Fixed Effect and Random Effect
Fixed Effect and Random Effect
0 Methodology
The Fixed Effect Model (FE) examines the relationship between indicator and outcome
parameters. The FE model shows the relationship among observation vary in the same direction
or observed the observation vary vary in one side not. The indicator factors may be influenced
by the unique characteristics of each chemical (for instance, being a male or female could impact
the assessment toward specific issue; or the political arrangement of a specific nation could have
some impact on exchange or GDP; or the strategic policies of an organization might impact its
stock cost). We recognize that something within the individual may affect or conditions the
indicator or result factors when using FE, and we want to account for this. This is why there is a
suspicion of a link between the substance's erroneous label and indicator factors.
The random coefficient model's premise of directionality between X and 0 is not required in the
fixed effects model. It's possible to write it as
Y ij =β 1+ β2 x ij + β 3 x ij Gij +σ j + μij 1
The σ jare between-context contrasts are supposed to be conventional Gaussian errors, while the
G ijare are considered to be conventional Gaussian errors. The additive term in G ij has been
removed from the randomized coefficient model, and the intercepting and slope can no longer
change randomly between contexts. In simple A fixed effects method is used to predict in which
the parameter estimates are fixed or non-random values, as defined in statistics. Random effects
models and mixed models, on the other hand, have all or some of the model parameters as
random variables. A fixed effect refers to a regression model wherein the group means are fixed
(quasi) as contrasted to a random effects in which the group means are a random sampling from
a population in numerous applications include econometrics and biostatistics. In general, data
can be classified based on a number of variables. For each grouping, the group means could be
described as fixed and random effects. Each group mean in a fixed - effect model is a cohort
fixed number. Fixed effects are the subject-specific means in panel data with longitudinal
observations for the same subject. Fixed effects estimator (also known as the within estimator) is
a term used in panel data analysis to refer to an estimator for the coefficients in the regression
model that include those fixed effects (one time-invariant intercept for each subject).
If the random coefficient model's assumption of orthogonality among x and 0 is implausible, as it
can be when the list of G i j variables is inadequate, the fixed effects model offers an alternative
that overcomes the problem. The fixed effects model in the Gaussian scenario is a standard linear
regression (Massin and Kop, 2014). The problem of none normality could be handled because
panel data is a combination of cross-section and time series. Checking the data for normality is
advised to ensure that your data analysis is accurate (Baltagi, 2021). In panel data analysis when
the number of T >10 then data is serious indulged in unit root to remove this we add more
variables in the model to eliminate the unit (Stock and Watson, 2012). But in our case T <10 and
we are including 9 variables in the model. In this analysis no need to check the stationary of the
data and we will directly approach to other necessary test such as OLS, Fixed effect model and
post estimation test.
We can examine the net influence of the indicators on the result variable by removing the impact
of those time-invariant features using FE. Another important assumption of the FE model is that
those time-invariant qualities are unique to the individual and should not be linked to other
personal characteristics. Because each substance is unique, the element's descriptive term and the
consistent that captures individual characteristics should not be linked to the others. If the
blunder terms are linked, then FE is not rational because derivations may not be correct and you
actually want to demonstrate that relationship, probably using arbitrary impacts; this is the basic
argument behind the Hausman test, which will be discussed later in this record. In this study we
choose the Fixe effect model because the value of Hausman test less than 0.5 percent. After this
we will check the model by using different test such as normality test and LM test to check
whether the residual of the model is correlated or not with each other. Further we used to check
which country get the more growth rate in terms of GDP or not.
2.0 Summary Statistics
In section we are going to discuss the growth indicator of Developing and Developed countries.
