Between 1980 and 2020, the study looks at "An Empirical Analysis of the Effects of Population Growth on Economic Growth in Ethiopia Using an Auto Regressive Distributive Lag (ARDL) Model Approach." The appraisal coefficient of population growth (POP) is positive and significant, according to the findings of this study. However, in response to the long-term association between population expansion and economic growth in Ethiopia, the macroeconomic variables were subjected to a limit test and a broader causality test. From the finding of bound test, since the value of F-statics is greater than the upper boundary line then there are long run equilibrium relationships between RGDP, Population size, foreign direct investment, personal remittance, population growth rate, rate of inflation and gross capital formation. According to the findings of the Grander Causality Test, Real Gross Domestic Product can cause Ethiopian population size (POP), but population number (POP) cannot cause Real Gross Domestic Product at the same time. Finally, in order to boost Ethiopia's economic growth, the government should implement policies that would attract foreign investors. The government should also establish a benchmark to ensure that the economy increases faster than the population.
Research
An Empirical Analysis of the Effects of Population Growth on Economic Growth in Ethiopia Using an Auto Regressive Distributive Lag (ARDL) Model Approach
https://doi.org/10.21203/rs.3.rs-646003/v2
This work is licensed under a CC BY 4.0 License
Version 2
posted
You are reading this latest preprint version
ARDL
Economic growth
Population growth
The world's population was estimated to be around six billion people as the twenty-first century began. According to UN projections, the population will reach more than nine point two billion by 2050, with a maximum of eleven billion by 2200. Over 90% of that population will live in underdeveloped countries (Todaro and Smith, 2006).
According to Anulawathie Menike (2018), population growth and economic development in a country have a close and reciprocal link. In one sense, the population is a supply of labor that could be used to boost the country's output. It might also be viewed as a consumer group that consumes and depletes a significant portion of the country's resources. Certain economists from the past have noted that a country's population increase and quick growth is linked to its economy.
Gideon Thuku discussed the Malthusian hypothesis, which suggested that food problem is associated with a high population increase, i.e., malnutrition and hunger while Bloom and Freeman (1998) has a different view that food problem is more problem on poverty and inadequate income than high population growth. When people could buy then the food crisis can be overcome as prices give ample incentives to produce. In addition, to meet their rising demand for food, most developing economies would have to expand exports, seek foreign aid, or borrow from abroad. A high rate of population increase has an impact to render more intense on the constraints to the development of savings, foreign exchange, and human resources has an adverse impact on improvement in food supplies (Gideon T., 2013).
More than 75 million people are added to the world's population every year. Developing countries account for nearly all of this net population gain (97 percent). Although such large increases are unusual, the challenge of population expansion is more than just a matter of numbers. It's a question of human well-being and development. Rapid population expansion has the potential to have major ramifications for humanity's overall well-being. If development entails the improvement in people’s levels of living their incomes, health, education, and general wellbeing and if it encompasses their capabilities, self-esteem, respect, dignity, and freedom to choose (Todaro and Smith, 2012).
Ethiopia's population is growing at an incredible rate. Ethiopian populations were estimated to be 11.75 million in the early twentieth century, with a 0.2 percent annual growth rate, and expected to double in 346 years. Ethiopia's population, on the other hand, has more than doubled in the last 60 years. Between 1960 and 1990, Ethiopia's population doubled. According to population census data from 1984, 1994, and 2007, Ethiopia's population was 39.37 million, 55.18 million, and 80.67 million, respectively. As a result, it will only take 27 years to double the population growth from the 2007 census report.
Kassahun Alemu (2014); uses a time serious data on “the impacts of population growth on the Ethiopian performance from 1970/71 up to 20012/13”. From the general regression results, He was concluded that, “population issue was very controversial because it has negative impact in the short run, but the implication is positive in the long run”. This invites to investigate other variables like investment, age dependency ratio and labour force that affect economic growth if its correlation can be positive in the long term.
In Ethiopia, monetary policy is a tool used by the government or the Central Bank of Ethiopia to achieve a set of goals aimed at promoting economic growth and stability. The National Bank of Ethiopia's key responsibilities in Ethiopia include monetary policy management and financial sector stability (NBE). With changing economic dynamics within the country and increased empirical and theoretical understanding of monetary policy around the world, Ethiopian monetary policy and the process of formulating it has changed (NBE, 2009).
Population increase and per capita output growth do not appear to be independent, according to the literature. The nature of their interaction appears to be highly dependent on the conditions and population age structure in each country and region. Because of Japan's aging population, a smaller cohort of working-age individuals will be required to sustain increasing numbers of pensioners and slower economic development unless productivity and per-capita output increase significantly. In many African countries, a new form of dependency problem arises, where a small working-age population is required to support a huge number of children with significant educational and health demands. When these youngsters join the labor force in the future, economic growth should accelerate (E. Wesley F. Peterson, 2017).
The trends suggest that the interaction of population pressures and the economy is a very important issue and may have contributed to perpetuation of poverty trap in Ethiopia. Specifically, this study tries to know the effects of total population on economic growth analyze the short run and long run relationship of economic growth.
1.2. Statement of the Problem
Ethiopia's population is growing at an incredible rate. Around 20th century, the number of Ethiopian populace considered to be 11.75 million with 0.2 percent annual growth rate and it anticipated to be doubled in 346 years. Ethiopia's population, on the other hand, has more than doubled in the last 60 years. Between 1960 and 1990, Ethiopia's population doubled. According to population census data from 1984, 1994, and 2007, Ethiopia's population was 39.37 million, 55.18 million, and 80.67 million, respectively. Therefore, in order to double the population growth from 2007 census report it just requires 27 years.
