Excerpt

## Table of contents

Acknowledgment

List of Tables

List of Figures

Acronyms

List of tables in the Appendix

Abstract

CHAPTER ONE: INTRODUCTION

1.1. Back ground of the Study

1.2. Statement of the Problem

1.3.Research Questions

1.4. Objective of the Study

1.5. Significance of the study

1.6. Limitation and Scope of the Study

CHAPTER TWO: RESEARCH METHODOLOGY

2.1. Research Design

2.2. Data Type and Source

2.3. Method of data analysis

2.3.1. Descriptive Analysis

2.3.2. Econometrics Analysis

2. 4. Definition and Measurement of Variables

2.5. Theoretical Framework for econometric model specification

2.6. Econometric Model specification

2.7. Time series properties

2.7.1. Stationary Tests

2.7.2. Phillips-Perron (PP) test

2.7.3. Cointegration test

2.7.4. Granger Causality Test

2.8. Estimation Techniques

2.8.1. Vector Auto-Regression (VAR) analysis

2.8.2. Impulse Response Analysis

2.8.3. Variance Decomposition

2.8.4. The Error-Correction Model

2.9. Chow Test for Structural Stability

CHAPTER THREE: RESULT AND DISCUSSION

3.1. Descriptive Analysis

3.1.1 Total Government expenditure and growth in Ethiopia

3.1.2. Composition of public expenditure in Ethiopia

3.1.3. Distribution of Government Expenditure

3.1.4. Sectorial composition of public expenditure in Ethiopia

3.1.5. Financing of Government Expenditure

3.1.6. Government budget deficit and means of financing

3.1.7. Trends of private Investment in Ethiopia

3.1.8. Investment during Derg period

3.1.9. Investment during post Derg period

3.1.10. Gross domestic savings and investment in Ethiopia

3.2 Econometrics Analysis

3.2.1Augmented Dicky-Fuller test

3.2.2. Phillips-Perron (PP) test

3.2.3 Cointegration Analysis

3.2.4 Coitegration (Long run estimation)

3.2.5 Granger-causality test

3.2.6 VAR Estimation and Diagnostic Tests

3.2.7 Regressions Estimation and Diagnostic Tests

3.2.8 The short run error correction result of private investment

3.3 The Effect of Government Expenditure on Private Investment

3.4. The Effect of Recurrent Expenditure on Private Investment

3.5. The Effect of Budget Deficit on Private Investment

3.6. The Effect of Domestic credit to private sector on Private Investment

3.7. Forecast Error Variance Decomposition Analysis (FEVD)

3.8. Impulse response Function

3.9. Chow test result

CHAPTER FOUR: SUMMARY, CONCLUSION AND POLICY IMPLICATION

4.1 Summary

4.2. Conclusion

4.3 Recommendation

4.4 Policy Implications

4.5 Areas for Further Research

Reference

## Acknowledgment

Anyone who has written a dissertation knows that you can’t get here on your own **!**

The first Gratitude goes to the Almighty God, who is helping me in every sector of my life.

Next to God, I am highly indebted to my advisor Mekonnen Bersisa (PhD candidate) without his encouragement, insight, guidance and professional expertise the completion of this thesis work would have not been possible. He encouraged me to explore possibilities, to seek answers with fervor and zeal. Our discussions stimulated the mind and led to new ideas. I learnt many aspects of the academia from him, and I have been privileged to have him as my major advisor. The encouragements and facilitations rendered by Abdi Teshome (MSC) who is my co –advisor were magnificent and I would like to extend my heartfelt thanks.

Last, but not least, I would like to thank my father Hailu Gobu and all individuals who relentlessly contributed to the completion of this research work in one way or another.

## List of Tables

Table 1. Average growth rate of real GDP, GDP per capita and Population, (%) 1974 to 2012

Table 2 Growth of Total Government Expenditure and GDP under the Two Regimes

Table 3.Capital and current expenditure (% of total expenditure) 1974 to 2012.

Table 4: Average share of various compositions of gov’t expenditure in total expenditure, 1975 to 2012.

Table 5 Sectorial classification of government expenditure as a percentage of total expenditure, 1974/75 to 2010/11

Table 6: Summary of budget deficit and means of financing (% of GDP), 1979 to 2012.

Table 7 Private, public and Total investment (% of GDP)

Table 8 Average gross capital formation and domestic saving (% of GDP), 1981 to 2012.

Table 9 Unit Roots Tastes Results at Level

Table 10 Unit Roots Tastes Results after differencing

Table 11: Phillips –perron (pp) unit root test at Level

Table 12: Phillips –perron (pp) unit root test at difference

Table 13: Johansens’ Cointegration test using Maximum Eigen value& the Trace statistic

Table 14: Normalized long run beta coefficient for private investment

Table 15 Adjustment (α) Coefficient for private investment

Table 16: Test result of weak exogenity

Table 17 Result of zero restriction tests on *B* coefficient for private investment

Table 18 Long run relationship regression result

Table 19 Error correction Model Results

Table 20: variance decomposition of private investment

## List of Figures

Figure 1 Average growth rate of real GDP, GDP per capita and Population, (% of GDP) 1974 to 2012

Figure 2 Trends in nominal GDP and Total government spending

Figure 3. Total government expenditure (% of GDP) and real GDP annual growth 1974 to 2012

Figure 4. Total government expenditure (% of GDP), 1974 – 2012

Figure 5.Real GDP Growth rate, 1974 to 2012

Figure 6 Revenue, Expenditure and Deficit share of GDP from 1979-2012

Figure 7 Trends in Investment (In Levels) in million birr

Figure 8 Trends in Investment (% of GDP)

Figure 9 shows the effect of a one standard deviation shock on recurrent government expenditure on private investment

Figure 10 shows the effect of a one standard deviation shock on budget deficit on private investment

Figure 11: shows the effect of a one standard deviation shock of Domestic credit to private sector on private investment

