See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/312091590

Public and Private Investment Analysis in Ethiopia

Article in Research Journal of Economics and Business Studies · April 2015

CITATIONS READS

5 11,520

1 author:

Kurabachew Menber
Punjabi University, Patiala
7 PUBLICATIONS 210 CITATIONS

SEE PROFILE

All content following this page was uploaded by Kurabachew Menber on 06 January 2017.

The user has requested enhancement of the downloaded file.

 Public and Private Investment Analysis in Ethiopia
Kurabachew Menber
Lecturer at Dire Dawa University in the Department of Economics, Ethiopia PhD Scholar in
Economics in Punjabi University, Patiala

Abstract
There has been a growing evidence of the link between investment and economic growth. But when it
is viewed at sector-wise, public and private investment, there has been a debate on the relationship
between them. In some countries there was evidence of the crowding out effect of public investment
on private investment, while there has been the crowding in evidence in other countries. But for sure, a
complementary relationship between the two is preferred for better economic growth. This study
found out the existence of the crowding out effect in the country. The study also evaluates the
macroeconomic determinants of private investment in Ethiopia for the last three and half decades
(1975-2010) in a short, medium and long run perspective by means of a regression analysis based on
the co-integration and Error Correction Model (ECM) of Engle and Granger (1987) combined with the
general to specific methodology and the distributive lag model. Econometric results of the study shows
public investment, external debt, gross private consumption level, credit availability to private sectors,
output and trade liberalization significantly determine private investment in the long run. The changes
in inflation rate, interest rate, gross private consumption, public investment and debt servicing have a
negative impact while credit availability, real exchange rate deprecation, output growth rate and
domestic savings have a positive effect in the short run. The error correction model estimation showed
the existence of a shorter period of adjustment between the short run and the long run dynamics of
private investment determinants in Ethiopia. Accordingly, to bring better economic growth in the
country redirection of public investment from non-infrastructural sectors to the social over heads is
required. An appropriate interest rate policy to boost saving on the one hand and measures to widen
and modernize the scope ofo credit disbursement to private sectors on the other hand is recommended.
Moreover, sound macroeconomic policies to cheek uncertainties, mechanisms to curve luxurious
private consumptions as well increasing efforts for debt cancellation and relief there by to reduce
future indebtedness are some of the presumed measures to be taken so as to boost private investment in
the country.

1.1. Background of the Study

Investment is the nucleus of an economy. It plays a crucial role in the models of economic growth. It is
an essential component of aggregate demand and fluctuations in investment have considerable effect
on economic activity and long term economic growth. The achievement of the desirable goals of stable
macroeconomic condition and faster growth partly depends on the level of investment. Since it is an
essential macroeconomic variable through which economic growth of a country is achieved, the
analysis of the sectoral scene of the country`s investment scenario is a prerequisite to wards designing
an appropriate investment policy. In the sense of this study sectoral analysis refers to the analysis of
investment based on ownership-that is public and private investment. And the analysis confined to the
identification of the relationship between private and public investment in the one hand and the
identification of short run and long run determinants of private investment on the other hand in the
country.

Total investment increased sharply since 2000 in the country. After an initial jump at the end of the
socialist regime in 1991, total investmentstagnated as a share of GDP at about 15–16 percent. In 2001,
at the end of the 1998–2000border conflict with Eritrea), total investmentrose strongly to reach some
20–21 percent of GDP in 2007. This compares reasonably wellwith similar countries in the region,
such as Kenya, Tanzania, Uganda, and Zambia, as wellas to the sub-Saharan African (SSA)
average(Tafere, 2008). But when it is viewed sector-wise, public investment mimicked total

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  26  
investment trend but not private investment.Public investmentaveraged about 12 percent of GDP
during this period, and was higher than private investmentby about 5 percent of GDP. About a fifth of
public investment is undertaken by publicenterprises, mainly the major utilities. According to
IMF(2005), large infrastructure investments are needed not only to support the creation of rural-urban
linkages and rural growth, but also to create an enabling environment for private sector development.
The issue weather government expenditures (particularly government investment increase or decrease
private investment: i.e., crowding in versus crowding out private investment has been a recent debate
among macroeconomists. It is said that inadequate provision of physical infrastructure (i.e., public
investment), has been a primary cause of poverty and unemployment in developing countries like
Ethiopia. Government investment expenditure is unavoidable for the creation of physical infrastructure
in such countries and thereby it works as a feeder for increasing gross state domestic product and
employment in private sector of the country’s economy. In Ethiopia, on one hand public investment
showed an increasing trend while private investment is declining as % of GDP, and the country is in
need of increased public investment in infrastructures. This makes investigation of the relationship
between the two sectors more interesting.

According to Aschauer (1998), investment in infrastructure (public investment) is necessary for a

strong, flexible, and growing economy. When he was dealing with its relationship with private
investment, the short- and long-term growth-maximizing effects of public investment increase as the
ratio of public to private capital stock rises to an optimal level (found to be about 61 percent); above
that level the growth effects decrease. He concludes that economic growth can be increased by
increasing the public capital stock and by redirecting government expenditure from noninvestment
consumption spending to public investment. Many other studies in developing countries also showed
the existence of complementary relationship between public and private investment. But also there are
studies which showed a supplementary relationship in some countries. For example, Cruz & Teixeira
(1999) showed that public and private sector investments were complementary in the long term and
substitutive in the short term for Brazil. Apergis (2000) found a crowding out effect of public
investment on private investment under the period 1981-1996 for Greece. The same was true in the
findings of Acosta & Loza (2005) for Argentina. It is therefore important to make an investigation on
the relationship between the two.