Kassahun Alemu analyzes "the impacts of population expansion on Ethiopian performance from 1970/71 to 20012/13" using time-series data. He concluded from the overall regression results that the population issue was very contentious because it has a negative influence in the near run but has a beneficial long-term implication. This opens the door to looking at other factors that affect economic growth, such as investment, the age dependency ratio, and the labor force, to see if the correlation can be favorable in the long run.
Gideon Thuku (2013) used data from 1963 to 2009 in his study. He used a VAR model in which he included explanatory variables such as real GDP growth, total population, population density, and urban population growth. The estimated Augmented Dickey Fuller test statistic value is bigger than the critical value at 95 percent confidence level, implying that both economic and population growth rates are stationary at levels, according to the study's empirical findings. He also omitted variables such as personal remittances and the rate of inflation, both of which are important in the analysis of population and economic growth. There is still a lot of disagreement about the effects of population growth on economic growth. As a result, the argument over the beneficial and negative effects of population expansion on the economy continues. On the plus side, economists argue that increased population leads to technological advancements and innovations. This is because population expansion fosters business competitiveness, and as a country's population grows, so does the size of its potential market (Gideon Thuku, 2013).
There appears to be considerable agreement in the literature that population increase and per capita production growths are not mutually exclusive. The nature of the interaction between them seems to be that it varies very much on the conditions and the age structure of the population in the various countries and areas. Because of the aging population in countries like Japan, a smaller cohort of working-age people will be required to support growing numbers of retirees and slowing economic growth unless productivity and per capita output rise significantly. In many African countries, a new form of dependency problem arises, where a small working-age population is required to support a huge number of children with significant educational and health demands. When these youngsters join the labor force in the future, economic growth should accelerate.
This rapid population growth may pose a threat unless the country's economic performance compensates. Econometric research has indicated that, this population growth has had a major negative impact in the short run but a favorable impact in the long run on the economic performance of the country (Kasahun Alemu, 2005).
According to the trends, the combination between population pressures and the economy is a major issue that may have led to Ethiopia's poverty trap perpetuation. As a result, the purpose of this article is to investigate the relationship between population growth and economic growth. If the population continues to increase at a rapid rate, so will economic growth.
1.3. Objective of the Study
1.3.1. General objective
This study looked into the "Effects of Population Growth on Economic Growth in Ethiopia Using the ARDL Model," with a focus on total population, real GDP, personal remittance, population growth rate, net inflow of foreign direct investment, inflation rate, and gross capital formation.
1.3.2. Specific objectives
-
To understand the effects of total population on economic growth.
-
To analyze the short run and long run relationship of economic growth with respect to population growth.
-
To identify the causal relationship between population growth and economic growth.
2.1. Theoretical Literature Review
Population of an area is the total number of all individuals that live in a caution point in time. Population of a country is the number of people and a tall account is taken of differences of quantity; it is of quality in close relation to numbers. Thus, population can be regarded as a label for a human aggregate (Thomas Frejka, 1973).
Real Gross Domestic product (RGDP) is an inflation-adjusted measure of the value of all goods and services produced by an economy in a given year (expressed in base-year prices) and is often referred to as “constant-price”, “inflation-corrected”, or constant dollar” GDP. Real Gross Domestic product (RGDP) makes comparing GDP from year to year and from different years more meaningful because it shows comparisons for both the quantity and value of goods and services (Akhilesh and Eric, 2020).
Gross Domestic Product (GDP) is the most important measure of the productive capacity of an economy and GDP can be measured in three different ways. The Commerce department measures GDP using three methods: the income method, the expenditure method and the product (or value added) method. The gross domestic product (GDP) is the value of all final goods and services produced. Final goods are those that are sold directly to final users as opposed to intermediate goods that are produced by one firm and used as an input by another (Roger Farmer, 1997).
Malthus model of population growth
Gideon Thuku discusses Malthus’s theory which believed that the world's population tends to increase at a faster rate than its food supply. Whereas population grows at a geometric rate, the production capacity only grows arithmetically. Therefore, in the absence of consistent checks on population growth, Malthus made the prediction that in a short period, scarce resources will have to be shared among an increasing number of individuals. The positive check to population is a direct consequence of the lack of a preventive check. When society does not limit population growth voluntarily, diseases, famines and wars reduce population size and establish the necessary balance with resources (Thomas Malthus, 1826).
Boserup’s Model
This model has been proposed by Easter Boserup (1965), who has rejected classical (Malthusian trap) theory and rather basically argued that we would never out strip out food supply. She considers population growth as autonomous force of exogenous factor that causes to technological progress in agriculture and postulates that aggregate agricultural production function in the long run will always shift up wards in response to population pressure to maintain output per capita. Generally, the reason why she rejects classical extensive margin is that land productivity can increase due to factors. Firstly, she argues increases in physical labor input to work longer hours will give increased output per hectare. Secondly, in the long-term technological progress would lead increased output per hectare that results in to shift in the production function (see, Dejene, et al., 2001).
2.2. Empirical Literature Review
Many researchers have been stated the relationship between population growth and economic growth as below.
Dao (2012) studied on the relationship between population and economic growth in developing countries. Dao adopted a methodology of least square estimation technique in a multivariate liner regression. From the result of the estimation, population growth has appositive impact on overall economic growth. However, the coefficient is always smaller than one suggesting that the additional people have a less than proportionate influence on economic growth.
Bucci (2003) investigated whether there is a long-run relationship between population (size and growth) and per-capita income focusing on human and physical capital as reproducible inputs. The study found out that population growth exerts a negative effect on economic growth. The study also extends edits analysis to the case where physical and human capital can interact with each other in the production of new human capital. If physical and human capitals are substitutes for each other, the increase of population size, together with the reduction of the aggregate human capital stock, determines an unambiguous decline of the per-capita level. On the other hand; if physical and human capital are complementary for each other, the final effect on the per-capita level of skills.
