion of investment from productive sector to speculative activities.
The finding was consistent with findings of Alemnew (2015) and Saxena (2021) in Ethiopia and they found that inflation rate had a negative effect on private domestic investment in both long run and short run.The same result was reported also in Malawi by Maluleke et al. (2023) and in Gambia by Ayeni and Nsiah (2020).However, findings were contradicted with Legass et al. (2022) and Waktola (2020) in Ethiopia; they found that inflation rate had a positive relationship with private investment.The consequence of their finding is that as prices of goods and services are rising higher in Ethiopia, a profit maximize individual sees it as an opportunity to make abnormal profits, thus venturing into such businesses in order to share in the perceived excessive gains.The findings were also contradicted with Ajide (2013), and Molapo and Damane (2015) conducted in other developing counties states that inflation affects significantly and positively the domestic private investment.

Public investment
As many scholars discussed in literatures, public investment and domestic private investment have both crowding-out and crowding-in effects.The effect of public investment in this result was negative and significant at 1% level of significance in both short run and long run which is similar with the prior expectation.Other things remaining the same, as public investment increases by 1%, domestic private investment declines by 6.71% (−0.067095) (Table 9) and 23.21% (−0.232112) (Table 10) in the short and long run, respectively.The result of public investment indicates that it has "crowding out" effect on the domestic private investment, as public and private sectors compete for the same resources in the economy.It means that there is a resource competition (finance and market) between public and domestic private investment.Firstly, Ethiopia is characterized by limited market size which indicates that public investment in productive sectors may displace private ventures.Secondly, financial crowding out effect is occur if the increase in public investment is paid by borrowing on the domestic financial markets, which leads to greater occurrence of limiting of credit to the private sector for there is limited credit pool in Ethiopia (due to the government is heavily borrowing from domestic financial institutions).
The same findings were reported in Ethiopia like Melaku (2020), Kibret (2018), Agidew (2014), Sisay (2010) and Shiferaw (2016) and in other developing countries like Ago and Dollo (2020), Akçay and Karasoy (2020).Nevertheless, this finding is contradicting to the result that obtained by Ayeni and Nsiah (2020),, Adugna (2013), Clemens et al. (2019), stated that, public investment can  (Ndikumana & Verick, 2008).The effect of FDI in this study was negative and insignificant in short run which is opposing with the prior expectation.However, it was positive and significant at 1% level of significance in long run which is similar with the prior expectation.This result indicate that if all other explanatory variables are held constant, as FDI increases by 1%, domestic private investment also increases by 37.22% (0.372170) in the long-run (Table 10).
The result was in line with Legass et al. (2021), Sisay (2010), Al-Sadig (2013) found that FDI encourages domestic private investment which supports the crowd-in hypothesis, they found that the positive effects of FDI on domestic private investment in low-income countries depend on the availability of human capital.In Ethiopian, since 1992 market oriented economic reforms have been given to taken place and emphasis has been attracting FDI (Asante, 2000).However, the result was contradicting with Mbaye (2014) stated that as foreigner invested foreign capital in lowincome countries, the domestic investor couldn't be competing in financial investment.FDI highly compete with labor force because it provides high salary and other benefits such as health assurance and allowance for his workers.

Domestic credit to the private sector
In this study, the effect of domestic credit to the private sector on domestic private investment was positive and significant at 5% level of significance in long run which is similar with the prior expectation.However, it is insignificant in short run.The result of long run indicates that, if the influence of another explanatory variable constant, as level of domestic credit to the private sector was increased by 1%, domestic private investment also increases by 54.98% (0.549756) (Table 10).The implication of this is that as the availability of credit increases, people have access to finance for the required project and raises domestic private investment rates.The result was consistence with Aklilu (2021), Maluleke et al. (2022), Nwankwo and Allison (2021), stated that, domestic credit to the private sector and domestic private investment are significantly positive relation in long run.The result was contradicting with Mbaye (2014), stated that, credit to private sector is surprisingly negatively associated to domestic private investment.This is due to funds to the private sector don't go to finance new investments.In developing countries, due to rampant poverty; most people would borrow to finance for other matters like education, healthcare and basic necessities.

Real effective exchange rate
The Effect of Real Effective Exchange Rate (REER) in this result was negative and significant at 1% level of significance in both short and long run which is similar with the prior expectation.This result indicate that if all other explanatory variables are held constant, as REER increases by 1%, domestic private investment declines by 1.54% (−1.536551) (Table 9) and 1.81% (−1.809484) (Table 10) in the short and long run, respectively.As foreign currency (the value of dollar) increases, domestic private investment decreases.This implies that real devaluation of exchange rate affects private investment negatively through raising the real cost of imported goods.Ethiopia imports a large amount of good for investment; depreciation of the nation's currency leads to raise the price of these imported goods and creates adverse shocks on the supply of input items in the investment activities.The basic inputs of investment activities are founded appropriate mixture of machineries; equipment's used for construction purpose obtained from imported items.The finding was similar with Nainggolan and Daulay (2015) and Agidew (2014).However, the result was contradicted with Rao and Tolcha (2016) which shows that REER appreciation in the time of higher export capacity and influences the rate of domestic private investments positively.Having a weaker currency relative to the rest of the world can help boost exports.