According to investigations in developing countries there are multi dimensions of variables which
determine private investment. From simple macro-economic variables like inflation, exchange rate,
term of trade, interest rate to the complex of dummy variables like passage of legislation, coups and
their after effects determines private investment.
A thorough investigation made by Serven (1996)with panel data set of 94 developing countries showed
a strong negative link between investment and macroeconomic instability which is measured by
volatility of five variables: inflation, relative price of capital, the growth of output(measured by real
GDP), the real exchange rate and the term of trade. An Empirical Analysis by Greene & Villanueva
(1991) in developing countries during 1975-87 indicates that the rate of private investment is
positively related to real GDP growth, level of per capita GDP, and the rate of public sector
investment, and negatively related to real interest rates, domestic inflation, the debt-service ratio, and
the ratio of debt to GDP. According to Daniel (2005) real GDP and its growth rate have a positive
impact in the long run on private investment in line with the prediction of the accelerator model. Public
investment in infrastructures also has high significant explanatory power. His result suggests a long
run positive impact of public investment on private investment. On the other hand foreign exchange
availability, affects positively in both periods while inflation rate-in both periods- and external debt to
GDP ratio –only in the long run-have a strong negative effect. The implication is that macroeconomic
instability and uncertainty has a negative impact on private investment.
Besides the investigation of the crowding in versus crowding out effect of public investment on private
investment, the identification the main determinant of private investment both in the long run and in
the short run is the second issue given an emphasis in this paper.

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  27  
1.2. Statement of the Problem
Economic theory assures that economic growth of a country is largely related to the level of capital
formation (investment). Investment plays a very important and positive role for the progress and
prosperity of any country. Many countries rely on investment to solve their economic problems such as
poverty, unemployment etc. Here note that not getting increased level of capital formation is the only
trick, but also designing an appropriate sectoral and regional investment pattern is important to benefit
from its end results. To this end the effect of public spending on private sector expenditure has
received considerable attention in the economic literature. This is because of the fact that mainly
public expenditure has a crowding- out effect. But for an economic growth to come economists argue
that there should be a complimentary rather than a substitutability relationship between the two. This is
because the “crowding-out” effect reduces the ability of the government to influence economic activity
through fiscal measures.
Public and private investments play a crucial role in achieving the goals of economic and
socialdevelopment. The nature of public and private investment gave rise to a strong argument in
economic theory and policy for a complementary relationship between them. Besides to this, there are
multi dimensions of variables which determine private investment in Ethiopia. Therefore it is
important to identify those significant variables which determine private investment in the country.
This helps to know where to focus and what policy measures to take so as to boost private investment
in the country. The purpose of this study is to identify the kind of relationship that existed between
these two sectors in Ethiopia and then to investigate the main explanatory variables which determine
private investment.

1.3. Objectives of the Study

The main objective of this study is to identify the crowding in/out effect of public investment on
private investment after their relationship identification and investigate the main determinants of
private investment both in the long run and in the short run in Ethiopia using a time series data.
Specifically, the main objectives to be attained are:
• Prior identification of the relationship between public and private investment in Ethiopia,
• Empirical investigation of whether public investment `crowds out’ or ‘crowds in’ private
investment in the country via econometric model;
• Identification of the main determinants of private investment in the country through empirical
model, private investment as dependent variable.

2. Data and Methodology of the Study

Basically the investigation is based on secondary time series data on the range of 1975 to 2010. Public
investment, private investment and gross domestic saving all as percentage of GDP, are taken from
World Bank national accounts and World Bank Global Development Finance. The external debt as
percentage of GDP is taken from ERS (Economic Research Service) International Macroeconomic
Data Set. The real interest rate is taken from National Bank of Ethiopia. The source of data for real
GDP growth rate and GDP itself is world pin table 7 designed by the IMF. All the other variables are
taken from The World Bank.
In this study empirical analysis was used to identify the crowding in/out effect of public investment on
private investment and for the investigation of private investment determinants in the country both in
the short run and in the long run.
This study used two models. The first model is used to identify the nature of relationship between
private and public investment. The estimation of the model under the given criteria helped to know the
direction of relationship between the two (that is whether public investment is weakly exogenous to
private investment or on the other way). Knowing the direction of relationship intern enables to solve
the crowding out versus crowding in controversy in the country. The first model was adapted from
Coutinho & Gallo (1996) while they used the model for the same purpose for developing countries.
Let “𝐼𝑝𝑟𝑡”and “𝐼𝑝𝑢𝑡” as private and public investment as percentage of GDP at time `t’ respectively,