Eli H. Tartiyus et al. (2015) conduct a study on the impact of population growth on economic growth in Nigeria from 1980 to 2010 using Error Correction model. The result of co-integration test shows the existence of long run equilibrium relationship between population, economic growth and other variables. The coefficients of population, fertility, and export have positive relationships with GDP Growth.
Shahjahan Ali et al. (2015) analyzed the empirical relationship between population growth and economic development of Bangladesh. The researchers were detected serial correlation by using the Durbin – Watson d statistics. From the result, the coefficient for population growth is negative, and significantly different from zero, meaning that population growth adversely affects the economic growth developed. These results demonstrate that rapid population growth is a real problem in Bangladesh.
Therefore, both the theoretical and empirical review shows that the topic, i.e. investigating the relationship between population growth and economic growth are sensitive issue especially countries like Ethiopia which has been recording an exponential population growth rate.
3.1. Data Type and Source
In this study, secondary data were used to analyze the relationship between variables considered. The time series data from 1980 to 2020 that are obtained from World Bank database of Ethiopia were used.
3.2. Data Analysis
After the collection of the secondary data; the model clearly specified to analyze the secondary data. In order to analyze the secondary data, the researcher used Eviews 10 software. Eviews 10 software is used for the flexible and comfortableness for tabulation, graphing and data analyzing process.
3.3. Definition of the Variables and Expected Sign
The dependent variable is Real Gross Domestic Product (RGDP) which is a measure of the total product of the country at a given period.
Total Population (POP)
Total population of a country consists of all persons falling within the scope of the census. This variable was expected to influence economic growth positively.
Foreign Direct Investment (FDI) net inflows refer to direct investment equity flows in the reporting economy. It is the sum of equity capital, reinvestment of earnings, and other capital. This variable was expected to have a positive effect on economic growth.
Personal Remittance (PRem)
Personal remittances comprise personal transfers and compensation of employees. This variable was expected to have a positive effect on economic growth.
Population Growth Rate (POPGR)
Population growth rate is calculated as the division of birth deaths over population size. The population growth rate was expected to have a positive effect on the economy.
Rate of Inflation (RI)
Inflation is a continuous increase in the overall price level in a country and measured in percent. This variable was expected to have a negative impact on Real GDP.
Gross Capital Formation (GCF)
measures the value of acquisitions of new or existing fixed assets by the business sector, government, and “pure” households or excluding their unincorporated enterprises less disposals of fixed asset. This variable is expected to influence economic growth positively.
3.4. The Autoregressive Distributed Lag (ARDL) Model
ARDL model is used in order to undertake this study. We applied ARDL model because it allow capture sufficient number of lags in data generating process. ARDL technique applied irrespective of whether the underlying variables are I(0) or I(1) or a combination of both but not I (2).This is helped to avoid the pretesting problems associated with standard co-integration analysis which was required the classification of the variables into I(0) and I(1). Since each of the underlying variables stands as a single equation, endogeneity is less of a problem in the ARDL technique because it is free of residual correlation. The major advantage of this approach was laid in its identification of the co-integrating vectors where there are multiple co-integrating vectors.
The equation expressed as follows: -
RGDP = f (POP, FDI, PREM, POPGR, RI, GCF)……………………….. (3.2)
lnRGDPt = β0 + β1lnPOP + β2FDI + β3PREM + β4lnPOPGR + β5RI + β6lnGCF + εt ….... (3.3)
Where β0…β7 are parameters of interest which are to be estimated and
RGDP = Real Gross Domestic Product
POP = Total population
FDI = Foreign Direct Investment net inflow
PREM = Personal Remittance
POPGR = Population Growth Rate
RI = Rate of Inflation and
GCF = Gross Capital Formation
Real Gross Domestic Product (RGDP), Total Population, Population Growth Rate and Gross Capital Formation have been used as a natural logarithm. Four of them are stationary after I have changed them to logarithm form.
3.5. Time Series and Econometric Issues
3.5.1. Stationary Test
Augmented Dicky Fuller (ADF) test was useful to know whether there exists a unit root of time series data. If the secondary data is stationary at level, then it is integrated of order zero I(0). If data is not stationary at level, we need to difference to make it stationary (Gujurati, 2004).
Augmented Dicky Fuller (ADF) test was useful to know whether there exists a unit root of time series data. If the secondary data is stationary at level, then it is integrated of order zero I (0). If data is not stationary at level we have taken first difference, second difference and so on until it become stationary (Gujurati, 2004).
The procedure of stationary testing using the ADF is similar to that of the DF test but rather applied to a particular model as specified below.
∆Yt = α + δt + λYt-1 + B1∆Yt -1 +… + Bj-I ∆Yt -j+1 + µ t …………………….…………. (3.6)
Simply, ∆Yt = α + δt + λYt−1 + \(\sum _{t=1}^{i}\beta 1\)+ ∆Yt−1 + µ t ………………………………………… (3.7)
Where α is a constant term, δ is the coefficient of the time trend, j is the optimal lag length, ∆ is the difference operator, t represents the time trend and µ represents the Gaussian white noise. The test for stationary was carried out under the null hypothesis λ = 0 as against the alternative hypothesis λ < 0.
Testing for stationarity of time series was mandatory to get a meaningful regressing analysis. According, the assumption of the neoclassical regression model necessitates that the variable was stationary when error term a zero mean and a constant variable.