Annual lending interest rate
The annual lending interest rate is the rate of interest per year an investor expects to receive for investment (Greene & Villanueva, 1990).It is determined by National Bank of Ethiopia.In this study, the effect of annual lending interest rate was negatively significant at 1% significance level in both short and long run which is similar with the prior expectation.This result indicate that if all other explanatory variables are held constant, as lending interest rate increases by 1%, domestic private investment declines by 76.02% (−0.760203) (Table 9) and 26.48% (−0.267475) (Table 10) in the short and long run, respectively.It implies that a high level of lending interest rates raises the cost of capital and thereby making investment less profitable and vice versa.This finding was consistent with the results of Akçay and Karasoy (2020), Afawubo and Mathey (2017).However, the finding contradicted with Shabbir et al. (2021), Legass et al. (2022), Atoyebi et al. (2012), Agu (2015) stated that the lending rate of interest has a positive relationship with the rate of private investment.For the reason that, in rising interest rates there could be the existence of joint venture business.

Unemployment rate
In this study, the effect of unemployment rate was positively significant at 1% significance level in short run which is reverses with the prior hypothesis.However, it is insignificant and positive relation in long run.The result of short run indicates that, if the influence of another explanatory variable constant, as level of an unemployment rate was increased by 1%, domestic private investment also increases by 22.10% (0.221016) (Table 9).This implies that, as unemployment rate increased the cost of input decreased because of working labor forces available cheaply (the investor paid cheap labor for production) that motivate investor to invest more.Since Ethiopia has high unemployment rate especially youth it encouraging the growth of domestic private investment.The result was similar with Ngoma et al. (2020), Waktole and Bogale (2018), Waktola (2020).However, the finding was contradict with Legass et al. (2022), Michaillat (2014), stated that, unemployment rate has negative effect on domestic private investment.Because, high rate of unemployment results an increase in social problems like crime and vandalism are more common in areas with high unemployment.If crime and vandalism are practiced in one country the private investors are fear to invest their resource in that country.

External debt service
To help low-income countries like Sub-Saharan African Countries particularly Ethiopia, the IMF and the World Bank created the Debt Sustainability Framework in 2005 to periodically assess the situation and provide recommendations to address any potential risks (World Bank, 2011).For economy of the debtor country to be sustainable, the World Bank recommended a maximum debt-service ratio of 10% for public debt.In this study, the external debt service was negatively affecting the domestic private investment and significant at 1% level of significance in long run which is equivalent with the prior expectation.However, it is insignificant in short run.The result of long run indicates that, if the influence of another explanatory variable constant, as level of external debt service was increased by 1%, domestic private investment decreases by 44.48% (−0.444782) (Table 10).This implies that external debt creates uncertainty in the macroeconomic environment and "crowding-out" effects for domestic private investment.As debt overhang explains large amount of debt eradicates the incentive for investors because returns from investors used for repay the existing debt and puts pressure on current and future tax burden on private investors (Ayeni & Nsiah, 2020;Agidew (2014).Uncertainly of time and amount of external debt transfers to the creditors as it be subject to future levels of world interest rates, the purchasing capacity of exports and the ability to reschedule the existed debt also have big impact on private investors.The result was consistence with Ayeni and Nsiah (2020), Agidew (2014), Kazeem (2022).Nevertheless, the finding was contradicts with Kibret (2018), Adugna (2013), Lawanson (2012), stated that, external debt service has significant positive effect on private investment.As long as it is used in productive investment it has favorable effect on the private investment in countries where there is serious shortage of finance.

Trade openness
Trade openness in Ethiopia can be classified in to export promotion and import substitution.Ethiopian government eradicated restriction on exporter to promote the export and have trade surplus (World Bank, 2015).In this study, the trade openness was positively affecting the domestic private investment and significant at 5% level of significance in long run which is equivalent with the prior hypothesis.However, it is insignificant in short run.The result of long run indicates that, if the influence of another explanatory variable constant, as level of trade openness was increased by 1%, domestic private investment also increases by 36.94% (0.369380) (Table 10).This is due to the fact that trade openness promotes technological progress, increases productivity and division of labor that can serve as a potential source of productivity, increasing in key markets, and rising competition of trade.The result was consistence with the previous scholars (Kibret, 2018, Ambe, 2017;Boachie et al. (2020), Sisay (2010); Kim et al. (2013); Ajide (2013); Adugna (2013), who found a positive and significant effect between trade openness on private investment.However, the finding was contradicts with Ouattara (2004) pointed that the impact of the trade openness on private investment is negative.Because of private investment in Senegal is highly sensitive to external shocks.