LOGFERTIL
CABALANCE FDI GDPPCG GFCAPITAL GOVCONS INFLATION ITY TRADE UNEMP
Mean 0.495984 1.835541 2.494893 24.29682 15.13281 35.41185 0.360470 49.20683 8.435182
Median -0.808336 1.652456 2.077486 22.01542 14.76167 6.161019 0.387921 49.81271 5.740000
Maximum 18.33270 6.464240 10.94373 43.71608 27.57302 1090.805 0.760498 95.74474 30.97200
Minimum -11.64663 0.008452 -9.042986 14.20328 6.647945 -0.310161 0.053846 16.52259 2.468000
Std. Dev. 4.633229 1.268896 3.258744 7.012473 4.799352 149.6685 0.147657 17.87698 6.856510
Skewness 1.398139 0.977460 -0.103545 0.859398 0.480398 6.269128 -0.048177 0.234817 2.173483
Kurtosis 6.810482 4.209553 4.914868 3.070286 2.646099 42.99199 2.996178 2.770930 6.871619
Jarque-
Bera 61.43210 14.53302 10.20142 8.137807 2.883026 4464.607 0.025571 0.750833 93.18526
Probability 0.000000 0.000699 0.006092 0.017096 0.236570 0.000000 0.987296 0.687003 0.000000
Sum 32.73492 121.1457 164.6630 1603.590 998.7654 2160.123 23.79100 3247.651 556.7220
Sum Sq.
Dev. 1395.343 104.6564 690.2619 3196.361 1497.196 1344040. 1.417169 20773.12 3055.763
Observatio
ns 66 66 66 66 66 61 66 66 66
The current account balance average mean value is 0.496 which is positive and indicate that the
developing countries current account improve in the last two decades. The mean value of foreign
direct investment is 1.835 which means that developing countries is emerging market and
attracting foreign direct investment in the past years. The gross fixed capital formation mean
value is also positive in the developing countries. The government spending rate is also positive
in developing or lower income countries in the last two decades. Due to increase in spending the
inflation rate mean value is positive which is equal to 35.412. The fertility rate in developing
countries average born baby per 1000 is equal to 0.360. The trade account balance mean value is
49.207 which show that the developing countries increase the export massively in the past 20
years. The last indicator unemployment rate which is important indicator and the mean value is
8.435 which show that the unemployment rate increases gradually past 20 years.
CABALANC LOGFERTILI
E FDI GDPPCG GFCAPITAL GOVCONS INFLATION TY TRADE UNEMP
Mean -0.254110 2.296394 1.043779 21.91177 19.10763 1.862081 0.207247 50.71249 7.320573
Median -0.615723 1.922245 1.156133 21.31271 19.28002 1.722899 0.203032 50.98751 7.154000
Maximum 7.752176 5.970383 3.286136 31.61752 23.94810 5.086918 0.323871 89.70895 11.84400
Minimum -5.048414 0.025914 -1.370308 16.22532 14.13553 -0.442322 0.079181 16.72361 2.580000
Std. Dev. 2.978032 1.580590 1.084899 3.129865 2.384955 0.973344 0.067592 19.14926 2.404613
Skewness 0.610706 0.557315 0.004398 0.773668 -0.208710 0.472898 -0.027343 0.171482 0.022118
Kurtosis 3.109362 2.477421 2.755790 3.895585 2.844469 4.528955 1.742405 2.303736 2.119579
Jarque-Bera 3.258246 3.283558 0.129385 6.925357 0.429931 7.003173 3.433163 1.305218 1.683712
Probability 0.196101 0.193635 0.937356 0.031346 0.806569 0.030150 0.179679 0.520686 0.430910
Sum -13.21373 119.4125 54.27652 1139.412 993.5967 96.82821 10.77684 2637.050 380.6698
Sum Sq. Dev. 452.3024 127.4115 60.02734 499.5989 290.0886 48.31732 0.233001 18701.40 294.8903
Observations 52 52 52 52 52 52 52 52 52
The average mean value of the current account balance is -0.254, which is negative and indicates
that the current account balance of developed countries has detreated during the last two decades.