3.5.2. Lag Selection Criteria
The issue of finding the appropriate lag length for each of the underlying variables in the ARDL model was very important. Because: we were needed to have Gaussian error terms (i.e. standard normal error terms that do not suffer from non normality, autocorrelation, etc). In order to select the appropriate number of lags for the long run underlying equation, the researcher determined the optimal lag length (k) by using the lower values of Akaike Information Criteria (AIC), Schwarz Bayesian Criteria (SBC) or Hannan-Quinn Criteria (HQC) (Nkoro & Kelvin, 2016).
We specify those common information criteria as follows:
$$AIC\left(m\right)=lo\text{g}\left|{\widehat{{\Sigma }}}_{m}\right|+\frac{2}{T}{mk}^{2} \left(1\right)$$
$$SC\left(m\right)=\text{log}\left|{\widehat{{\Sigma }}}_{m}\right|+ \frac{logT}{T}{mk}^{2} \left(2\right)$$
$$HQC\left(m\right)=-2\text{log}\left|{\widehat{{\Sigma }}}_{m}\right|+ 2klog(\text{l}\text{o}\text{g}(T\left) \right(3)$$
Where \({\sum }_{m}\) is likelihood value, T is number of observation, while m is the order and k is number of variables.
Given a set of candidate values for the data, the preferred value is the one with the minimum the value of AIC, SC and HQC always suggests the best lag length to be chooses in each model.
3.5.3. The Long Run ARDL Model
To examine the existence of the long run relationship between the variables the following version of ARDL model was used.
To solve the problem of Spurious Regression; the research changed Real Gross Domestic Product, Total Population, Population Growth Rate and Gross Capital Formation in to a natural logarithm. Four of them are stationary after the researcher were changed them to logarithm form.
3.5.4. Short Run and the Error Correction Model (ECM)
Once the long run model was estimated, the next task was to model the short run dynamics of the model by estimating an Error Correction Model associated with the long run estimates. This was specified as follows:-
\({ECM}_{t-1}\)is the error correction term that will be obtained from ARDL long run dynamics of the model and it is expected to have negative sign. It indicates the speed of adjustment back to long-run equilibrium after a short-run shock.
3.5.5. Bound Test
The general version of the ARDL bound test modeling was required for the construction to check the long run relationships. The following unrestricted error correction model: -
∆Xt = δ0i + \(\sum _{i=1}^{h}a\)1 ∆Xt−1 + \(\sum _{i=0}^{h}a\)2 ∆ Yt−1 + δ1 Xt−1 + δ2 Yt−1 + µ1t --------------------------- (6)
Where m is the ARDL model maximum lag order, ∆ represents the first difference operator, Xt vector of dependent variables, Yt vector of k determinants of Xt and u1t refers to residual error term which was assumed to be white noise error, having zero mean and constant variance.
The ARDL bound approach to co-integration was involved testing the presence of long run relationship among the variables. In this approach long run and short run parameters of the model was estimated using Error Correction Model. ARDL formulation can be written as follow: -
∆lnRGDPt = βo + \(\sum _{\text{i}=1}^{\text{n}}{\beta }\)1t∆lnRGDPt−1+ \(\sum _{\text{i}=1}^{\text{m}}{\beta }\)2t∆lnPOPt−1 + \(\sum _{\text{i}=1}^{\text{m}}{\beta }\)3t∆FDIt−1 + \(\sum _{\text{i}=1}^{\text{m}}{\beta }\)4t∆PREMt−1 + \(\sum _{\text{i}=1}^{\text{m}}{\beta }\)5t∆lnPOPGRt−1 + \(\sum _{\text{i}=1}^{\text{m}}{\beta }\)6t∆RIt−1 + \(\sum _{\text{i}=1}^{\text{m}}{\beta }\)7t∆lnGCFt−1 + 𝛼1lnRGDPt−1 + 𝛼2lnPOPt−1 + 𝛼3FDIt−1 + 𝛼4PREMt−1 + 𝛼5lnPOPGRt−1 + 𝛼6RIt−1 + 𝛼7lnGCFt−1 + εt ............... (7)
The null hypothesis that the coefficients of the lag level variables are zero and alternative hypothesis of the lag level variables are non zero. The null of non existence of the long-run relationship is defined by:-
HO: α1 = α2 = α3 = α4 = α5 = α6 = α7 = α8 = 0
H1: α1 ≠ α2 ≠ α3 ≠ α4 ≠ α5 ≠ α6 ≠ α7 ≠ α8 ≠ 0
The rejection or acceptance of the null hypothesis was based on the whether or not the Pasaran F-statistic critical values falls above or below the upper and lower critical value bounds Pesaran et.al (2001).
3.5.6. Causality Test
Granger causality was used to determine whether one time series is useful in forecasting another (Enders 1995). Granger causality was tested the direction of causation between population growth and economic growth. Auto regressive distributed lag model was used to perform Granger causality tests. Thus Granger causality test was applied.
RGDP = αo + \(\sum _{i=1}^{p}a\)1i LNPOP + \(\sum _{i=0}^{q1}a\)2iFDI + \(\sum _{i=1}^{q2}a\)3iPREM + \(\sum _{i=0}^{q3}a\)4i LNPOPGR + \(\sum _{i=1}^{q4}a\)5i RI + \(\sum _{i=0}^{q5}a\)6iLNGCF + µ1t …………………………………………………………..……. (8)
3.5. Diagnostic Tests
A) Normality test
Since a number of the most common statistical tests rely on the normality of a sample or population, it was often useful to test whether the underlying distribution is normal, or at least symmetric. To see whether the values of the data’s are normally distributed or not; for this thesis Jarque - Bera test statistic of normal distribution was used.