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  28  
then the assumption of the long run relationship of the two is characterized by their ratio being
constant, i.e.
𝐼𝑝𝑟𝑡  /  𝐼𝑝𝑢𝑡   = 𝑘. ….……………………………………………………..…………… 3.1
This expression, taking logs can be written as,
𝑙𝑜𝑔𝐼𝑝𝑟𝑡  – 𝑘 − 𝛽𝑙𝑜𝑔𝐼𝑝𝑢𝑡   = 𝐸𝐶𝑡...………………………………………………………...3.2
, where𝐸𝐶𝑡 is error correction at time t, provided that the two variables are co-integrated. But the more
general error correction model is the following:
𝛥𝑙𝑜𝑔𝐼𝑝𝑟! 𝐴 𝐴! 𝛥𝑙𝑜𝑔𝐼𝑝𝑟!!! 𝛿 𝜀!!
= ! + ! 𝐸𝐶!!! + 𝜀 … … … … … … … . .3.3
𝛥𝑙𝑜𝑔𝐼𝑝𝑢! 𝐵! 𝐵! 𝛥𝑙𝑜𝑔𝐼𝑝𝑢!!! 𝛿! !!
Note that,“𝐸𝑐𝑡 − 1 = 𝑙𝑜𝑔𝐼𝑝𝑟𝑡 − 𝑘 −  𝛽𝑙𝑜𝑔𝐼𝑝𝑢𝑡“is obtained from the residual of a first stage
regression of 𝐼𝑝𝑟𝑡 on 𝐼𝑝𝑢𝑡 with a constant (intercept) k, that is:
𝑙𝑜𝑔𝐼𝑝𝑟𝑡 =  𝑘 +  𝛽𝑙𝑜𝑔𝐼𝑝𝑢𝑡   + 𝑢𝑡  
From the estimation of this error correction model through OLS, there will exist four cases:
1. If δ1 =0 and δ2=0, there is no relationship between the two.
2. If δ1 =0 and δ2≠0, private investment leads public investment (private investment is weakly
exogenous).
3. If δ1≠ 0 and δ2=0, public investment leads private investment (public investment is weakly
exogenous).
4. If δ1≠ 0 and δ2≠0, there is an inter dependence between the two and based on the signs of the
coefficients one can decide on the crowding in/out effect.
Note that, not only must co-integrated variables respect the above vector error correction model but the
converse must also hold: variables in a model where both the difference and levels coexist must be co-
integrated. In this case, one of the two or both coefficients δ1 andδ2 are different from zero. This is the
hypothesis of weaklyexogenous. But if both are equal to zero, then there is no co-integration between
the two variables (Coutinho & Gallo, 1996).
The second model is all about the identification of those variables which affect private investment in
Ethiopia. To test the relationship between private investment and its different determinants the
following model in functional form is used:
𝐼𝑝𝑟 = 𝑓  (𝐼𝑝𝑢, 𝑟𝑒𝑟, 𝑑𝑠, 𝑡𝑟𝑑𝑙, 𝑟𝑖𝑟, 𝑑𝐼𝑓, 𝐺𝐷𝑃, 𝑝𝑐𝑜𝑛, 𝑔𝑑𝑝𝑟, 𝐸𝑥𝑑𝑏, 𝑑𝑠𝑒𝑟𝑣, 𝑝𝑐𝑟𝑑, 𝑑, 𝜀) … … … 3.4  
Where Ipr is private investment as percentage of GDP, rer –real exchange rate, ds-domestic saving as
percentage of GDP, trdl-trade liberalization (import Plus export as percentage of GDP), “𝑟𝑖𝑟”-real
interest rate, “𝑝𝑐𝑜𝑛”-gross private consumption as percentage of GDP, “𝑑𝑙𝑓”-domestic inflation
rate,  “𝑑𝑠𝑒𝑟𝑣”-debt servicing as percentage of GDP, “𝑔𝑑𝑝𝑟”-growth rate of GDP, 𝑐𝑟𝑑-credit
availability to the private sector as percentage of GDP, d-dummy variable for policy regime (1 after
1992 and 0 before) and ε is the error term. The dummy variable is included to illustrate the effect of
liberalization policy to private investment after 1992.
The methodology followed in this section of the study has four stages as it is indicated in(Acosta &
Loza, 2005). First, the order of integration of each of the series used in the analysis was determined by
applying the stationary or unit root tests. For this purpose the Augmented Dickey Fuller (ADF) test
was used at the beginning. But in case if there is a structural break in a model the Philips-Perron unit
root test is preferred. Hence at the beginning both probabilities were examined and Chow test for
structural break was conducted for the year where there has been a change in investment policy that
was 1992. The detail for the ADF and Philips person test of unit roots are discussed in the next section
3.3.3.1. In this stage, the order of integration for each series has been identified.
In the second stage, an investment equation that represents the long run determinants of private
investment was estimated. This was done in such a way that Engle-Granger and Johansson’s co-
integration test were conducted on the integrated of order one𝐼(1) series. Both tests proved the
existence of long run equilibrium relationship between these variables. The result from the estimation
of this co-integration equation enabled the identification of long run determinants of private
investment by applying the general to specific methodology. The Johansen co-integration test is used
for the mere identification of the existence of co-integrated relationship between the hypothesized

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  29  
independent variables and the dependent variable. Further inference was done by the Engle-Granger
technique, since it is the analysis of a single equation model.
The aim of the third stage was identification of short run determinants of private investment. This was
done through the use of an auto regressive distributive lag model by using the methodology of “from
general to particular”. In this stage those variables that are significant in the private investment
equation were identified, with their respective lags at difference. Both the difference and their
respective lags of the integrated and level series were included in the estimation process. This made
identification of all variables that significantly affect private investment in the short run easy.
In the last stage, an error correction model which relates the estimated long run investment equation
with the short run one was identified and analyzed. This error correction model helps to identify
weather there exist a disequilibrium in the short run model from the long run and also if exist how the
speed of adjustment takes place.
While each model was estimated a diagnostic test was conducted on the residuals and in the model to
cheek weather the estimation was accurate or not. These diagnostic tests include Ramsey RESET test
for regression specification errors, Jarque-Bera test for normality, Breusch-Godfrey Serial Correlation
LM testfor autocorrelation, ARCH and White tests for heteroskedasticity and the stability tests of
CUSUM, CUSUM square of residuals and recursive residuals test were made. All the tests were
passed and the regressions were not spurious.

3. Results and Discussions

3.1.1. Normality and StationaryTests
The Jarque-Bera test statistic for normality and distribution graph of the Kernel Density showed the
necessity of taking the log transformation of the variables which were non-normal in distribution.After
these variables were cheeked for normality each of them were tested for stationary by using the ADF
test and PP test.Public investment, private investment, credit to private sectors, trade liberalization,
debt servicing, gross private consumption, external debt all as % of GDP, and log of output(GDP) and
real exchange rate series were integrated series of order one I(1), being non-stationary in level while
their first differences are stationary.Gross domestic saving as % of GDP, real growth rate of GDP,
domestic inflation rate and real interest rate series were stationary in level.
3.2. Models Estimation and Results
3.2.1. Estimation Result for Relationship Identification
This section of the paper examines the relationship between private and public investment as
percentage of GDP through the error correction model specified in the previous chapter (equation 3.3)
and further confirmed by a co-integration test of Ipr and Ipu. The identification of the existing
relationship between the two helps to assess whether public investment acted as a catalyst or competed
for scarce resources in the country. The error correction model specified in equation three in the
previous chapter as matrix form can be written as two system of equations:
𝛥𝑙𝑜𝑔𝐼𝑝𝑟! = 𝛼! 𝛥𝑙𝑜𝑔𝐼𝑝𝑟!!! + 𝛼! 𝛥𝑙𝑜𝑔𝐼𝑝𝑢!!! + 𝛿! 𝐸𝑐!!! + 𝜀!!
… … .          4.2
𝛥𝑙𝑜𝑔𝐼𝑝𝑢! = 𝛽! 𝛥𝑙𝑜𝑔𝐼𝑝𝑟!!! + 𝛽! 𝛥𝑙𝑜𝑔𝐼𝑝𝑢!!! + 𝛿! 𝐸𝑐!!! + 𝜀!!
WhereEct-1 is obtained as defined before. The estimates of these models have given the following
coefficient results.
Table 4.3 Coefficient estimates of relationship identification equation 4.1.
variables Coefficient Std. Error t-Statistic Prob.
𝛥𝑙𝑜𝑔𝐼𝑝𝑟!!! 0.028570 0.059180 0.482769 0.6310
𝛥𝑙𝑜𝑔𝐼𝑝𝑢!!! -0.185062 0.168990 -1.095108 0.2777
𝛿! -0.317940 0.141548 -2.246156 0.0283
𝛥𝑙𝑜𝑔𝐼𝑝𝑟!!! 0.039026 0.046842 0.833147 0.4080
𝛥𝑙𝑜𝑔𝐼𝑝𝑢!!! -0.186864 0.133757 -1.397046 0.1674
𝛿! 0.136594 0.112037 1.219184 0.2274
Note: For the complete output see Table A2 in the Appendix