Foreign direct investment has a mean score of 2.296, indicating that developed countries are
markets that have attracted opportunity for foreign direct investment in recent years. In advance
countries, the mean value of gross fixed capital formation is also positive and which is lower
than developing countries. In the previous two decades, government spending has increased in
developed countries as compared to developing countries. The inflation rate has a positive mean
value of 1.862 which is smaller than lower develop countries. In underdeveloped countries, the
average number of babies born per 1000 people is 0.356. In underdeveloped countries, the
average number of babies born per 1000 people is 0.207. The average value of the trade account
balance is 50.712, indicating that developed countries' exports have increased dramatically in the
last 20 years but lower than developing countries. The most recent indicator is the
unemployment rate, which is an important indicator with a mean value of 7.321, indicating that
the unemployment rate has gradually increased over the last 20 years but less than developing
countries.
Correlation
CABALANC GFCAPITA LOGFERTI
Probability E FDI GDPPCG L GOVCONS INFLATION LITY TRADE UNEMP
CABALANCE 1.000000
-----
UNEMP -0.156677 -0.087081 -0.229497 -0.423645 0.296249 -0.023583 0.180833 0.049484 1.000000
0.0990 0.3613 0.0149 0.0000 0.0015 0.8050 0.0564 0.6044 -----
The above table shows the correlation matrix which shows the relation among the variables and
the p value shows that the relationship among the variable is significant or not. The variables
government consumption and GDP growth rate is negative and significant relationship at 5
percent level. The gross capital formation shows the positive relationship with GDP and shows
the significant at 5 percent level. The inflation coefficient is negative and shows the significant
relationship at 10 percent level. The log fertility rate shows the negative relationship but
insignificant effect on the variable. The trade coefficient is positive but statistical insignificant at
5 percent level. Unemployment rate is negative effect on GDP and shows the significant effect at
5 percent level. Now we are going to estimate and classify the countries results through panel
regression.
C -3.04385 0.0814
The trade account coefficient of developing and developed countries is negative and significant
impact on the developing countries case and insignificant impact on the developed countries
case. The combine effect coefficient is negative and shows the insignificant effect on the model.
The coefficient of unemployment rate is positive in case of developing countries and shows the
insignificant behavior. The coefficient of unemployment rate in developed countries is positive
and shows the significant impact on the model but the combine effect coefficient is positive and
insignificant impact on the model.
2.4.1 GDPPCG
GDP shows the value market values of the all the final goods and services produce in the
economy in given year. The growth rate of GDP is used to measure the performance of the
country and it’s widely used around the world. It considers both individual and public utilization,
just as corporate and public venture, just as commodities short imports. In the above table
GDPPCG is the dependent variable and the other variables are said to be independent or
regressors.
2.4.2 Capital Account
The current account is based on the official reserve the government which includes short terms
loan, long terms loan and official reserve place in the central bank. A positive worth
demonstrates that the nation is a net exporter of labor and products, while a negative worth
shows that it is a significant exporter of labor and products. The positive coefficient of the capital
account means the impact of current account balance on GDPPCG is positively and 1 unit
change in capital account leads to 0.0141 impact on the dependent variable and shows the
insignificant effect on the model because the p value is greater than 5 percent. Now we simply
look the graphical relationship between capital account and GDPPCG.
GDPPCG CABalance
20
16
12
-4
-8
-12
European Union - 91
European Union - 11
United Kingdom - 01
United States - 91
United States - 11
Saudi Arabia - 91
Saudi Arabia - 11
South Africa - 01
Argentina - 91
Argentina - 11
Indonesia - 11
Indonesia - 91
Germany - 91
Germ any - 11
Australia - 06
Canada - 11
M exico - 91
M exico - 11
France - 01
Turkey - 11
Japan - 91
Turkey - 91
China - 01
Japan - 11
Korea - 01
Brazil - 96
Brazil - 16
India - 01
Italy - 01
Russian Federation - 01
The graph show the positive relationship between GDPPCG and capital account balance when
the positive surplus accumulate in the current account leads to increase the growth rate of the
GDP.