## Acronyms

illustration not visible in this excerpt

## List of tables in the Appendix

Appendix A1: Result of granger causality test

Appendix A2: VAR diagnostic tests

Appendix A3: VAR estimation result

Appendix A4: variance decomposition result

## Abstract

*This study attempts to investigate the effect of government expenditure on private investment in Ethiopia over the period 1980-2012. The central question of this study is weather government expenditure has a positive or crowding in effect (complementary hypothesis) or a negative or crowding out effect (the substitutability hypothesis )on private investment in Ethiopia. To achieve its objective it adopted a modified flexible accelerator model to enlighten on the economic relationship between private investment and the other variables and used the modern technique of vector auto regressive model (VAR) and vector error correction model(VECM)as its methodology. The study also used the Johansen-Juselius (1990) cointegration analysis of a multivariate system of equation to estimate the long run relationship between government expenditure and private investment to determine the order of integration of the variable and Granger-Causality test was undertaken to determine causal relationship between the variables. In addition to this the study employs the Augmented Dicky-Fuller (ADF) unit root test and phillip perron test. The statistical tests reveal that all-time series data are non-stationary in their level and they become stationary after diffrencing.i.e.they are integrated of order one I(1).The johansen-juselius cointegration test shows that the series are cointegrated and then employs the vector error correction model moreover the study applies the impulse response function (IRF)and forecast error variance decomposition (FEVD) to investigate the effect of government investment shocks on private investment. And the empirical findings support the complementary hypothesis between government capital expenditure and private investment and that tends to crowd-in private investment in Ethiopia. And the empirical finding of recurrent part of government expenditure shows a mixed effect of complementary hypothesis and substitutability hypothesis which tends to crowd-in and crowd out effect .Thus government expenditure have a positive as well as negative effect on private investment and finally the study is used CHOW test in order to know whether structural break has an effect on private investment or not and the result depict that there is a structural break that have a positive effect on private investment of Ethiopia.*

Keyword: *Government expenditure, private investment, VAR, crowding-In, crowding out, Ethiopia*

## CHAPTER ONE: INTRODUCTION

### 1.1. Back ground of the Study

The interest of economists in the relationship between government spending and private investment is motivated mainly by the controversy over the crowding out or crowding in effect of government spending on private investment. With the renewed interest in the role of the private sector as an engine of economic growth, the examination of this relationship is given further impetus. The idea of a private sector led economic growth in Ethiopia is therefore traceable to the observed success of the major industrialized countries; which attributed to the resilience of their organized private sector (Nasir & Muhammad, 2009).

As a result of the poor performance of the economy over the period in which government played the leading role in the economy, there was a change in the expected role of the government. To this end, market oriented structural reform programs such as privatization and deregulation were adopted to ensure a reduction in the role of government in the economy. The guiding principle in this redefined role of government was that government should concentrate its resources in areas that compliments rather than crowd-out private sector investment, thereby creating an enabling environment for the private sector investment (Nasir & Muhammad, 2009).

Some economists have argued that increase in government spending can be an effective tool to stimulate aggregate demand for a stagnant economy and to bring about crowed-in effects on private sector. According to Keynesian view, government could reverse economic downturns by borrowing money from the private sector and then returning the money to the private sector through various spending programs. High levels of government consumption are likely to increase employment, profitability and investment via multiplier effects on aggregate demand. Thus, government expenditure, even a recurrent nature, can contribute particularly to private investment and economic growth. On the other hand, endogenous growth models like Barro predict that only those productive government expenditures will positively affect the long run growth rate(Barro, 1990).

In the neoclassical growth model of Solow productive government expenditure may affect the incentive to invest in human or physical capital, but in the long-run this affects only the equilibrium factor ratios, not the growth rate, although in general there will be transitional growth effects. Others have argued that increase in government expenditures may not have its intended salutary effect in developing countries, given their high and often unstable levels of public debt. Thus government consumption crowds out private investment, dampens economic stimulus in short run and reduces capital accumulation in the long run(Solow, 1956).

Further Vedder and Gallaway argued that as government expenditures grow incessantly, the law of diminishing returns begins operating and beyond some point further increase in government expenditures contributes to economic decline and stagnation(Gallaway & Vedder, 1998).

The Ethiopian economy is a mixed system in which the government and the private sector co-exist. The two could play complimentary roles to enhance economic growth. Thus, it is in line with this that the use of government expenditure to enhance private investment is being advocated.

To address the inefficiencies in public expenditure management in Ethiopia, the federal government introduced wide range of policies and institutional reforms, geared towards privatizing the economy, particularly since 1992 when the structural adjustment program (SAP) was implemented. The Transitional Government of Ethiopia, which was established after the downfall of the military regime in May 1991, initiated a new market-driven economic policy followed by a comprehensive structural and economic reform program. One of the major objectives of the reform program was to rectify the fiscal ills & attain a consolidated government budget. This objective called for rationalizing the state’s role in the economy, implying reorientation of government expenditure, and at the same time enhancing revenue performance with the support of International Monetary Fund and the World Bank as well as other multilateral and bilateral donors(MoFED, 2010).

Notwithstanding the developments stated above, total expenditure during the period 1991/92 to 1997/98 increased at an annual average rate of 16.7 percent per annum. Though both recurrent & capital expenditure have been rising, capital spending accounted for a larger part of the increase in total spending. Similarly recurrent expenditure as a ratio to GDP has declined during the post-reform period while that of capital expenditure increased from their pre-reform levels. Much of the increased in capital expenditure was accounted for by spending on roads, energy, education and health sectors. On the recurrent side, wages and operating expenses and debt servicing took the lion’s share during the period under review(MEDaC, 1999)

Recent developments also indicate a trend of increasing government expenditure, especially expenditure on priority sectors of education, health and roads. Total government expenditure as share of GDP has increased to 33.5 percent annual average in 2001/02-2004/05 from 28.8 percent during 1991/92-1995/96 periods. And it also emphasized on the evolution of a private sector led market oriented economy with competition as a driving force. The key elements of this strategy include privatization, deregulation, and reducing the influence and involvement of government in the economye (1974-1991), Ethiopia’s private sector hardly existed. The socialist regime provided little rooms for the private sector to flourish until the period it vanished in 1991. Following the fall of the Derg regime in 1991, a shift in policy from command economy to free market economy was introduced in the country. This shift has opened opportunities for the private sector to have an active role in various sectors of the country’s economy. Since then, a lot of efforts have been exerted and various forms of incentive packages have been provided by the government to encourage domestic private investment in the country. Despite the incentives taken by the government, the sector’s contribution towards the economy of the country has remained very poor by international standards, even when compared with Sub-Saharan countries (World Bank., 2010)

Similarly the trend of domestic private investment as a percentage of GDP is a good evidence of how low the sector’s contribution to the economy is. For instance, from 1992-2000 and 2001-2010 domestic private investment as a percentage of GDP were 2.6 and 1.2 respectively. Particularly, in the last five years domestic investment has been reduced though the country’s economy was growing continuously. For instance, from 2006-2010 the average domestic investment as percentage of GDP was only 0.5% while average economic growth for the same period was about 11%. Similarly while the country’s domestic private investment to GDP is low, the resource gap between savings and domestic investment is very high. For instance, the resource gap between savings and investment in 2009/10 was 19.4 % which is very high in comparison to the international standard (MoFED, 2010).