B) Serial correlation Test
Error terms from different time periods or cross-sectional observations was correlated; Serial correlation occurs in time-series studies when the errors was associated with a given time period carry over future time periods. There are different types of serial correlation such as first-order serial correlation, in which the errors in one time period are correlated directly with errors in the consequent time period and the positive serial correlation in which errors in one time period are positively correlated with errors in the next time period. (Williams, 2015)
In order to detect whether the serial correlation exists in the given econometric model or not, the Breusch-God Frey Lagrangian Multiplier (LM) test was used.
C) Model Specification Test
Model specification was detected using Ramsey Reset test.
3.6. Stability Test
Pesaran and Pesaran (1997) suggest using Brown et al. (1975) stability test also known as cumulative (USUM) and cumulative sum of squares (CUSUMSQ) test.
4.1. Descriptive Analysis
From Figure 1, we observed that Real Gross Domestic product and number of population have different rate of growth. Specifically, the number of population has almost the same rate while the rate of Real GDP increases with constant phase and dramatically changes.
In 1980, the numbers of population in Ethiopia were around 35 million people and 9 years later in 1988 the number of population were increased to 45 million people. Then, the number of population was reached 55 million within 6 years in 1994; consequently, in 2002 the number of population increased 70,142,091.
The Real GDP of Ethiopia in 1980 was 108,920 in constant prices and was reached 310,115 in constant price in 2006. Due to the millennium development goal and political strength; the GDP of Ethiopia were again doubled within two years and reached 626,977 in constant price. In 2015 again the Gross Domestic Product were reached 1,449,397.
In Figure 2, we discusses about population growth rate where population growth rate is calculated as the division of birth deaths over population size. In 1980, the population growth rate of Ethiopia was around 1.88% and in 1992 the population growth of Ethiopia rate was doubled and becomes 3.6%. Starting from 1992 the population growth rate of Ethiopia was gradually going down to 3.57% in 1993, 3.47% in 1994, 3.32% in 1995, and 3.16% in 1996.
4.2. Results from Stationarity Test
If the variables are not stationary at level then apply difference to make stationary.
From Table 1; RI and PREM are stationary at level, while LNRGDP, LNPOP, FDI, LNPOPGR and LNGCF are stationary after first difference as indicated by ADF test statistic values which is greater than the corresponding critical values. Thus, we need to conduct an ARDL based co-integration (bound) test since variables are integrated at different order, i.e. I (0) and I (1).
|
Variables |
Values |
ADF test at level I(0) |
ADF at first difference |
Order of Integration |
|||
|
Intercept |
Trend and Intercept |
None |
Intercept |
Trend and Intercept |
|||
|
LnRGDP |
ADF test-static |
3.84 |
1.03 |
3.69 |
5.19 |
5.82 |
I(1) |
|
5% Critical Value |
2.94 |
3.53 |
1.95 |
2.94 |
3.53 |
||
|
LnPOP |
ADF test-static |
2.62 |
1.73 |
14.69 |
17.72 |
7.14 |
I(1) |
|
5% Critical Value |
2.95 |
3.53 |
1.95 |
2.94 |
3.54 |
||
|
FDI |
ADF test-static |
2.03 |
5.32 |
2.91 |
3.39 |
4.19 |
I(1) |
|
5% Critical Value |
2.96 |
3.54 |
1.95 |
2.96 |
3.57 |
||
|
PRem |
ADF test-static |
3.74 |
4.90 |
3.08 |
5.26 |
5.15 |
I(0) |
|
5% Critical Value |
2.96 |
3.56 |
1.95 |
2.96 |
3.57 |
||
|
LnPOPGR |
ADF test-static |
3.21 |
3.08 |
0.76 |
4.10 |
4.09 |
I(1) |
|
5% Critical Value |
2.95 |
3.55 |
1.95 |
2.94 |
3.54 |
||
|
RI |
ADF test-static |
4.73 |
5.02 |
1.23 |
8.57 |
8.44 |
I(0) |
|
5% Critical Value |
2.94 |
3.53 |
1.95 |
2.94 |
3.54 |
||
|
LnGCF |
ADF test-static |
0.69 |
1.81 |
2.53 |
7.32 |
7.75 |
I(1) |
|
5% Critical Value |
2.94 |
3.53 |
1.95 |
2.94 |
3.53 |
||
| Mackinnon (1996) Critical Values | |||||||
Source: Authors own computation
4.3. Results of Lag selection criteria
The guideline for selecting the lag selection criteria is depends on the lower value of AIC (Akaike Information Criteria), SC (Schwarz information Criteria) and HQ (Hannan-Quinn information Criteria). Thus, the optimal lags are selected based on minimum value of those three information criteria. The VAR (Vector Auto regression) model is one of the methods of obtaining the number of lag selection criteria. The VAR model tells us to use lag 3 as the maximum lag for ARDL model because lag 3 has the lowest values of AIC. From the below table, lag 3 values of AIC (Akaki Information Criteria), SC (Schwarz information Criteria) and HQ (Hannan-Quinn information Criteria) are less than that of lag 2 and lag 1 values.
From the guide line of lag selection criteria, the lower the values of AIC, SC, and HQ then the better the model. Since the values of lag 3 of AIC is 43.47600*, SC is 50.24994* and HQ is 45.84028* were lower than lag 2 values of AIC (49.56121), SC (54.17980) and HQ (51.17322) and as well as lower than the values of lag 1.
Generally, from the VAR criteria lag 3 is the best number of lag and the better the model based on the guide line. In VAR criteria; the AIC (Akaki Information Criteria) SC and HQ values of lag 3 are lower than that of lag 2 and lag 1.
4.4. ARDL Long Run and Bound Test
Once the ARDL model was developed; then the next steps are identifying the bound test results, the long run values of coefficient and long run equations.