2.4.3 Foreign Direct Investment
Foreign Direct investment considered to one of the major part which is contribute in the
development of countries. The investment comes from outside from the country is called FDI.
The inflow of FID is higher means the greater foreign reserve available and increase in
employment in the economy ultimately increases the level of GDP. The coefficient of FDI 0.120
and shows the positive effect on the growth rate off the GDP and the coefficient value meaning
that 1 unit change in FDI leads to 0.120 percent impact on the growth rate of the GDP but the p
value is higher than critical region or lies outside the critical region which indicates the
coefficient insignificant impact on the model. The below graph indicate that there is exhibit
positive relationship between FID and GDP growth and with the passage of time the FDI inflow
in the economy leads to positive impact on the GDP.
GDPPCG FDI
12
-4
-8
-12
Russian Federation - 01
Australia - 06
European Union - 91
European Union - 11
France - 01
India - 01
Indonesia - 91
Indonesia - 11
Italy - 01
Mexico - 91
Mexico - 11
Saudi Arabia - 91
Saudi Arabia - 11
South Africa - 01
United Kingdom - 01
United States - 91
United States - 11
Argentina - 91
Argentina - 11
Brazil - 96
Brazil - 16
Canada - 11
China - 01
Germany - 91
Germany - 11
Japan - 91
Japan - 11
Korea - 01
Turkey - 91
Turkey - 11
2.4.4 Gross Fixed Capital
The Gross Fixed Capital formation means to increase the way of investment by using its
resources for the development or promotion of investment and make the arrangement of new
capital such situation is called gross capital fixed capital formation. Without new investment the
growth rate of the GDP is not increase and its directly influence the growth of the GDP. The
gross fixed capital formation is most volatile variable in the economy as the economy suffer in
negative shots it leads to decrease the investment multiple times and vice versa. The Viner
(1936) said that there is multiple time increase in growth rate of the GDP as 1 unit change in the
capital formation or adding new capital in the economy. Further it is necessary to increase the
capital stock year by year to maintain the growth rate of the GDP. The coefficient of Gross fixed
capital development is positive and but insignificant impact on the model.
GDPPCG GFCapital
50
40
30
20
10
0
E u ro p e a n U n io n - 1 1
U n ite d K in g d o m - 0 1
Eu rop ea n U n io n - 9 1
U n ited S ta te s - 9 1
U n ited S ta te s - 1 1
S a u d i A ra b ia - 1 1
Sa ud i Ara b ia - 91
S o u th A frica - 0 1
-10
In do n e sia - 1 1
In d o n e sia - 9 1
Arg en tin a - 91
Arg en tin a - 11
G e rm a n y - 9 1
G e rm a ny - 11
A u stra lia - 0 6
C anada - 11
M exico - 1 1
M e xico - 9 1
F ra nce - 01
T u rke y - 9 1
T u rke y - 1 1
Ja p a n - 1 1
Ja p a n - 9 1
C h ina - 0 1
K o re a - 01
Bra zil - 96
Bra zil - 16
In d ia - 0 1
Ita ly - 0 1
Russian Federation - 01
The above graph shows positive relationship between gross capital formation and GDP and
depicts that when the amount of capital or investment increase the growth rate of the GDP also
increase in the same direction and vice versa.
2.4.5 Government Consumption
The government spending also directly influence on the growth rate of the GDP. If the
government wants to increase the economic activity the government cut the tax rate to increase
the growth rate of the economy. Further if government to see the economy is suffers in the
inflation then use the strict fiscal policy in this station the economic activity reduces. Further if
the government investment or spending on development projects such as dams, roads, hospital
and other big projects then its leads to positive contribution in GDP growth but if the government
spending on other purpose or non-development projects such as military spending and debit
servicing this situation is leads to negative impact on the GDP growth of the economy. The
public authority utilization use coefficient is negative which is equivalent to - 0.471 and shows a
critical impact in light of the fact that the p-esteem is not exactly the percent level. It occurred
because of non-advancement use which implies that the public authority hugely spent on
enormous non-useful exercises for this situation GDP development rate shows a negative
connection among spending and creation (Mo, 2007).