Understanding this problem, the government has prepared a strategy document namely Growth and Transformation Plan (GTP) in 2010 that addresses many issues of the country including domestic investment and its constraints as one agenda (FDRE, 2010). Low per capita income of citizens, limited saving behavior, poor and limited financial institutions and lack of infrastructure are some of the factors identified by the government in its GTP as bottlenecks to the country’s domestic private investment(MoFED, 2010).

However, as Ascharer (1989) noted, the precise effect of government expenditure on private investment depends on the type of government expenditure being considered. To the best of the knowledge of the researcher, there is paucity of literature conducted particularly on Ethiopia. Certain categories of government expenditure crowd out private investment while others complement or crowd-in private investment. Amidst the prevailing contradicting view; it is unquestionable to conduct an empirical study to understand the effects of government expenditures on private investment (Aschauer, 1989).

### 1.2. Statement of the Problem

Undertaking macroeconomic policy, various development strategies and economic policies introduced by the government so as to increase private investment did not bring a desired result. As it is well established in the literatures the failure of the government to achieve rapid and sustained private investment in Ethiopia spurred the debate on whether the government or the private sector should spearhead the nation’s economic growth process. To reconcile this issue, in the last five years the government dominated the economic activities of the country by changing the expenditure level. For example; growth in recurrent expenditure registered an average of 13.8 percent per annum over the past seven years (2003/04-2009/10) while the growth rate in capital expenditure averaged 30 percent per annum during the same period (MoFED, 2011).

Consequently, the share of capital expenditure to total expenditure increased from 40 percent in the fiscal year 2003/04 to 55 percent in 2009/10. While the share of recurrent expenditure to total expenditure declined from 58 percent in 2003/04 to 45 percent in2009/10. Even if the government undertakes such tremendous increases and decrease in its expenditure components its contribution to increase private investment has been remained in significant. Thus private investment has been persistently low in Ethiopia; For instance, from 1992-2000 and 2001-2010 private investment as a percentage of GDP were recording 2.6 and 1.2 respectively. It was identified that this low performance of private investment is a factor responsible for the lowest share of private investment as a percentage of GDP. For instance, from 2006-2010 the average share of private investment as percentage of GDP was only 0.5% while average economic growth for the same period was about 11% (MoFED, 2011)

Standing with the above juncture, private sector operators argued that the factors which militate against their contributions to the economy include high cost of doing business, unstable macroeconomic policies, infrastructural bottlenecks, faltering consumer spending, and lack of capital investment and stifling effect of multiplicity of taxes. Hence low productivity/un competitiveness of the private sector is therefore as a result of the hostile business environment(Nasir & Muhammad, 2009).

Further more political instability and social unrest can also be taken as potential sources of uncertainty and may even be highly damaging factors of the private investment environment. According to Seyoum (2002), there are two important mechanisms through which socio-political factors could influence private investment in developing countries. The first relates to extreme cases of instability that lead to changes in the rules of the game and threatens investors of possible confiscations. The other and perhaps the most common one relates to the unpredictability of the political environment say due to repeated changes of government or officials of key government institutions which undermines the responsiveness of private investors to economic incentives or reform measures. The basic idea is that investors may not regard government policy under political instability as credible, hence creating a value for waiting. This indirectly weakens the effectiveness of fiscal policy measures taken by government bodies to improve the share of private investment in the economy.

As to the researcher knowledge, researches conducted on the effects of government expenditures on private investment didn’t exist in Ethiopia. But there are other limited related researches. For example the research that was undertaken by Zerfu. (2001) assesses the Macroeconomic determinants of private investment in Ethiopia using time series data for the period 1965-1999. The result from this study indicated that GDP, public investment on infrastructure and foreign exchange availability all have positive effects on private investment. On the contrary, inflation rate and external debt to GDP ratio showed a negative effect on private investment. Another research which was conducted by Getnet (1992) on the determinants of private investment in Ethiopia using annual data for the period 1970-1989 identified a negative relationship between public investment and private investment. This result demonstrated a crowding out of public investment on private sector activities. However, on estimating a reduced form of the investment equation, public investment turned out to be insignificant while real Gross National Product (RGNP) growth was seen to have a negative effect on private investment. This result may not seem plausible, as it is inconsistence with the theoretical economic a priori.

In addition Mitiku (1996) carried out a study using survey and time series data for the period 1975- 1994, was also another attempt to investigate the determinants and constraints of private investment in Ethiopia, The results based on the time series data indicated that private investment in Ethiopia is determined by the availability of finance, the real exchange rate, investment policy (private investment policy), debt-service payment and debt-overhang. On the contrary, real interest rate, growth of GDP per capita, public investment and changes in terms of trade did not affect private investment during the period of study.

Without prejudicing the importance attached to the mobilization of resources, current circumstances obliged the proper allocation and efficient utilization of government expenditure as the reward is greater. Likewise, the penalty for bad policy in this respect is greater than ever before in the realm of globalization. In a nutshell, government expenditure could adversely affect domestic private investment if its allocation and utilization is not properly addressed. However, there is also lack of unanimity in the empirical findings as to the crowding-in/ crowding-out effects of Government expenditure on private investment. Research outputs often conflict with one another on the issue. For instance, while some have identified positive effects of government investment spending on private investment (Greene & Villanueva, 1991 and Ghura & Goodwin 2000), others such as Balassa (1988) cited in Bashier Al-Abdulrazag (2009), have found the other way round(Bashier, 2009).

As argument noticeably asserted in the above paragraph and the literatures of different researches undertaken in other countries on this area forwards an implication as understanding the underlying relationship between government expenditures and private investment is essential for devising and implementing policies. Coincident with the above fact, the distinct feature of this study is that it tried to investigate effects of government expenditures on private investment in Ethiopia; which is expected to bridge the existing gap in the area of the study.

### 1.3 Research Questions

This study sought to address the following research questions, that is, investigate the effects of government expenditures on private investment.

- What is the trend of government expenditure and private investment looks like over the study period? Does liberalization policy boosts private investment?