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  30  
The value of δ1 is significantly different from zero at 5% significant level but δ2 is insignificant. The
implication is that public investment is weakly exogenous to private investment. In other words, the
result showed that public investment leads private investment but not the other way. There is a one
way relationship. But the insignificance of the other variables coefficients makes it impossible to
distinguish weather public investment is crowding in or crowding out private investment an adjusted
R2 below 25% makes it insufficient to make conclusions.
Besides this error correction model, the Engle granger and CRDW tests of co-integration between
𝐼𝑝𝑟and 𝐼𝑝𝑢, i.e., private and public investment as percentage of GDP, including a dummy variable to
account for the structural break showed the existence of a long run equilibrium relationship between
the two.
Addition of a dummy variable representing the liberalization policy in 1992 to account for the above
observed structural break and the instability gives the following result for the model:
𝐼𝑝𝑟𝑡 =  𝛽0 + 𝛽1𝐼𝑝𝑢𝑡 +  𝑑𝑢𝑚𝑚𝑦 +  𝜀𝑡 ………..4.3
Table 4.5 Coefficient estimates for model 4.3
Dependent variable 𝐼𝑝𝑟𝑡  
Variable Coefficient Prob.
Iput -0.302947 0.0044
Dummy 4.665229 0.0000
constant 7.842833 0.0000
2
Adjusted R 0.429775
DW 1.127647
F-statistic 12.43593 0.0000
The Engle-granger test for the above model depends on unit root test on the residual regression from
the above model. The result obtained from the residual regression was:
𝛥𝑢𝑡 =   −0.575660 ∗ 𝑢𝑡 − 1  
𝑡 =       −3.743808    
𝑅2 =  0.291819  
The Engle–Granger 1 percent critical τ value is −2.5899. Since the computed τ (= t) value is much
more negative than this, the implication is that the regression of 𝐼𝑝𝑟𝑡 and 𝐼𝑝𝑢𝑡 including the dummy
are I(0); that is, they are stationary. This result reconciles with the error correction model estimated
above. Private investment and public investment have long run equilibrium relationship. The
coefficient of 𝐼𝑝𝑢𝑡 in equation (4.3) indicates that public investment has a possibility of crowding out
effect on private investment in the country.
3.2.2. Long Run Determinants of Private Investment
The method used to identify the determinants of private investment both in the long run and short run
is the estimation of the general function proposed in equation 3.4 in the following steps. Using the
methodology “from general to particular”, it is concluded that private investment is positively co-
integrated with output (log of real GDP), the domestic financing opportunities (credit to private
sector), trade liberalization (import plus export as % of GDP), while it is negatively co-integrated with
public investment, and gross private consumption. The co-integration equation shows that the
existence of a crowding out effect of public investment on private investment. In the long run, a unit
increase in public investment as percentage of GDP is accompanied by 0.6 unit decrease in private
investment as percentage of GDP. The result shows the existence of a sort of competition of public
investment to private investment for scarce resources. In support of the accelerator theory output has a
positive effect on private investment in the long run. A percentage increase in output leads to 0.014unit
increase in private investment as percentage of GDP.
Table 4.7 Private investment in Ethiopia, 1975- 2010- co-integration
Variable Coefficient probability
Public investment -0.597360 0.0000***
Log of GDP 1.361738 0.0001***
Credit to the private sector 0.029386 0.0003***

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  31  
Private consumption -0.189679 0.0031***
Trade liberalization measure 0.103210 0.0621*
constant 12.77949 0.0104**
Adjusted R2=0.775542 Observations=36
test statistic probability
JarqueBera(normality test) 1.944832 0.378168
Serial correlation(LM test) 1.647069 0.438878
ARCH LM test 0.441386 0.506454
White heteroskedasticity test 9.309144 0.503037
Ramsey RESET test 0.503687 0.777366
***significance at 1% *Significance at 10%

The estimate from the above Engle-Granger co-integration techniques confirms most of the empirical
results found in the literature: the perspectives of growth (output), profitability (trade liberalization),
and viability (domestic financing) of the economic system are the main variables that guide the
investment decision in the long run. The private consumption behavior is also significantly affects
investment decisions in the long run.
The more valid co-integration technique, the Johansen co-integration test, also confirmed the existence
of a long run equilibrium relationship with greater precision.
Table 4.8 Johansen co-integration test result
Eigenvalue Likelihood 5 Percent Critical 1 Percent Critical Hypothesized No. of
Ratio Value Value CE(s)
0.871553 308.8797 192.89 204.95 None **
0.829949 239.1037 156.00 168.36 At most 1 **
0.822404 178.8672 124.24 133.57 At most 2 **
0.654900 120.1069 94.15 103.18 At most 3 **
0.562590 83.93357 68.52 76.07 At most 4 **
0.474745 55.81948 47.21 54.46 At most 5 **
0.447369 33.92784 29.68 35.65 At most 6 *
0.293315 13.76363 15.41 20.04 At most 7
0.056013 1.959842 3.76 6.65 At most 8
*(**) denotes rejection of the hypothesis at 5 %( 1%) significance level

L.R. test indicates 7 co-integrating equation(s) at 5% significance level.