GDPPCG GovCons
30
25
20
15
10
-5
-10
Russian Federation - 01
Australia - 06
European Union - 11
Germany - 11
India - 01
Japan - 91
Saudi Arabia - 91
Argentina - 91
Argentina - 11
Brazil - 96
Brazil - 16
Canada - 11
China - 01
European Union - 91
France - 01
Germany - 91
Indonesia - 91
Indonesia - 11
Italy - 01
Japan - 11
Korea - 01
Mexico - 91
Mexico - 11
Saudi Arabia - 11
South Africa - 01
United Kingdom - 01
United States - 91
United States - 11
Turkey - 91
Turkey - 11
The above graph shows that there is negative relationship between government spending and
growth rate of the GDP. It is happened when the government diverts the resources to some
wastage of non-productive development purpose.
2.4.6 Inflation
The inflation means the overall increase in the general price level or in other terms the inflation
occur when the prices increase rapidly varies from time to time periods. It is also most important
factor which disturb the economics activity and the central bank control this factor by using
different tools. The inflation effect the buying poser of the consumer when the inflation increases
the buying power of the consumer down because the prices of the basket increase. The
coefficient of the expansion shows the adverse consequence on the GDP which implies that 1
unit change in the inflation adverse consequence on the GDP which is equivalent to - 0.004 and
shows the critical impact on the model. Our result is coordinate or backing with the investigation
of Banerjee and Marcellino (2006).
GDPPCG Inflation
1,200
1,000
800
600
400
200
-200
Russian Federation - 01
Australia - 06
Brazil - 16
China - 01
Germany - 91
India - 01
Indonesia - 91
Indonesia - 11
Italy - 01
Japan - 91
Japan - 11
Korea - 01
South Africa - 01
Turkey - 91
Turkey - 11
Argentina - 91
Argentina - 11
Brazil - 96
Canada - 11
European Union - 91
European Union - 11
France - 01
Germany - 11
Mexico - 91
Mexico - 11
Saudi Arabia - 91
Saudi Arabia - 11
United States - 91
United States - 11
United Kingdom - 01
12
-4
-8
-12
Russian Federation - 01
Australia - 06
India - 01
Indonesia - 91
Indonesia - 11
Italy - 01
United States - 91
Argentina - 91
Argentina - 11
Brazil - 96
Brazil - 16
Canada - 11
China - 01
European Union - 91
European Union - 11
France - 01
Japan - 91
Japan - 11
Korea - 01
Mexico - 91
Mexico - 11
Saudi Arabia - 91
Saudi Arabia - 11
South Africa - 01
Turkey - 91
Turkey - 11
United Kingdom - 01
United States - 11
Germany - 91
Germany - 11
The above graph also shows the negative relationship between GDP and fertility rate which
shows the number of birth rate leads to decrease the growth rate of GDP.
2.4.8 Trade
Trade is most important part of the economy it is the external part of the economy based on the
export and import. If the export is greater than import then we can say that trade account is
surplus and if the export is less than import then the economy suffer in deficit trade. The
coefficient of trade is negative and shows the insignificant effect on the model. The coefficient of
trade account is negative due to export is lower than import or weaker export and strong demand
of import goods which results the foreign resources flight outside the country in the form of
foreign payment (Zestos and Tao 2002).
Trade GDPPCG
100
80
60
40
20
-20
Russian Federation - 01
Australia - 06
Brazil - 96
Germany - 91
Indonesia - 11
Italy - 01
Japan - 91
South Africa - 01
United States - 91
United States - 11
Argentina - 91
Argentina - 11
Brazil - 16
Canada - 11
China - 01
European Union - 91
European Union - 11
France - 01
Germany - 11
India - 01
Indonesia - 91
Japan - 11
Korea - 01
Mexico - 91
Mexico - 11
Saudi Arabia - 91
Saudi Arabia - 11
United Kingdom - 01
Turkey - 91
Turkey - 11
The above graph shows that the number of import increase leads to increase the deficit in trade
and services account results it will leads to negative impact on the growth rate of the GDP of the
economy.