- What are the variables affecting private investment?

- What is the relationship between categories of government expenditure and private investment? Does components of government expenditure crowd in or crowd out private investment in Ethiopia?

- Which type of government expenditure complement private investment and which has crowding out effects?

### 1.4. Objective of the Study

The general objective of the study is to determine the effect of government expenditures on private investment in Ethiopia.

The specific objectives of the study are`

- To investigate the causal relationship between disaggregate government expenditure and other variables on private investment in Ethiopia.

- Analyze the relative effects of various components of government expenditures on private investment in Ethiopia (crowding in and crowding out)

- Analyze the Effect of Budget Deficit on Private Investment in Ethiopia.

- Determine the effects of government domestic credit to the private investment in Ethiopia.

- To review and analyze the trend of government expenditure and private investment in Ethiopia under the two regimes

(The Derg and The post Derg)

This study put tentatively the following hypothesis in line with the objectives stated above.

Hypothesis 1: The recurrent government expenditure and private investment will have a negative relationship.

Hypothesis 2: There will be positive relationship between capital government expenditure and private investment

Hypothesis 3: There will be a negative effect of inflation on private investment.

Hypothesis 4: There will be a positive relationship between economic growth and private investment.

Hypothesis 5: There will be a negative relationship between budget deficit and private investment.

Hypothesis 6: Structural break before reform program is expected to have negative relationship with private investment while the after reform program will be expected to have positive relationship with private investment.

### 1.5. Significance of the study

Albeit it’s significance for policy makers and to the best of my knowledge there is no research that was undertaken on the effect of government expenditure on private investment in Ethiopia. In this respect, existing literatures done in different countries have different result and conclusion for the effect of government expenditure on private investment. Hence, such work can usually be conducted in the context of country specific to capture the effects of government expenditure on private investment.

One more advantage of this study is that it incorporates the most recent data and employs quantitative analysis and a more advanced econometric technique (Johansen approach to cointegration) to study the effect of government expenditure on private investment. Thus, the immediate outcome of this study will be to provide pertinent result and policy implication to policy makers by bridging the aforementioned gap. Besides, the researcher believe that the study will provoke and pave a way for further study in the area as it reveals the difficulty in resolving the empirical question of the effect of government spending on private investment.

### 1.6. Limitation and Scope of the Study

One limitation of this study arises from lack of clear agreement on which categories of government expenditure can affect domestic private investment (or how to measure its constituents). Economists are not yet certain about the relative importance of each component which clearly or perhaps not so clearly influences private investment. Without such knowledge the area of disturbing doubt is uncomfortably large. Besides, the definition of particular expenditure as productive or unproductive is open to debate.

The other limitation of the study was that it does not explicitly consider the quality of government spending, which was probably the most important factor. The caliber of the civil service and the military and the conditions in which they function have impact on creative and efficient use of public expenditures. Unproductive public spending can take various forms, including spending on wages and salaries of unproductive or ghost workers. Public spending is also unproductive when government expenditures do not reach designated spending objectives. This happens, for example, when government officials are corrupt and seek bribes for preferentially selecting beneficiaries of government programs, for authorizing private investment projects, for allowing participation of government enterprises in joint ventures with private investors, or for allowing access to inputs provided through state enterprises.

Apart from these, the econometric model does not consider the incidence of revenue structure and financing issues. It is believed that when government expenditure is not covered by own revenue, the nature of expenditure financing has a bearing on economic growth. Moreover, the nature of own revenue itself (whether it is distortionary or not) has a role in explaining the effect of government expenditure on domestic private investment.

The scope of this study is limited to an empirical analysis of the effects of government expenditure on private investment between 1980 and 2012. The major focus is on the effects of government expenditure on private investment. The choice of this study period is based on the availability of data.

## CHAPTER TWO: RESEARCH METHODOLOGY

### 2.1. Research Design

This study aimed at analyzing the effect of government expenditure and other variables on private investment in Ethiopia. Quantitative data was used in the study to answer the research questions. The study used the time series data of 33 years for the period 1980 to 2012 the collected data were analyzed using Vector auto **-** regressive (VAR) modeling technique and the error **-** correction model was generated after undergoing time series properties tests.

### 2.2. Data Type and Source

This study purely depends on secondary data. Different data sets were collected from various sources such as Central Bank of Ethiopia (CBE); Ministry of Finance and Economic Development (MoFED); Ethiopian Economic Association (EEA).These three agencies served as the main sources of data for federal government capital and recurrent expenditures and their various sub-components. Private investment data and other explanatory variables data were sourced from World Bank (WB) development indicators.

### 2.3. Method of data analysis

#### 2.3.1. Descriptive Analysis

In this study, to analyze the effect of government expenditure on private investment descriptive analysis was employed as starting point of analysis part. Here the theoretical relationships and empirical analysis of the variables and private investment were dicussed.In order to analyze the effect of government expenditure on private investment and to look at their trend different tools like tabulations, percentage and graphs were employed to make an inference.

#### 2.3.2. Econometrics Analysis

In addition to descriptive analysis, econometric analysis was used for analyzing data. These help to capture the degree of influence and significant effects of government expenditure and other variables on private investment. In doing so the government expenditure and other macroeconomic variables that are supposed to affect private investment are used as independent variable and private investment as dependent variable. The collected data were analyzed using Vector auto **-** regressive (VAR) modeling technique and the error **-** correction model was generated after undergoing different time series properties tests.

### 2. 4. Definition and Measurement of Variables

**Private investment (PI)**: (i.e. Gross domestic investment) is the total change in the value of fixed assets plus change in stocks. Gross capital formation is used as proxy to private investment. The researcher used domestic private investment as dependent variable and other explanatory variables. The explanatory variables that may affect the decision making of domestic private investment in the literatures are very wide and only variables having sound theoretical explanations and complete data will be selected. In this section the researcher attempt to describe the theoretical explanations and empirical evidences of the explanatory variables selected for this study.

**Government capital Expenditure (GCE):** It is the government expenditure on capital overheads. It was measured by the total government expenditure less recurrent expenditure.

**Government recurrent Expenditure (GRE):** It is the current expenditure for purchase of goods and services at all levels of government. It encompasses purchases of materials, office supplies, fuel and lighting, salaries and wages, travel services and payment of rent. It was measured by recurrent expenditure on labour costs and other goods and services.