Real GDP
The growth and the level of the GDP provide an indication about the investment opportunities open to the economy.A broad measure of an economies size is its output.This output is basically and mostly the results of private investment or capital accumulation.In this study, the real GDP was positively affecting the domestic private investment and significant at 5% level of significance in long run which is equivalent with the prior expectation.However, real GDP is insignificant in short run.The result of long run indicates that, if the influence of another explanatory variable constant, as level of real GDP was increased by 1%, domestic private investment also increases by 13.44% (0.134372) (Table 10).The GDP shows an increase in sales and profits.The higher real GDP per capita is assumed increase effective demands for goods and services and thereby motivate domestic private investors.As income rises, capacity to manage resources to domestic saving rises and able to get more finance and begin new investment projects.Thus, it also creates consistent with expectations of neoclassical investment theory, positive association between private investment and income growth rate.Higher-income countries are more likely to put more of their money into domestic investments, which can then be used to help finance private investment (Mbaye, 2014).The result was consistence with the previous empirical results from Ethiopia (Legass et al., 2022;Ago & Dollo, 2020;Kibret, 2018, Ambe, 2017;Adugna, 2013), Ayeni and Nsiah (2020) from Gambia, Ajide (2013) from Nigeria, Ngoma and Berke (2020) evidence from Sub-Saharan Africa.

Conclusion and recommendation
The primary objective of the study was to investigate the trends and determinants of domestic private investment in Ethiopia by using a time series data from 1992 to 2022.The study used the ARDL bounds testing approach to examine the determinants of domestic private investment.The findings of short run demonstrate that inflation rate, public investment, real effective exchange rate and annual interest rate negatively influenced domestic private investment.Contrariwise, unemployment rate was positively influence domestic private investment while external debt service was insignificant.Foreign direct investment was insignificant and negative relationship with domestic private investment in short run.The results of long run shows that external debt service, inflation rate, public investment and real effective exchange rate were negatively affect domestic private investment.However, domestic credit to the private sector, foreign direct investment, real GDP and trade openness were positively affect domestic private investment.While total government expenditure, unemployment rate and lending interest rate were insignificant.The study generalized that the macroeconomic variables were highly affected the domestic private investment of Ethiopia in long run than short run.
Based on the findings the study recommended that since inflation has a negative relation with domestic private investment in short run and long run, policymakers should understand the cause for inflation volatility and keep in a stable manner.Inflation should be kept at a manageable level because with the uncertainties that it hurts the progress of domestic private investment.As trade openness was positively influence the domestic private investment, the government of Ethiopia should be encourage the relationship between international trade and domestic private investment to boost economic growth of the country.As a consequence of external debt service has negative impact on domestic private investment, the government of Ethiopia must first pursue debt restructuring of cancellation before their goal to promote industrialization through private investment can be achieved.This study focused on the determinants of domestic private investments in Ethiopia for only 31 years (192 to 2022).The author suggests that the future research can expand the study period (more than 31 years) and include other the rest macroeconomic, political and institutional variables.
Figure 1.Trends of domestic private investment in Ethiopia (1992 to 2022).Source: Author computation; from various years' report of EIC and MoFEC

Figure
Figure 3. Stability test result using CUSUM and CUSUM Square.Source: Own computation using Eviews 12.

Table 5 .
TheADF (1981)andPP (1988)statistic used is a negative number.The more negative indicates the stronger the rejection of the hypothesis.The result indicates that most variables such as domestic private investment (in both ADR & PP), annual inflation rate (in PP), government total expenditure (in PP), Foreign Direct Investment (in PP), public investment (in both ADR & PP), Real GDP (in both ADR & PP), interest rate (in both ADR & PP) and Trade openness (in both ADR & PP) were stationary at the level (Table

Table 5 )
. A variable is non-stationary if the estimated ADF and PP test is

Table 4 . Optimal lag length selection criterion
*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, HQ: Hannan-Quinn information criterion.

Table 9
provides the outcomes of the short-run error correction terms.The Error Correction Model (ECM) is a non-spurious regression model as indicated by the R-squared and Durbin-Watson statistics.It leads the variables of the estimation to restore back to equilibrium or it repairs disequilibrium.This happens when the sign of error correction (CointEq) is negative and significant.In this, the CointEq is

Table 7 . Diagnostic testing results for misspecification, auto-correlation and heteroskedasticity Test F-statistic & Observed R-squared Probability
Note: Figures in parenthesis representsF-statistic & observed R-squared and their P-values.

Table 10 . Results of estimated long-run coefficients ARDL(1, 1, 0, 1, 0, 1, 1, 1, 0, 1, 0, 1) selected based on Schwarz Criterion(SC) = Level Equation Dependent Variable: LNPINV
private investment in a situation where resources are not fully employed.Because public investment is acting as a crowding-in serve as a catalyst to private investment through external economies, like the provision of infrastructure such as transport (road access), communication and electric power. raise