2.4.9 Unemployment
The unemployment rate is the rate which reflects how much the working age population is
unemployed despite they are ready for working in the economy. The unemployment coefficient
is positive value 0.011 and insignificant effect on the model because the p value is greater than
critical boundary. But Okun law says that 4 percent decrease in GDP as 1 percent increase in
unemployment. In this study this coefficient turns to be positive and shows the insignificant
behavior.
GDPPCG Unemp
40
30
20
10
-10
Russian Federation - 01
Australia - 06
European Union - 91
European Union - 11
France - 01
India - 01
Indonesia - 91
Turkey - 91
United Kingdom - 01
United States - 91
United States - 11
Argentina - 91
Argentina - 11
Brazil - 96
Brazil - 16
Canada - 11
China - 01
Germany - 91
Germany - 11
Indonesia - 11
Italy - 01
Japan - 91
Japan - 11
Korea - 01
Mexico - 91
Mexico - 11
Saudi Arabia - 91
Saudi Arabia - 11
South Africa - 01
Turkey - 11
The above graph shows that there is ambiguous relationship between unemployment and growth
rate of GPD which totally negate the negative relationship between unemployment and growth
rate of the GDP.
Chi-Sq.
Test Summary Statistic Chi-Sq. d.f. Prob.
12 Mean 0.003872
Median -0.006848
Maximum 5.672856
8 Minimum -8.422331
Std. Dev. 1.884439
4
Skewness -0.432677
Kurtosis 6.210495
0 Jarque-Bera 51.59520
-8 -6 -4 -2 0 2 4 6
Probability 0.000000
The Jarque-Bera test, a type of Lagrange multiplier test, is a normalcy test. Many evaluation
measures, such as the t test and the F test, presume normality; the Jarque-Bera test is commonly
used to confirm normality before one of these exams. Alternative parametric tests, such as
Shapiro-Wilk, are problematic when n surpasses 2,000, thus it's usually used for large data sets.
The skewness and kurtosis of data are compared to see if they are similar to the normal
distribution. The data could be presented in a variety of ways, including.
A skew of zero shows that the distribution is symmetrical around the mean, and a kurtosis of
three indicates how much data is in the tails and how high the distribution expands. It shouldn't
need to know the information's mean or standard error to conduct the test. The above histogram
shows that the data is normally distributed and the value of Jarque-Bera test is greater than p
value which indicated that we can reject the null hypothesis which stated that data is not
normally distributed. So, there is sufficient evidence that we can accept the alternative
hypothesis. Now we test the cross-sectional dependency test which shows that error is serial
correlated or not.
Cross-sections included: 20
Total panel (unbalanced) observations: 112
Test employs centered correlations computed from pairwise samples
The above table shows the cross-sectional dependency test and null hypothesis shows that no
cross-section dependence among residual exist and alternative hypothesis stated that cross
sectional dependence serial correlation exists. The result shows that Breusch-Pegan LM test p
value is less than critical value and which indicate that we cannot reject the null hypothesis. We
can conclude that cross section dependence among error is exhibit.
References
Lacalle-Calderon, M., Perez-Trujillo, M., & Neira, I. (2017). Fertility and economic
development: Quantile regression evidence on the inverse J-shaped pattern. European Journal of
Population, 33(1), 1-31.
Massin, S., & Kopp, P. (2014). Is life satisfaction hump-shaped with alcohol consumption?
Evidence from Russian panel data. Addictive Behaviors, 39(4), 803-810
.
Mo, P. H. (2007). Government expenditures and economic growth: the supply and demand
sides. Fiscal studies, 28(4), 497-522.
Zestos, G. K., & Tao, X. (2002). Trade and GDP growth: causal relations in the United States
and Canada. Southern Economic Journal, 859-874.