The eigenvalue statistic rejects the null hypothesis of the existence of zero or one co-integrating
relationships in the private investment equation, but accepts the alternative existence of six co-
integrating relationships at least at the 5% significance level. The long run relationship is guaranteed
by the existence of at least one co-integrating vector which confirms the final model estimated by the
general to specific model in equation 4.4 is co-integrated.
3.2.3. Short Run Determinants of Private Investment
Once the long term relationship is obtained, it is interesting to estimate a partial adjustment model (or a
short term relationship) between private investment and its main determinants. This is done via
estimation of a distributive lag model from general to specific method by including those level
stationary variables not included in the long run equation. The Distributed Lags (DL) specification
proposes combining integrated variables with lags and the stationary variables, including the
dependent variable. The general form of these models is the following (Acosta 2008):
Dyt= b0 + A(L)Dyt-1 + B(L)Dqt+ht,,
Where yt represents private investment, qtthe vector of independent variables, and ht the error term.
By using the methodology “from general to particular” , by dropping those variables which are found
to be insignificant in the hypothesized model, the short run behavior of private investment ends up
with the estimation of the following preferred model. Dropping of insignificant variables results in
higher adjusted R2 and low AIC and SIC for subsequent models till the preferred model is found. The
preferred model that represents the short run behavior is the following:

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  32  
 𝛥𝐼𝑝𝑟𝑡 =α0+α11ΔIprt-1+𝛼!" 𝛥𝑙𝑔𝑑𝑝!!! + !!!!(𝛼!! 𝛥𝐼𝑝𝑢!!! +𝛼!! 𝛥𝑐𝑟𝑑𝑝!!! +𝛼!! 𝛥𝑑𝑠!!! +
𝛼!! 𝛥𝑟𝑖𝑟!!! ) + 𝛼! 𝛥𝑝𝑐𝑜𝑛! +𝛼! 𝛥𝑟𝑒𝑟! +𝛼! 𝛥𝑑𝑠𝑒𝑟𝑣!!! +𝛼!" 𝛥𝑔𝑑𝑝𝑟!!! +𝛼!! 𝛥𝑑𝐼𝑓!!! +
+𝛼!" 𝑑𝑠! +𝛼!" 𝑟𝑖𝑟! +𝛼!" 𝑑𝐼𝑓! +εt
As the result on the table shows, the tests concerning the behavior of the errors 𝜀𝑡are also included.
The hypothesis of serial correlation in the error term, that would lead to biased estimates due to the
presence of lags of the dependent variable in the right hand side of the equation, is rejected (Breusch-
Godfrey Test).
Similarly, both the heteroskedasticity (Breusch-Pagan Test) and non-normality (Jarque-Bera Test)
hypotheses are rejected. So the estimates are both consistent and efficient.
Table 4.9 first differences private investment in Ethiopia, 1975-2010 - distributive lags and Error
correction
Variable Coefficient Prob. Erorr correction
𝐸𝐶!!! -0.847623***
ΔIprt-1 -0.421478 0.0140
𝛥𝐼𝑝𝑢! -2.309675 0.0001 -0.700817***
𝛥𝐼𝑝𝑢!!! -0.398877 0.0309
𝛥𝑙𝑔𝑑𝑝!!! -26.55735 0.0102
𝛥𝑝𝑐𝑜𝑛! -0.991221 0.0005 -0.334818***
𝛥𝑟𝑒𝑟! 6.536649 0.0008 1.010622***
𝑟𝑖𝑟! 1.011489 0.0015
𝛥𝑟𝑖𝑟! -0.445276 0.0034
𝛥𝑟𝑖𝑟!!! -0.933637 0.0017
𝛥𝑐𝑟𝑑𝑝! 0.689983 0.0086
𝛥𝑐𝑟𝑑𝑝!!! 0.751319 0.0059
𝑑𝑠! -1.824332 0.0012
𝛥𝑑𝑠! 1.606994 0.0008 0.221640**
𝛥𝑑𝑠!!! 0.662172 0.0020
dIf 0.784682 0.0033
𝛥𝑑𝐼𝑓!!! -0.533810 0.0032
𝛥𝑑𝑠𝑒𝑟𝑣!!! -3.250312 0.0067
𝛥𝑔𝑑𝑝𝑟!!! 0.099347 0.0293
α0 11.14225 0.0050
F 6.159164 0.002622 9.9178321***
Adjusted R2 0.777527
Jarque-Bera(normality test) 0.816735 0.664735
Breusch-Pagan(ARCH Test) 0.028355 0.985922
Breusch-Godfrey(Serial correlation LM test) 2.400588 0.301106
Ramsey RESET test 4.370405 0.224145
***significant at 1%, ** significant at 5%
The result shows that there is evidence of partial adjustment in private investment in Ethiopia (only the
first lag of the difference is significant at the 5% level). On the other hand, other variables also
influence the short term behavior of private investment.
Output affects private investment at first lag negatively (a one unit change in GDP leads to a -0.26 unit
change in private investment as percentage of GDP at one lag). But the growth rate has a positive
impact (a unit change in real GDP growth rate leads to a 0.1 change in private investment as
percentage of GDP). The implication is that it takes time that output to be converted to investment.
Moreover, since private investment is measured as % of GDP, the negative sign is an indicator of the
trend in private investment was not going with the output trend. Output was increasing at more
amounts. The negative impact vanished in the long run and turned out to be positive, which supports