4.4.1. ARDL Bound Test
ARDL Bound Test is helped to know whether there is co-integration between the variables or not. THE model developed by Pesaranet al. (2001) can be applied irrespective of the order of integration of the variables (irrespective of whether regressors are purely I(0), I(1) or mutually co-integrated but not I(2).
|
F – Bound Test |
Value |
Null Hypothesis: No levels of relationship |
||
|---|---|---|---|---|
|
F-statistic |
17.90609 |
Asymptotic: n=1000 |
||
|
10% |
1.99 |
2.94 |
||
|
5% |
2.27 |
3.28 |
||
|
1% |
2.88 |
3.99 |
||
| Source: Authors own computation | ||||
From Table 2, since the values of Pasarian F-statistic are 17.90609 and the Pasarian upper bound critical values at 5% are 3.28 then the Pasarian F-statistic is greater than the Pasarian upper bound critical values at 5%. Therefore, the result shows that there are levels of long run relationships.
Generally, the researcher rejects the null hypothesis (there are no levels of relationships) and accepts the alternative hypothesis (there are levels of relationships). There have been existences of long run relationships between the variables. The Pasarian F-statistic values have way higher values than the Pasarian upper bound critical values at 1%, 5% and 10% level of significance.
4.4.2. ARDL Long Run
After checking the existence of long run relationships, the next step is estimation of the long run coefficients. In most of regression analysis, researchers have been predicting the values of the unknown dependent variables based on the known values of the independent variables. The result of the best-fitted ARDL model by information criteria is given in table below.
|
Levels Equation |
||||
|---|---|---|---|---|
|
Case 2: Restricted Constant and No Trend |
||||
|
Variable |
Coefficient |
Std. Error |
t-Statistic |
Prob. |
|
LNPOP |
0.390532 |
0.078782 |
4.957101 |
0.0001 |
|
FDI |
1.30E-10 |
2.18E-11 |
5.992954 |
0.0000 |
|
PREM |
0.000167 |
6.59E-05 |
2.529943 |
0.0204 |
|
LNPOPGR |
-0.137015 |
0.171129 |
-0.800652 |
0.4332 |
|
RI |
-0.011479 |
0.001740 |
-6.598554 |
0.0000 |
|
LNGCF |
0.495917 |
0.050657 |
9.789717 |
0.0000 |
|
C |
2.100256 |
1.127419 |
0.088925 |
0.9301 |
|
EC = LNRGDP - (0.3905*LNPOP + 0.0000*FDI + 0.0002*PREM -0.1370 |
||||
|
*LNPOPGR - 0.0115*RI + 0.4959*LNGCF + 2.1003 ) |
||||
| Source: Authors own computation from ARDL long run | ||||
From Table 3, the researcher developed the long run ARDL model equation as follows:
LNRGDP = 2.1003 + 0.3905* LNPOP + 1.30e-10 FDI + 0.000167 PREM – 0.137*LNPOPGR - 0.0115* RI + 0.4959* LNGCF
The number of population (POP), which is the main variable for this study, was found to have a positive and statistically significant at 1% significant level in explaining economic growth. As it expected; population size has a positive effect on economic growth of Ethiopia. From the result of long run ARDL model, population size is statistically significant at 1% level having probability values of 0.01%. We can interpret it, as a one percent increase in population number will result an increase in economic growth by 39.05 percent.
In this study as it anticipated; Foreign Direct Investment net inflow have a positive effect having 1.30e−10 coefficient and statistically significant having a probability values of 1%. Holding other thing constant; a one unit increase in the Foreign Direct Investment net inflows results in an increase in economic growth by 1.30e−10.
The personal remittance (PREM) coefficient is positive and statistically significant at 5% level of significant having p-values of 0.0204. Thus, personal remittance directly increases the economy of remitted person. Without the other factors, a one percent change in personal remittance will result to increase an economic growth by 0.0167 percent.
Gross Capital Formation has a positive effect on economic growth. Gross Capital Formation is statistically significant since probability value is 0%, which is, much less than the five percent level. In this study; when Gross Capital Formation increased by 1% then economic growth increased by 49.59 percent holding other things constant.
Inflation rate has a negative coefficient. It affects economic growth negatively because its probability value is less than 5% level. Rate of Inflation has a p-value of 0%: which is less than 5% level and it was statistically significant. Thus, a 1% increase in inflation rate will results in a decrease of economic growth by 1.1479%.
However, Population Growth Rate (POPGR) has a negative coefficient and also not statistically significant. Meaning that; Population growth rate does not affects economic growth because its probability values are 43.32% which is greater than 5% confidence level.
4.5. Results of Short Run and Error Correction Model
The next step was estimation of the short run coefficients a long with the short run Error Correction Term (ECT) term. Since, the long run model is clearly specified and the effect of the regressors on the regress and is clearly interpreted. Thus, the ECT shows the short run dynamics of the model be side with the long run adjustment.