**Inflation rate (INF):** inflation results from the macroeconomic effect of government spending and is the fourth variable that the researcher was used in the study as a proxy to measure macroeconomic stability of the country. There is no uniformity on the theoretical explanation of the variable and its effect on domestic private investment. Some models such as the cash-in-advance models (e.g. Stockman, 1981) forwarded that inflation raises the cost of acquiring capital which then lowers capital accumulation. This model further states that the existence of high inflation may make it difficult and costly for economic agents to extort the right relative price which could then lead to misallocation and inefficient resources. However, other models like the Tobin-Mundell model argues that higher anticipated inflation lowers the real interest rate which then causes to be made portfolio adjustments away from real money balances to real capital which then expected higher inflation to raise real investment (Ghura & Goodwin, 2000) Empirical studies such as (Bakare A. , 2011) and (Léonce, 2000) reported that inflation has a negative effect to private investment.

**Economic growth (GDP)**: The gross domestic product GDP is used as proxy for economic growth. This indicates the level of output in the economy. Its rate of growth is therefore an indication of the rate of growth of the economy and it is one of the most commonly variable used as explanatory variable to measure its effect on domestic private investment. Some literatures such as Fielding, (1997), Greene and Villanueva (1991) explained that private investment is positively related with real GDP growth of one country.

This is because countries with higher income level inclined to allocate more of their wealth to domestic savings which could be then used to help in financing private investment. Empirical results such as Ajide and Lawanson (2012) from Nigeria, Outtara (2004) from Senegal and Asante (2000) from Gahana have evidenced that real GDP growth rate helps domestic private investment. Ghura and Goodwin (2000) on their part revealed that real GDP growth has stimulating effect on private investment in Asia and Latin America though its effect in Sub_ Sahara Africa was found insignificant. But (Ndikumana & Sher, 2008) found the positive and significant relationship in Sub-Sahara Africa (SSA) which contradicts the findings of (Ghura & Goodwin, 2000).

**Domestic Credit to the Private Sector(DCPS) ;** Domestic credit to the private sector refers to financial resources provided to the private sector, such as through loans, purchases of non-equity securities and trade credits and other accounts receivable, that establish a claim for repayment.

Domestic credit thus represents the investible resources devoted to productive activities in the economy. Credit provision to the private sector is a critical determinant of the quantity as well as the quality of private investments that are undertaken. In a situation where policies are implemented that facilitate increased private sector participation such as increased financial sector liberalization, these should have a positive effect on domestic credit to the private sector, leading in turn to greater investment and in turn economic growth. Domestic credit is widely used in the literature including by Badawi., (2003), (Boopen & Khadarro., 2007)

**Public Investment(PUI):** It comprises all additions to the stocks of fixed assets (including purchases and own-account capital formation), land improvements fences, ditches, drains etc.-plant, machinery and equipment purchases (including imports), construction of roads, railways, schools, hospitals, office buildings, less any sales of second-hand and scrapped fixed assets, by government units and non-financial public enterprises. Outlays by government on military equipment are excluded. Fixed Investment by the public sector is expected to have a positive effect on real output. In Neo classical growth theory, increases in the level of public capital stock (public investment) are argued to lead to increases in the capital stock and this in turn should lead to increases in real output. This effect however only occurs if the public investment is coupled with increased private investment or if the reductions in real output due to any reductions in private investment are offset by the increase in output due to an increase in public investment. Public sector gross fixed capital formation is used in the literature as a measure of public investment by among others (Badawi., 2003),

**Public Consumption (PUC);** It comprises government purchases of goods and services, including office supplies and maintenance charges, wages and salaries of employees and expenditures on national defense. Public consumption leads to an increase in real output in the economy. This is particularly the case when it is used to encourage private investment to meet the additional demand, by way of supply of goods and services purchased by the government. However if public consumption is financed by borrowing from the domestic economy it can reduce private investment. Similarly, if say for political economy reasons, public consumption is encouraged at the expense of public investment, it can have potentially deleterious effects on real output. Public consumption has been used in the literature by (Nazima & Adiqa, 2011).

**Budget Deficit (BD ):** It refers to the total government revenue less total expenditure. It measures that portion of expenditure which is financed through borrowing and by printing of notes and coins. It was measured by deducting total expenditure (re- current and capital) from total revenue (tax and non **-** tax).

**Structural Break (SAP):** This variable will be included as a dummy variable introduction of the structural adjustment programme in 1992 had as its major policy objective that is a reduction in government participation in the economy, while at the same time giving priority to the private sector to lead economic growth process. This called for a substantial reduction in government expenditure and thus a structural break in the economy. The dummy variable D, in our model captures the effect of structural break as a result of government partial disengagement from the economy.

### 2.5. Theoretical Framework for econometric model specification

The theoretical literature on the investment maintains that government expenditure can either crowd **-** in or crowd **-** out private investment depending on how this policy is designed and implemented (Keynes, 1936). The analytical framework underlying this position is fashioned in this study in line with the variants of the flexible accelerator investment macro-economic theory designed to investigate the quantitative effect of disaggregate government spending along with a set of variables on private investment and variant of the neoclassical flexible accelerator model discussed by (Ramirez, 1994) and, (Charles & M.Veeman., 1996) in order to know the casual relationship between Government expenditure and private investment.

To derive the theoretical framework for this study, the partial adjustment model is expressed as:

Where is desired or optimal investment at a time , t is actual investment at time t,and is the partial adjustment coefficient reflecting the assumption that the rate at which firms move from actual level of investment to the desired or optimal level is gradual involving lags. Equation (1) implies that the change in investment is a partial adjustment to the gap between the desired and actual investment. Regarding production function, the study used a Domar type production function that relates output to the stock of capital.