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  33  
the accelerator hypothesis (output has a reciprocal relationship with investment). The overall impact of
output on private investment is positive.
There is also evidence that supports the theory of a “crowding-out” effect of the public investment on
private investment in the short run. A unit change in public investment as percentage of GDP at level
and at one lag respectively, leads to a 2.3 and a 0.4 change in private investment as percentage of GDP
in the opposite direction. As the co-integration model shows, in the long term this effect persists and
there is evidence of a negative relationship between public and private investment both in the long run
and in the short run. This suggests that there is a sort of competition for resources between the public
and the private sectors, both in the long run and in the short run. The implication is that there might be
a high level of public investment in non-infrastructure which competes resource for wide spread
involvement of the private sector in various business activities at greater extent in the country.
As it was expected gross private consumption has a negative impact on private investment. A unit
change in gross private consumption as percentage of GDP results in almost a similar opposite unit
change (-0.99) in private investment as percentage of GDP. The negative effect also persists in the
long run as it has been indicated from the estimation of the long run model above.
In the short run, currency deprecation has a positive effect on private investment, a unit change in real
exchange rate results in 6.5 unit changes in private investment as percentage of GDP. This finding is in
support of the view that devaluation reactivates the exportable sector of the economy and it also made
the acquisition of local assets by foreign companies at much lower price (Acosta, 2008). Hence it
enables to attract more foreign private investors in the country. But this positive effect did not persist
in the long run. It has no significant impact in the long run.
The expected rate of return is also an important determinant of private investment in the short run as it
can be shown from the interest rate. A unit rise in the real interest rate leads to the same level of
increase in private investment as percentage of GDP. But the change in the interest rate, theoretically,
is considered as the measure of uncertainty in the economy. A unit change in real interest rate at level
and at one lag respectively leads to -0.4 and -0.9 unit changes in private investment as percentage of
GDP.
Domestic financial availability is another variable which plays an important role for private investment
in Ethiopia. Domestic saving and credit availability to the private sectors significantly affects private
investment at least in the short run significantly. For instance, in the short run, a change in domestic
credit to private sectors as % of GDP both at level and first lag has a positive effect. A unit change in
domestic credit to private sectors as percentage of GDP at level and first lag results 0.69 and 0.75
change in private investment as percentage of GDP in the same direction. On the other hand, domestic
saving like output, has no an instantaneous effect on investment. It takes time to be converted in to
investment. The estimate of the short run model shows that a unit increases in domestic saving as %
GDP results a negative change in private investment as percentage of GDP by 1.8 units. But it has a
positive significant after effects. A unit change in domestic saving as a percentage of GDP both at
level and first difference respectively is accompanied by 1.6 and 0.6 unit change in private investment
as percentage of GDP.
According to Lesotlho (2006), on the one hand, higher expected inflation lowers the real interest rate,
causing portfolio adjustments away from real money balances to real capital, thereby raising real
investment, but on the other hand a high change in inflation rates are an indicator of macroeconomic
instability, which can have adverse impact on private investment. The finding in this study is similar
with this argument. The inflation rate itself has a positive effect (0.8) while its change at one lag has a
negative impact (-0.53). But in the long run inflation rate has no effect on private investment in the
country.
Lastly, debt servicing is one of the variables which significantly determine private investment in the
short run in Ethiopia. A unit change in debt servicing as percentage of GDP, leads to -3.25 units
changes in private investment as percentage of GDP.
Other variables like trade liberalization, external debt and policy change (investment liberalization) are
found to be insignificant variables to determine private investment in the short run. The complete
output for the short run model estimation is given in the Appendix Table A3

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  34  
3.2.4. Error Correction Specification and Results
Finally, it is necessary to compile in a single model both short and long run determinants of private
investment. For that, an Error Correction specification can be used, taking into account the speed of
adjustment to the long run trend of the series. This type of model which includes the long run and the
short run at once helps to correct the potential biases in the estimation of the coefficients in models
with differences that do not take into account co-integration relationships; when these long term
restrictions between the variables are ignored, there could be an omitted variables bias.
The initial proposed specification that includes both the preferred short and long term models (only
including the variables significant in both previous specifications) is the following:
𝛥𝐼𝑝𝑟𝑡 = 𝛼0 + 𝛿𝐸𝐶!!! + 𝛼11𝛥𝐼𝑝𝑟𝑡 − 1 + 𝛼!" 𝛥𝑙𝑔𝑑𝑝!!!
!