From the long run model and the short run; foreign direct investment net inflow has a positive effect on economic growth. However, in the short run, FDI is not statistically significant having p-values of 25.98%. That means, foreign direct investment net inflow does not have affect economic growth in the short run.
|
Variable |
Coefficient |
Std. Error |
t-Statistic |
Prob. |
|---|---|---|---|---|
|
D(LNRGDP(-1)) |
0.437488 |
0.096066 |
4.554035 |
0.0002 |
|
D(FDI) |
2.62E-11 |
2.26E-11 |
1.161598 |
0.2598 |
|
D(PREM) |
0.000226 |
2.73E-05 |
8.306190 |
0.0000 |
|
D(RI) |
0.002391 |
0.000611 |
3.910762 |
0.0009 |
|
D(RI(-1)) |
-0.007993 |
0.001024 |
-7.808011 |
0.0000 |
|
D(RI(-2)) |
-0.003631 |
0.000712 |
-5.102695 |
0.0001 |
|
D(LNGCF) |
0.159154 |
0.032433 |
4.907130 |
0.0001 |
|
D(LNGCF(-1)) |
-0.380966 |
0.062225 |
-6.122378 |
0.0000 |
|
D(LNGCF(-2)) |
-0.214824 |
0.036065 |
-5.956529 |
0.0000 |
|
CointEq(-1)* |
-0.584298 |
0.082597 |
-14.00088 |
0.0000 |
|
* p-value incompatible with t-Bounds distribution. |
||||
| Source: Authors own computation | ||||
In the short run model, the first lag difference of RGDP is statistically significant having p-values of 0.02%, which is less than 5%. Allowing everything constant; when the short run coefficient of real gross domestic product could increase in one percent then economic growth increase by 43.7488 percent.
From the long run model, personal remittance has a positive effect on economic growth and its short run effect is also positive. From the above short run result of Auto Regressive Distributive lag (ARDL) model, personal remittance is significant having p-values of 1%percent, which is, much less than 5% levels and have a positive coefficient. Other things remaining constant, it tells us that when personal remittance increases by 1% then economic growth increase by 0.0226%.
The short run rate of inflation has positive and statistically significant affect economic growth. From most of the theory and general truth, rate of inflation affects economic growth negatively. Nevertheless, in the short, rate of inflation has a positive effect on economic growth. Assume other thing constant, 1% increase in inflation rate would account for 0.2391 percent increase in the economic growth of Ethiopia.
Gross capital formation has a positive effect on economic growth and statistically significant having p-values of 0.01%. A 1 unit increase in gross capital formation can lead to an increase in economic growth by an amount of 0.159154.
The Error Correction Term (ECT), which is estimated -0.584298 is negative as it was known and it is statistically significant. Thus, it indicates that 58.42% of the disequilibrium is adjusted before the next time period (year).
Generally, in the short run foreign direct investment, personal remittance, rate of inflation and gross capital formation have a positive coefficient.
4.6. Granger Causality Test
In chapter three, it was discussed that Granger Causality test is a test which enable us to identify the direction of the causality from where the lagged values of one variable in the past or present causes the future and; where the present and the future values responds to the cause in the past.
From Table 10; since the associated probability values of population growth to RGDP are 13.67% which was greater than 5%. Therefore, we cannot reject the null hypothesis rather we can accept the null hypothesis which said POP does not Grander Cause RGDP. On the other hand the associated probability values of RGDP to POP are 0.13% which was less than 5%. Therefore, we can reject the null hypothesis and we can accept the alternative hypothesis which was stated as RGDP does Grander Cause POP. Generally, there exists a uni-directional flow of causality from Real Gross Domestic Product (RGDP) to Population Number or from economic growth to population growth.
|
Pairwise Granger Causality Tests |
|||
|---|---|---|---|
|
Date: 06/10/20 Time: 12:39 |
|||
|
Sample: 1980 2019 |
|||
|
Lags: 3 |
|||
|
Null Hypothesis: |
Obs |
F-Statistic |
Prob. |
|
POP does not Granger Cause RGDP |
38 |
2.11501 |
0.1367 |
|
RGDP does not Granger Cause POP |
8.13111 |
0.0013 |
|
| Source: Authors own computation | |||
To summarize, the results from the causality test of the model for the period under investigation shows that an increased inflow of Real Gross Domestic Product (RGDP) can lead to an increase in population number (POP) of Ethiopia, while the increase in population number (POP) cannot lead to an increase in Real Gross Domestic Product (RGDP) at the same time.
4.7. Results of Model Diagnostic Tests
After estimation of long run, and short run then the next is the crucial step in any model was the application of the model diagnostic tests on the estimates of the residual of the ECM model. Diagnostic checks are very important in this analysis because if there is difficulty in the residuals from the estimation of an ARDL model, it is an indication that the model is not efficient.
i. Serial Correlation Test
From Breusch – Godfrey Serial Correlation LM Test; the researcher did not reject the null hypothesis. The guideline is that if the probability values are greater than 5% then we cannot reject the null hypothesis. Since the Breusch – Godfrey Serial Correlation LM test has probability chi-square (3) values of 9.92%, which is more than 5% level.
|
Breusch-Godfrey Serial Correlation LM Test: |
||||
|---|---|---|---|---|
|
F-statistic |
2.197354 |
Prob. F(3,14) |
0.1338 |
|
|
Obs*R-squared |
11.52455 |
Prob. Chi-Square(3) |
0.0992 |
|
| Source: Authors own computation | ||||
ii. Normality Test
From the Figure 3, the ARDL) model is well behaved and normal since its p-values (5.9685%) are greater than the 5% then we can deduce that the model is normal.
iii. Model Specification Test
From the below Table 6, the Ramsey RESET test also indicates that there is no specification error rather the ARDL model is stable since we accept the null of no model specification problems given its p-values (55.8) are greater than the 5%.
|
Ramsey RESET Test |
|||
|---|---|---|---|
|
Value |
Df |
Probability |
|
|
t-statistic |
0.594892 |
22 |
0.5580 |
|
F-statistic |
0.353897 |
(1, 22) |
0.5580 |
|
F-test summary: |
|||
|
Sum of Sq. |
Df |
Mean Squares |
|
|
Test SSR |
0.000771 |
1 |
0.000771 |
|
Restricted SSR |
0.048709 |
23 |
0.002118 |
|
Unrestricted SSR |
0.047938 |
22 |
0.002179 |
| Source: Authors own computation | |||
4.8. Stability Test
The researcher performed CUSUM tests of structural breaks for the long run equations to check the quality of the fitted ARDL model. Picture 4 shows that there were no structural breakdowns as shown in the figure. Since the CUSUM test of the residual lies in between the two critical lines; therefore, the parameters are stable. Finally, since the blue line lies within the two red lines then the model ARDL was more stable.