The desired capital stock was assumed to be proportional to the expected output as given in equation (2). Where represent input **-** output ratio which is assumed to be constant. This specification assumes a fixed factor proportions production function. This is justified on the grounds of existence of surplus labour in Ethiopia and thus production is constrained by the size of capital stock. Suppose that because it takes time to build, plan and install new equipment, the actual stock of capital adjust to the difference between the desired capital stock at the current period and the actual stock in the previous period. The actual private capital stock in period t can be expressed as:

Where is the current period’s capital stock, is the previous period’s capital stock and is the rate of depreciation. Equation (3) represents the simplest version of the flexible accelerator model. To see its implication, it can be noted that by definition, gross private investment at a time is given as:

By introducing the lag operator *L* given as and assuming steady state situation, equation (4) becomes; Where is the desired investment at time. By combining equation (2) and (5) the result is expressed as: Substituting equation (6) into (1), a dynamic flexible accelerator model is obtained as given by equation (7) The dynamic flexible accelerator model imply that the rate at which firms move from actual level of investment to the desired or optimal level is gradual involving lags, and that the variation in investment depends on output. This study adopted the approach used by(Blejer & kahn, 1984), which allows private investment to vary with economic conditions. The study further formed a hypothesis that the size of partial adjustment coefficient, depends on both fiscal and non **-** fiscal variables. This stipulation can be formally expressed as; Where is a vector of fiscal variables is a vector of non – fiscal variables Fiscal and non **-** fiscal variables are expressed in relation to the size of the discrepancy between actual and desired level of investment. In this specification, the above hypothesized factors affect investment through the process of adjustment from actual investment towards desired levels. In a linear form, equation (8) can be represented as Where is the intercept and are the coefficients of fiscal and non **-** fiscal variables, respectively, while is a white noise error term by substituting equation (9) into (1) and solving for, the obtained equation is expressed as:

In the equation (10), the desired investment is not observable. By substitution of equation (6) into equation (10) the result is expressed as:

Equation (11) imply that investment is not only explained by changes in output as proposed by the classical flexible accelerator model Shapiro (1992), but also by a vector of both fiscal and non **-** fiscal variables. In the spirit of Blejer & kahn (1984), these variables are found to be affecting the adjustment process from actual to the desired or optimal level of investment.

### 2.6. Econometric Model specification

The model to be estimated in this study is derived following Blejer & kahn (1984) and Aschauer (1989a), and Ekpo (1996) by disaggregating government expenditure into its various components and examined their separate effects on private sector investment. The disaggregate and aggregate effect of fiscal and non-fiscal variables was therefore captured through the variations in output. Structural reform has influence on private investment. In Ethiopia, structural reform was taken place in 1992 a vector of dummy variable were included in the model to capture the effects of these reforms on private investment.

The general form of the model is as given in equation (12)

Adopting this pattern therefore, the present study specifies the following models. This can be estimated A log Transformation of all variables is used in order to standardize the variables. All variable series are in logarithmic form. This is because the series expressed in logarithmic form have roughly constant variances, while the variances of a level series tend to increase with the sample size. The functional relationship into logarithmic form takes the following form:

Where is the private investment Gross domestic product government capital expenditure government recurrent expenditure inflation BD budget deficit Domestic credit to private sectors Public Investment public consumption budget deficit and dummy variable for structural adjustment program and μt is Error term encompassing all other factors determining private investment but not captured in the model.

However, the investment model given in equation (12) imposes stringent condition that investment is the dependent variable while the variables in the right hand side of equation the (12) are independent. This condition is introduced in an adhoc manner since economic theory does not provide adequate evidence on the granger causality between these variables. Investment could be impacted by changes in these variables and on the other hand its variations could also affect these variables (Sims C. , 1972) argued that the division of variable into endogenous and exogenous variables, as done in the structural models, is arbitrary and that VAR models could avoid that by treating all variables as endogenous. It is further asserted that in the VAR model, cross variable effects are automatically included as each variable is regressed on its own lagged value and lagged values of all other variables. Sims (1972) Modeling philosophy was adopted in this study and the modified flexible accelerator model only played the role of identifying variables of interest.

### 2.7. Time series properties

#### 2.7.1. Stationary Tests

Time series analysis was central to empirical modeling of the effects of government spending on the private investment. The non **-** random behavior of the time series data could undermine the usefulness of the standard econometrics methods if it was applied directly without considering time series properties of the data(Russell, 1993) . To test for stationarity in the variables used in the study, the formal statistical tests for the presence of a unit root were undertaken. The two main methods which were applied are Augmented Dickey **-** Fuller (ADF) and Philips Perron (PP) tests as explained (Dickey & Fuller, 1981) and (Philips and Perron, 1988) respectively. This was due to the fact that, the data generating process was not an AR (1) process. Therefore, the ADF was correctly specified in the higher **-** order case (Engle & Granger, 1987). The ADF procedure attempts to retain the validity of the tests based on white – noise errors in the regression model by ensuring that the errors are indeed white- noise. On the other hand, Philips Perron (PP) procedures corrects the problem of serial correlation through a parametric correction to the standard statistic(Stock J. , 1994)

(i) ADF without intercept and trend

(ii) ADF with an intercept but no trend

(iii) ADF with both the intercept and trend

Where is the variable in question

is time trend

is the lag length and

is the error term assumed to be white noise

The ADF tests the null hypothesis of a unit root/non stationarity ) against an alternative hypothesis that no unit root is present/stationary in the autoregressive equations:

The calculated t-statistic of the estimated is compared with Mackinnon critical values. If the absolute value of the calculated t statistic is greater than the critical value, then the null hypothesis of a unit root (non-stationarity) is rejected and we conclude that the series is Stationary. In this case the level of the time series is said to be integrated of order zero I (0).

#### 2.7.2. Phillips-Perron (PP) test

The PP test differs from the ADF test in that it does not assume white noise residuals, but corrects the problem serial correlation in the residuals. The test uses a non-parametric method to account for serial correlation in the residuals. It does not augment the Dickey-Fuller test equation when accounting for serial correlation but it instead adjusts the test ( statistic to account for serial correlation. The modified statistic of the PP test however follows the same distribution as the ADF statistic.

As with the ADF test, in the PP test, the null hypothesis of a unit root/non stationary ) is tested against the alternative hypothesis that no unit root is present/stationarity

The calculated t-statistic of the estimated is compared with its respective critical value. If the absolute value of the calculated t statistic is greater than the critical value, then the null hypothesis of a unit root (non-stationarity) is rejected and we conclude that the series is stationary. In this case the level of the time series is said to be integrated of order zero I (0).

#### 2.7.3. Cointegration test

Cointegration refers to the existence of a long-run equilibrium relationship between variables. The idea of long-run equilibrium implies that two or more variables may wander away from each other in the short-run but move together in the long-run (Enders W. , 1995) . The use of cointegration technique allowed the study to capture the (Adam, 1998) equilibrium relationship between non–stationary series within a stationary model following Adam (1998). It permitted the combination of the long-run and short-run information in the same model and overcame the problem of losing information which could have occurred when attempting to address non stationary series through differencing (Adam, 1998) Cointegration technique made it possible to capture the information of non-stationary series without sacrificing the statistical validity of the estimated equation(Stock & Watson, 1998)

Two main tests for cointegration, namely Johansen cointegration test and the Granger two **-** step methods were used. Johansen‟s methodology, which was expressed as a vector auto regression (VAR) of order P is given b Where is nx1vector of innovations This VAR can be re-written as Where an If the coefficient matrix reduced rank r<n then there exist n x r matrices and each with rank r, such that and is stationary. r is the number of coinegrating relationship.