+ (𝛼!! 𝛥𝐼𝑝𝑢!!! +𝛼!! 𝛥𝑐𝑟𝑑𝑝!!! +𝛼!! 𝛥𝑑𝑠!!! + 𝛼!! 𝛥𝑟𝑖𝑟!!! ) + 𝛼! 𝛥𝑝𝑐𝑜𝑛!   + 𝛼! 𝛥𝑟𝑒𝑟!
!!!
+ 𝛼! 𝛥𝑑𝑠𝑒𝑟𝑣!!! + 𝛼!" 𝛥𝑔𝑑𝑝𝑟!!! + 𝛼!! 𝛥𝑑𝐼𝑓!!! + +𝛼!" 𝑑𝑠! +𝛼!" 𝑟𝑖𝑟! +𝛼!" 𝑑𝐼𝑓! + 𝜀𝑡
The variable ECt-1(the error correction term) is the deviation (or gap) from the long term trend in the
previous period:
 𝐸𝐶𝑡 − 1 = 𝐼𝑝𝑟𝑡 − 1 + 0.597360𝐼𝑝𝑢𝑡 − 1 − 1.361738𝑙𝑔𝑑𝑝𝑡 − 1 + 0.029386𝑐𝑟𝑑𝑝𝑡 − 1
+ 0.189679𝑝𝑐𝑜𝑛𝑡 − 1   −  0.103210𝑡𝑟𝑑𝑙𝑡 − 1)    
, and  𝛿𝐸𝐶!!! represents the long term speed of adjustment coefficient. But by applying the general to
specific methodology on this proposed model leads to the estimation of the following preferred error
correction model. The criteria used to drop insignificant variable is the same as before, adjusted R2,
AIC and SIC supplemented by residual and stability tests on the models specified each time a variable
is dropped.
ΔIprt=α0+δ𝐸𝐶!!! + !!!! 𝛼!! 𝛥𝑑𝑠!!! + 𝛼! 𝛥𝐼𝑝𝑢! + 𝛼! 𝛥𝑐𝑟𝑑𝑝!!! + 𝛼! 𝛥𝑝𝑐𝑜𝑛! +𝛼! 𝛥𝑟𝑒𝑟! +𝛼! 𝑑𝑠! +εt
All the residual, stability, and specification errors conducted for the above subsequent models was also
conducted for this preferred error correction model and is found to be free from any error(see in the
appendix table 4). The error correction outcome for this equation is given in table 4.7 fourth column.
All the short term results remain. This “medium term” model predicts no partial adjustment of private
investment to the previous period, since now the model is corrected by incorporating the deviation
from the long term trend. Changes in Public investment and private consumption are negatively
significant. Whereas change in real exchange rate, domestic saving and its change (at level and lag)
and change in credit to private sectors (at lag) are found to have positive effect on the change in private
investment. The significance of the error term (ECt-1) indicates the presence of disequilibrium in the
short run. As can be observed, due to an increase in private investment above the long term trend, the
preferred model predicts that more than four fifths of the gap (85%) is closed in one year.
4. Summary and Conclusion
This study analysed the nature of relationship between private and public investment in the country by
giving prior consideration. With this regarded, an error correction model developed by Coutinho and
Gallo (1996) was estimated. The estimation of this error correction model shows the existence of one
way relationship between the two. According to the finding public investment determines private
investment but there is no evidence on the counter way.
After the relationship between private investment and public investment were identified, the next
target of the study was to know the impact of public investment on private investment (whether it has a
crowding in or crowding out effect) on the one hand, and to know those factors which determines
private investment both in the long run and in the short run on the other hand.
To this end, an investment function with a list of variables which are suspected to influence private
investment with reference of many literatures was formulated. Then, a preliminary analysis of these
variables was conducted through descriptive statistics, normality test (Jarque-Bera) and stationary tests
for further analysis. Some are fond to be normal distributed while the non-normal ones are normalized.
The ADF and PP test conducted on the series showed that private investment, public investment,
external debt, debt servicing, trade liberalization, gross private consumption, domestic credit to
private sectors all as percentage of GDP, and real exchange rate and log of real GDP were non-

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  35  
stationary. They were all integrated of order one I(1). But real interest rate, inflation rate , gross
domestic saving as percentage of GDP and growth rate of GDP were stationary at level. They were
integrated of order zero 𝐼(0)
After the stationary test, Johansen`s co-integration test was conducted on the integrated one variables.
It showed the existence of at least seven co-integration equations. The Engle-Granger co-integration
technique by using the methodology of “general to specific” was applied to develop a model which
represents the main determinants of private investment in the country in the long run.
It was found that public investment had a crowding out effect on private investment. There should be
re-examination of the public investment structure so that there should be a complementary relationship
between the two rather than substitutability. There might have been an excess of public investment
structure areas that could be a business for the private sector or else it is competing for resources for
domestic private investors. External debt and gross private consumption had a negative impact, while
availability of credit to private sectors; output and trade liberalization had a positive impact on private
investment in the country. The external debt effect on the country’s economy is in support of the debt
overhang theory and the liquidity constraint hypothesis.1 In poor countries like Ethiopia, provision of
more financial resources for the private sectors is important to boost their investment activities. The
implication of the positive effect of trade liberalization measure is that more of the investment
activities are on those business activities which produce tradable goods. The same logic applies for the
effect of the real exchange rate in the short run.
Once the long run determinants of private investment were identified, a short run distributive lag
model was developed by using the “general to specific methodology” and estimated to explore what
affects private investment in the short run. From the estimation of a preferred model through this
process, it was found that changes in inflation rate and interest rate, changes in gross private
consumption, changes in public investment and debt servicing have negative impact on change in
private investment. But changes in credit availability to the private sectors and domestic savings, the
change growth rate of output, and deprecation of the real exchange rate had a positive impact in the
short run. Interest rate and inflation rate at level seems to have a positive impact but it vanishes and
turned out to be negative through time. The same is true for output and savings at level in the opposite
direction.
In the short run, the changes in interest rate and inflation rate which are proxy measures of uncertainty
in the economy have negative impact on private investment. While the real interest rate as a measure
for rate of return and inflation rate as an incentive for future profitability from increased price of
produced goods affects private investment positively2.
Finally, an error correction model which combines both the short run and the long run determinants of
private investment was estimated. The error correction model showed the existence of disequilibrium
in the short run model. But it also showed that an increase in private investment above the long term
trend, the preferred model predicts that more than four fifths of the gap (85%) was closed in one year.
It took less than two year adjustment period. The negative impact of change in public investment and
gross private consumption; and the positive effect of real exchange deprecation and change in
domestic savings persisted in the error correction model (in the medium term). The crowding in versus
crowding out controversy for the case of Ethiopia ends up here. Public investment had a crowding out
effect on private investment in the country under the period of investigation.
Crowding out effect could happen as a result of more public expenditure in these areas. In this case, it
results in displacement effect by competing for resources that could have been used by the private
sector. But on the other side, more public investment in infrastructure areas has the crowding in effect
on private investment. Therefore according to this theoretical perspective, there should be more
reduction in public investment in non-infrastructure areas that the private sector can take part in. On

                                                                                                                       
1
The  debt  overhang  theory  states  that  countries  having  high  external  debt  to  GDP  ratio,  finds  less  funds  available  to  
provide  an  environment  conducive  for  business  and  promote  investment.  The  liquidity  constraint  on  the  other  hand  
states  that  an  increase  in  external  debt  servicing  leaves  less  revenue  for  developing  countries  to  service  their  debt.  
2
Asante  (2000)  approximate  uncertainty  with  the  changes  in  inflation  rate,  interest  rate  and  exchange  rate.  