In this study, we investigated the effect of population size on economic of Ethiopia using an ARDL model approach from the period 1980 through 2019. From the result of ADF unit root test, inflation rate and personal remittance are stationary at level, while growth of real GDP, population size, FDI, population growth and gross capital formation are stationary after first difference.
From the result of ARDL model bound test result; the values of F-statistic (17.90609) value is greater than the upper boundary line (3.28) which indicates the existence of co-integration or long-run relationship between variables determined in this study.
From the result of the study, population growth has a positive effect on economic growth and statistically significant. This shows that an increase in population growth have an effect on the economic growth. From the result of fitted long run ARDL model, population growth was statistically significant at 5% level having probability values of 0.01%. A 1% increase in population number will result an increase in economic growth by 39.05 percent. In the long run, rate of inflation has a negative effect on economic growth. Both foreign direct investment net inflows and personal remittance are statistically significant and have a positive effect on economic growth. Moreover, the coefficient of error correction term is negative and statistically significant indicating any deviation from the long run equilibrium will adjusted by the speed of 58.42%.
Finally, the researchers recommended the concerned stakeholders that; the population size significantly affects the economic growth of the country; the government should plan a strategy, which will maintain the current positive effect of population size on economic growth of the country. In addition to that, by focusing on the activities, which are attractive for FDI that requires an encouraging environment and personal remittance, which have significance effect for economic growth on either solving shortage of foreign currency or as a means of capital for receivers.
Availability of data and materials
- The authors want to declare that they can submit the data at any time based on publisher’s request. The datasets used and/or analyzed during the current study were available from the authors on reasonable request.
Competing interests
- The authors have not declared any conflict of interest.
Funding
- The authors have contributed for the fund and there were no any institutions or university contribution for the support of the paper.
Authors' contributions
- The authors contributed from the beginning of the research up-to the end of the paper work and the contributions were full of confidence.
Acknowledgements
- My special thanks go to Almighty God for giving me ability and patience to accomplish this piece of work. My deepest appreciation and special thanks go to my major advisor, Dr. Berhanu Alemu and to my co-advisor Mr. Melkamu Woldie, for their invaluable and constructive comments on the manuscript of the research thesis development. Particularly, my warm thank goes to my Mother Mulu Regasa Negeri for her support and providing me morale in the one to be exemplary. My Abayinesh Regassa Negeri also deserve my gratitude for their warm appreciation. Last but not least, let me take this opportunity to forward my thanks to my friends, staff workers, sources of my secondary data organizations and relatives who supporting a lot directly or indirectly in my work of this manuscript.
Consent for publication: Not applicable.
- Akhilesh Ganti and Reviewed by Eric Estevez (2020) Definition of Real Gross Domestic Product: updated Mar 7, 2021
- Boserup E (1965) The Conditions of Agricultural.Growth,691–701
- Brown RL, Durbin J, Evans JM (1975) Techniques for testing the constancy of regression relationships over time. J Roy Stat Soc: Ser B (Methodol) 37(2):149–163
- Felmingham B, Jackson N, Weidmann K (2004) Population Neutralism: A Test for Australia and its Regions. University of Tasmania
- Bucci A (2008) Population growth in a model of economic growth with human capital accumulation and horizontal R&D. J Macroecon 30(3):1124–1147
- Tartiyus EH, Dauda TM, Peter A (2015) Impact of population growth on economic growth in Nigeria. IOSR Journal of Humanities and Social Science (IOSRJHSS) 20(4):115–123
- Nkoro E, Uko AK (2016) Autoregressive Distributed Lag (ARDL) cointegration technique: application and interpretation. Journal of Statistical and Econometric methods 5(4):63–91
- Obere A, Thuku GK, Gachanja P (2013) The Impact of Population Change on Economic Growth in Kenya
- Damodar N (2004) G. Basic econometrics
- Menike HA (2018) A literature review on population growth and economic development. International Journal of Humanities Social Sciences and Education 5(5):67–74
- Alemu K (2014) Impact of Population Growth on the Ethiopian Economic Performance. LAP LAMBERT Academic Publishing
- Malmberg B, Tegenu T (2007) Population pressure and dynamics of household livelihoods in an Ethiopian Village: an elaboration of the Boserup-Chayanovian framework. Popul Environ 29(2):39–67
- Malthus TR (2018) Essay on the Principle of Population. Yale University Press
- Dao MQ (2012) Population and economic growth in developing countries. International Journal of Academic Research in Business and Social Sciences 2(1):6
- Pesaran MH, Shin Y, Smith RJ (2001) Bounds testing approaches to the analysis of level relationships. J Appl Econom 16(3):289–326
- Peterson EWF (2017) The role of population in economic growth. Sage Open 7(4):2158244017736094
- Williams R (2015) Serial correlation. Notre Dame, USA: University of Notre Dame
- Farmer RE (1997) Money in a real business cycle model.Journal of Money, Credit, and Banking,568–611
- Ali S, Alam K, Islam S, Hossain M (2015) An empirical analysis of population growth on economic development: The case study of Bangladesh. International Journal of Economics, Finance and Management Sciences 3(3):252–259
- Todaro MP, Smith SC (2006) Economic development 8th edition. Manila, Philippines: Pearson South Asia Pte. Ltd
- Frejka T (1994) Long-range global population projections: Lessons learned. The future population of the world: what can we assume today, 3-15