Under the null hypothesis matrix can be decomposed in to a product of two non-null matrixes such that matrices and hold adjustment coefficients and long run parameter respectively and are both nxr. The elements of are known as the adjustment parameters in the vector correction model, and each column of is a cointegrating vector. It has been shown that for a given *r,* the maximum likelihood estimator of defined the combination of that yielded the *r* largest canonical correlations of with after correcting for lagged differences and deterministic variables(Johansen, 1995).Johansen proposed two different likelihood ratio tests of the significance of these canonical correlations and thereby the reduced rank of the matrix. The trace test and maximum Eigen value test are shown in equation (21) and (22), respectively.

The Johansen procedure uses the Trace, and Maximum Eigen value, test statistics respectively Where T is the sample size and is the ith largest canonical correlation. The trace test tested the null hypothesis of *r* cointegrating vectors against the alternative hypothesis of *n* cointegrating vectors. The maximum Eigen value test, on the other hand, tested the null hypothesis of *r* contegrating vectors against the alternative hypothesis of *r* +1cointegrating vectors.

The residual based cointegration test introduced by Engle & Granger (1987) by analogy of equation (23) involves testing the significance of the coefficient in the Ordinary Least Squares (OLS) regression of: Where is the residual. The test postulates that if the residuals from the OLS estimation of the non-stationary variables are stationary, then the series are cointegrated. If the residuals exhibited a stationary trend, it implies that the error **-** correction model (ECM) could not be run. Instead, estimation could be done on the variables at their first difference. However, the long-run characteristics of the data would be lost. Therefore, the study used the Johansen cointegration method to test for the long-run relationship between the variables.

#### 2.7.4. Granger Causality Test

Given two variables and is said to granger-cause , if lagged values of help in the prediction of , or if the coefficients on the lagged values of ,are statistically significant in the equation of .

The test equation would thus take the form;

The null hypothesis is that and its rejection implies that can be said to Granger-cause

### 2.8. Estimation Techniques

#### 2.8.1. Vector Auto-Regression (VAR) analysis

The estimation process encountered a challenge of determining what variable adequately captures fiscal policy stance so that it may be included in the empirical equation. The literature has demonstrated that no single variable appear to be the best variable to estimate in ascertaining the effect of government expenditure on private investment. Further, economic theory does not provide adequate information on Granger causality between government expenditure and private investment. Therefore, following (Fu, Taylor, & Yucel, 2003) and Sims (1980), the study adopted a VAR model for estimating simultaneous shocks to more than one variable and used that to investigate unexpected and equivalent structural shocks. VAR modeling techniques was used to achieve the five objectives out of the six objective stated in chapter one. These objectives include: To investigate the causal relationship between disaggregate government expenditure and other variables on private investment in Ethiopia, analyze the relative effects of various components of government expenditures on private investment in Ethiopia (crowding in and crowding out), analyze the Effect of Budget Deficit on Private Investment in Ethiopia, determine the effects of government domestic credit to the private investment in Ethiopia and to show the impulse response function (IRF) of government expenditure, GDP and inflation on private investment to examine the positive and negative response of private investment to change in different variables in the model for ten years’ time horizon. Use of VAR in the study was on the justification that it is a theory free method used for the estimation of economic relationships(Sims, 1980).

The study mainly considered fiscal variables in the VAR since the main focus was on the government expenditure and its effects on private investment. Three different types of VAR exist: The reduced form VAR, the recursive VAR and the structural VAR. The recursive and structural VAR have the same form at the level of matrix equations. The reduced VAR sidestepped the need for structural modeling by modeling every endogenous variable in the system as a function of the lagged values of itself and of all the endogenous variables in the system Engle & Granger (1987). The reduced form and the recursive VAR models are statistical models that utilize no economic structure beyond the choice of variables. The compact form of a VAR model is represented as:

Where nx1 vector of constant terms is are n x n matrix of coefficients is n x1 vector of endogenous variable and is a vector of serially uncorrelated error terms that have a mean of zero and a covariance of matrix Φ. In the VAR model, each variable was regressed on a constant variable , P lags of itself, and p lags of each of the other variables in the model and the disturbance term .

The longer lag lengths are normally appropriate since they fully capture the dynamics of the system being modeled and increasing the parameters. However, given the data limitations, lag length determination became a major challenge. This is because, longer lags reduce the degree of freedom and the problem is further compounded by data limitations. Therefore, there was a need to have a trade **-** off between having a sufficient number of lags and a sufficient number of parameters to estimate. The choice of p (the number of lags) was determined using the Akaike information criteria (AIC) and the Schwartz information criteria (SIC).

Recursive VAR was used to examine the interrelationships among a set of economic variables and to analyze the dynamic effect of random disturbance on the system of variables. Various diagnostic tests were performed to find out the stability and statistical soundness of the estimated model. These included multicollinarity, Arch residual test; Ramsey reset test and white noise tests, among others see the test result in appendix A2 on the back side. In the framework, each variable irrespective of whether it was measured at levels or a given difference level was treated systematically. This implies that all variables in the system contained the same set of regressors (Mc Coy, 1997). There were no exogenous variables and no identifying restrictions. Economic theory therefore played only the role of specifying what variables was to be included. The variables of interest in this study to test for the effects of government expenditure and other fiscal variables on private investment. Where, private investment (PI), Public Investment ) public consumption (PUC), Domestic credit to private sectors Economic growth (GDP), inflation (INF), capital government expenditure (CEX), recurrent government expenditure (RECX), Budget deficit (BD) and(D) dummy variable for structural adjustment program (SAP).

#### 2.8.2. Impulse Response Analysis

VAR usually yield coefficient estimates which are meaningless because of the lack of theoretical underpinning. However, the coefficient estimates were used in the derivation of impulse responses and in forecasting error decomposition. Impulse response analysis linked the current value of the error **-** term to the future values of or similarly, the current and past values of the error **-** term to the current values of Forecast error decomposition measured how important the error in the equation was for explaining unexpected movements in the variable (Stock & Watson, 1998). An impulse response enabled the study to trace the effect of one time shock to one of the innovations on the current and future values of the endogenous variable. These impulse responses were obtained from a Vector **-** Auto Regression Moving Average (VARMA). The coefficient of the VARMA representation, described how a shock to a particular variable at one moment in time shifts the expected time path of each variable, in the model compared with its expected evolution had the shock not occurred.