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  36  
the same side, government should invest more in those infrastructure areas which bring a sentiment for
the private sector to invest more in those profitable business areas.
From the analysis of private investment determinants, it is found that domestic financial availability
like domestic savings and the availability of credit to private sectors play a stimulant effect on
promoting private investment in the country. Therefore there should be more concern to strengthen the
habit of savings on one hand modernization and widening of the scope of credit availability in the
country at a greater extent. This can be best dealt with a well-designed interest rate policy.
There should be an increasing effort for debt relief and cancellation to reduce debt burdens and to
increase the level of investment for a given level of future indebtedness.
Lastly, instability in the economy like changes in the inflation rate and the real interest rate has a
deterrent impact on private investment. Uncertainty in an economy hurts private investment at large.
Flourishing a stable economy through sound macroeconomic policies is crucial to raise private
investment in the country at large.

Reference
Acosta, P., & Loza, A. (2005). Short run and Long run Determinants of Private Investment in
Argentina. Jornal of Applied Economics, Vol. VIII(No.2), pages389-406.
Afonso, A., & St.Aubyn, M. (2008). Macroeconomic Rates of Return of Public and Private
Investment: Crowding-in and Crowding-out Effects. Working Papers, Technical University of Lisbon,
Department of Economics.
Apergis, N. (2000, July). Public and Private Investments in Greece: Complementary or Substitute
‘Goods’? Bulletin of Economic Research, Vol. VXII(No. 3), pages 225-234.
Asante, Y. (2000). Determinants of private investment behaviour. University of Ghana Legon,
Department of Economics. Nairobi: African Economic Research Consortium.
Aschauer, D. A. (1998). How big should the public capital stock be?: the relationship between public
capital and economic growth. (W. Karl, Ed.) Th e Public Policy Brief Series, Vol. XIV(No. 1), pages
7-33.
Aysan, A. (2006). How to Boost Private Investment in the MENA Countries: The Role of Economic
Reforms. University of Maryland, Department of Economics. Annapolis: College Park.
Coutinho, R. M., & Gallo, G. M. (1996). Do public and Private Investment Stand in Each Other`s
Way? Washington DC: World Bank.
Cruz, B. d., & Teixeira, J. R. (1999, April). The impact of public investment on private investment in
Brazil, 1947-1990. CEPAL Review, Vol. XXI(No.67), pages 75-84.
Daniel, Z. (2005). Determinants of private investment in Ethiopia. Adiss Ababa University,
Department of Economics. Addis Ababa: Adiss Ababa University.
Erden, L., & Holcombe, R. G. (2005, September). The Effects of Public Investment on Private
Investment in Developing Economies. Public Finance Review, Vol. 33(No. 5), pages 575-602.
Greene, J., & Villanueva, D. (1991). Private Investment in Developing Countries: An Emperical
Analysis. IMF Staff Papers, Vol. 38(No. 1), pages 321-348.
Harris, R. (1995). Using Co integration Analysis in Econometrics Modeling. New York: Prentice Hall.
Hatano, T. (2010, February). Crowding-in Effect of Public Investment on Private Investment. Public
Policy Review, Vol. VI(No. 1), pages 105-120.
Heng, T. K. (1996). On the Crowding in of private capital by public capital. Nanyang Technological
University, Nanyang Business School. Tokyo: EPLA Preess.
Hosamane, D. M., & Niranjan, R. (2010, March 1). Determinants of investment pattern in Indian
manufacturing industries a panel data study. Retrieved February 28, 2011, from Free Library:
http://www.thefreelibrary.com/Determinants+of+investment+pattern+in+Indian+manufacturing+indus
tries...
Karim, S., Hassan, J., & Zain, M. (2007). Private and public investment in Malaysia: Substitutability
or complementarity? The IUP Journal of Applied Economics, Vol. VI(No. 2), pages 17-34.

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  37  
Kings’ori, Z. (2007, September 11). Factors Influencing private investment in Kenya, Makerere
University, January. Retrieved February 14, 2011, from Google:
http://www.scribd.com/doc/2369078/Private-Investments-in-Kenya
Kollamparambil, U., & Nicolaou, M. (2011, February). Nature and association of public and
privateinvestment: Public policy implications for South Africa. Journal of Economics and International
Finance, Vol. III(No. 2), pages 98-108.
Ku-tepeli, Y. (2005, july 25). Effectiveness of Fiscal Spending: Crowding out and/or crowding in?
Retrieved January 21, 2011, from Google: http://www.bayar.edu.tr/~iibf/dergi/pdf
/C12S12005/YK.pdf
Lesotlho, P. (2006). An investigation of the determinants of private investment: The Case of
Botswana. University of the Western Cape, Department of Economics. University of the Western
Cape.
Monadjemi, M. S., & Huh, H. (1998). Private and Government Investment: A Study of Three OECD
Countries. International Economic Journal, Vol.3(No.2), 93-105.
Ribeiro, M. B., & Teixeira, J. R. (2001). An econometric analysis of private-sector investment in
Brazil. C E P A L Review, vol. 1(No.74), 154-166.
Sackey, H. A. (2006, January 13). Private Investment for Structural Transformation and Growth in
Africa: Where do Small and Medium-Sized Enterprises Stand? Retrieved 2011, from
http://www.uneca.org/aec/documents/Harry%20A%20Sackey.pdf.
Sadraoui, T., & Chockri, A. (2010, Novmber). A dynamic panel data analysis for relationship between
private and public investment in R and D. Journal of Economics and International Finance, 2(11), 278-
289.
Seruvatu, E., & Jayaraman, T. K. (2001). Determinants of Private Investment Fuji. Reserve Bank of
Fiji, Economics Department. Suva: Reserve Bank of Fiji.
Serven, L. (1996). Macroeconomic uncertainty and private investment in developing countries - an
empirical investigation. The World Bank.
Tafere, K. (2008). The Sources of The Recent Inflationary Experience In Ethiopia. Thesis, Addis
Ababa University School of Graduates, Economics Department, Addis Ababa.
Workie, M. (1996, October). Determinants and Constraints of Private Investment in Ethiopia. (A.
Geda, Ed.) Ethiopian Journal of Economics, Vol. V(No. 2), pages 57-81.

www.theinternationaljournal.org  >  RJEBS:  Volume:  04,  Number:  06,  April-­‐2015                                                                            Page  38  
View publication stats