#### 2.8.3. Variance Decomposition

While the impulse response functions traced the effects of shocks to one endogenous variable on the other variables in the VAR, variance decomposition separated the variation in an endogenous variable into the component to the VAR. To determine what proportion of the variance in a series was due to its own shock and other identified shocks, forecast error variance decomposition technique, which allocates weights to each identified shock in the system at every forecast horizon for a particular variable was used (Odour, 2008). Over a short horizon the own shock dominates the variance forecast and shocks to other variables in the system may gain importance relative to own shock, as the horizons lengthens. The study carried out variance decomposition to determine the proportions of the shocks in private investment that were accredited to fiscal policy variables and therefore established their relative importance in determining private investment in Ethiopia.

#### 2.8.4. The Error-Correction Model

The regression of private investment against government expenditure and other variables enabled the study to get the effects of these variables on private investment. However, the study could apply ordinary least squares method (OLS) directly since all variables were stationary at levels. Differencing the time series prior to estimating model could have only described the relationship between changes in variables and regarded the long-run and short run relationship between private investment and fiscal variables. ECM model adopted by this study assumes that the short-run effects occur when the economy is still in disequilibrium, and that the long-run effect occurs when the economy moves to equilibrium (Enders W. , 1995).The long-run equation (1) was lagged one period to derive the error correction mechanism, which was included in the error correction model. The coefficients in the ECM, describe the effect of a unit change of a given fiscal variable on the private investment.

### 2.9. Chow Test for Structural Stability

A series of data can often contain a structural break, due to a change in policy or sudden shock to the economy, i.e. the 1992 structural adjustment program (SAP) in Ethiopia was one of this changes. In order to test for a structural break, we often use the Chow test, this is Chow’ first test (the second test relates to predictions). The model in effect uses an F-test to determine whether a single regression is more efficient than two separate regressions involving splitting the data into two sub-samples (Chow & C.Gregory, 1960). This could occur as follows,In the first case we have just a single regression line to fit the data points (scatterplot), it can be expressed a

In the second case, where there is a structural break, we have two separate models, expressed as: This suggests that model (26) applies before the break at time t, and then model (27&28) applies after the structural break. If the parameters in the above models are the same, i.e. then models (26) and (27&28) can be expressed as a single model as in case (26), where there is a single regression line. The Chow test basically tests whether the single regression line or the two separate regression lines fit the data best. And the result is calculated by the following formula

Where RSSc= residual of regression using all the data i.e. year=1980-2012.

RSS1=residual of regression before structural change i.e. year<1980-1992

RSS2= residual of regression after structural change i.e. year>=1992-2012.

Eventually, if computed F value does not exceed tabulated F value at conventional significance levels (1% ,5% and 10%) the null hypothesis which states structural stability (parameters are stable over time ) would be not to be rejected which justifies using of pooled regression for 33 years. On contrary if computed F value exceed tabulated F value at conventional significance levels (1% ,5% and 10%) the null hypothesis states structural break (parameters are stable over time ) would be rejected and implies that the regression is made independently for time periods before and after structural adjustment program(Gujarati D. , 2004).

## CHAPTER THREE: RESULT AND DISCUSSION

### 3.1. Descriptive Analysis

Ethiopia is one of the least developed nations in the world. As per the findings of the studies which have been conducted in the country, the various civil wars that hit the country in the 1970s and 1980s, and the frequent droughts that have been occurred since the 1960s up to the present are believed to be the reasons behind the underdevelopment of the country (Tsegay, 2008).

Ethiopia had two different regimes with different policies and ideologies since 1975: the Derg regime (1975 to 1990) and the Ethiopian peoples’ Revolutionary Front (EPRDF) since 1991. The first regime followed command economic system while the later initiated economic reforms to address the long-term structural problems of under-development and followed a market oriented economic policy whereby the role of the private sector has been given priority. Pursuant to this fact, its economic performance has been different and fluctuating. Table 1 below and Fig 1below summarizes the average growth rate of real GDP, GDP per capita and population for the period ranging from 1974 to 2012

Table 1. Average growth rate of real GDP, GDP per capita and Population, (%) 1974 to 2012

illustration not visible in this excerpt

Source: Own calculation based on world development indicator World Bank database, 2011.

As can be seen from the above table during the entire period under study, the Ethiopian economy had an annual average real GDP growth rate of 4.5%, while population and per capita GDP grew by 2.6% and 2.4% respectively. Due to civil war, recurrent drought, high population growth and inappropriate economic policy and management, the performance of the Ethiopian economy was not satisfactory during the Derg regime. Average real GDP and per capita GDP growth rate were -0.13% and - 4.17% per annum for the period 1980 to 1985, respectively. This is the period of Derg regime when the government of Ethiopia, was implemented a socialist ideology. Following the end of the civil war on May1991, the EPRDF came into power with market oriented policy that brought about improvement in the economic performance of the country.as shown from the graph below the real GDP and GDP per capita during the derg regime it was small and even below from zero specially between 1980-1990 and in the early 1990’s and this figure was changed and showed an improvement after the failure of the derg government.

illustration not visible in this excerpt

Figure 1 Average growth rate of real GDP, GDP per capita and Population, (% of GDP) 1974 to 2012

Abbildung in dieser Leseprobe nicht enthaltenSource: own calculation based on data obtained from Mo FED 2010 The average growth rates of real GDP and GDP per capita have increased to 9.06% and 6.23% respectively. During this period, the importance of private sector with significant intervention of government in development has been recognized.

In view of the above scenarios, one can deduce that any economic outcome in Ethiopia is largely associated with the political process. Accordingly, the issue of government spending and private investment should be addressed in relation to the above two distinct regimes. The discussions in the subsequent subsections of this chapter also follow this framework. Besides, since the way expenditures are financed have a strong bearing on the effect of government spending on private investment, attempts have been made to link the financing issue to the gist of the study.

**[...]**

- Quote paper
- Frew Hailu (Author), 2014, Assessement of the effect of government expenditure on privat investment in Ethiopia, Munich, GRIN Verlag, https://www.grin.com/document/282156

Comments