Corresponding Author: Reyna Karlina
Email: reynakarlina.2021@student.uny.ac.id
INTRODUCTION
A problem that is very difficult for every
country to avoid is unemployment. Unemployment is the difference between the
level of the labour force and the level of employment opportunities; if the
supply of labour is higher than the demand for labour, then this can cause an
increase in the number of unemployed (Yuksel & Adali, 2017). The main
cause of unemployment is the imbalance between demand and supply of labour (Alam et al., 2020).
The increasing number of unemployed is also
considered a government failure to improve society's welfare. The increasing
number of people who do not have jobs can trigger increasing poverty rates,
which will affect a country's economy. Thus, unemployment is a major factor in
a region's economy because it can slow down the pace of a country's economy (Manaa & ul Haq, 2020).
The issue of unemployment is a crucial
problem and is also very worrying for every region, including West Java
Province. According to Kata Data (2022), it is reported that West Java is the
province with the highest number of unemployed as of February 2022 at 8.35% of
the total population, after Banten Province at 8.53% of the total population. Excessive unemployment levels in a region
negatively impact the economy, causing unstable economic conditions (Maqbool et al., 2013). For
example, increasing the number of poverty and the amount of government
spending, low per capita income, and decreasing tax revenues.
The educated unemployed category dominates the number of unemployed in West
Java Province. The educated unemployed are people who have obtained a secondary or higher
level of education and are looking for or available for work (Reimeingam, 2014). Due to the crisis in the job market,
graduates' lives have become more difficult than ever (Aktar et al., 2021). This condition occurs because competition
for labour is increasing while the number of available jobs is increasingly
limited. Therefore, this research focuses on the educated unemployed, high
school/vocational school graduates, and above. Data on the comparison of educated unemployment with open
unemployment is as follows.
Figure 1.
Data on the number of unemployed in West Java Province 2017-2021
Source:
Survey data from BPS West Java Province (2022)
Figure 1
shows that the number of open unemployed in West Java Province from 2017 to
2021 continues to increase. The highest increase occurred in 2020, from 1,901,498 to 2,533,076, in
other words, an increase of 34.2%.
This increase was caused by Covid 19, which caused the economy to experience a
contraction, so many employees were laid off. From 2017 to 2021, it shows that
the number of unemployed in this province is dominated by educated high
school/vocational school graduates and diploma/university graduates. On average, the number of educated
unemployed is 1,271,129, or 60.2% of the total open unemployment. It is a very
worrying condition that many high school/vocational school and
diploma/university graduates considered skilled workers still have not found work.
Widowati et al. (2022) state that educated unemployment is a serious problem. This is when many educated people do not have jobs, which can increase the country's
socioeconomic chaos. Socially, educated unemployment can hinder a
country's economic development, reduce productivity, and increase social
inequality. Four factors are thought to influence the number of educated
unemployed, namely: 1) Economic Growth as measured through Gross Regional Domestic Product (GRDP), foreign investment, Regency/City Minimum Wage, and Human
Development Index (HDI). Data on Economic Growth, foreign investment, MSEs, and
HDI are as follows.
Table 1 . Trends
in GDP Growth, Foreign Investment, MSEs, and HDI in West Java Province
|
Category
|
2019
|
2021
|
Growth
|
|
GRDP (%)
|
5.02%
|
3.74%
|
-25.50%
|
|
Foreign Investment (Rp Million)
|
3,267.25
|
2,821.42
|
-13.65%
|
|
UMK (Rp. Thousand)
|
2,731,081
|
3,073,079
|
12.52%
|
|
HDI (Unit)
|
72.03
|
72.45
|
0.58%
|
Source:
Survey data from BPS West Java Province (2022)
Table 1 shows that the average GRDP growth in 2021 has decreased
from 5.02% in 2019 to 3.74% in 2021, or a decrease of 25.50%. On average, foreign investment decreased
from IDR 3,267.25 billion in 2019 to IDR 2,821.42 billion in 2021 or a decrease
of 13.65%. On average, MSEs have increased from IDR 2,731,081 in 2019 to IDR
3,073,079 in 2021 or 12.52%. Meanwhile, the HDI has increased from 72.03 in
2019 to 72.45 in 2021 or an increase of 0.58%.
Economic growth can produce high commodity
demand shocks (Wang & Liao, 2022). The soaring demand for these commodities
can also encourage companies to increase their production capacity so that the
demand for educated workers will increase. This can reduce the number of
educated unemployed. Previous research results prove that high economic growth
can reduce the number of educated unemployed (Lativa &
Susilastuti, 2022; Kinasih &
Nihaya, 2022; Yunitasari et
al., 2021; Chris, 2015). However, contrary to the
research results by Alam
et al. (2020), economic growth can increase
unemployment. Meanwhile, according to (Dachito
et al., 2021), Magazzino
(2014) states that economic growth does
not affect educated unemployment.
Foreign investment is investment from abroad that can help the
government develop and establish new companies ( Sadikova
et al., 2017). Projects
funded by foreign investors generate more market share of both skilled and
unskilled labour (Irpan
et al., 2016). Increasing employment
opportunities can reduce the number of educated unemployed in the region. The
research results by Ma'in
et al. (2021) prove that foreign investment can
significantly reduce graduate unemployment. However, this differs from the
results of research by Eneji
et al. (2013), which states that foreign
investment does not affect educated unemployment.
Wages are an interesting issue because most of the unemployed
prefer the informal sector to fulfil their working life even though they are
still looking for other jobs compared to the formal sector, which pays minimum
wages (Harahap, 2018). Wage determination in this research focuses on the Regency/City Minimum Wage (UMK). The higher the UMK setting in
an area can attract the interest of prospective workers to apply for jobs,
thereby increasing job competition. Meanwhile, the amount of energy required is
very limited. Regions with a high MSE can increase operational costs for
companies, especially employee salary costs. This condition impacts Termination
of Employment (PHK) so that the number of educated unemployed increases. The
results of research by Lativa and Susilastuti (2022) and Setyadi
et al. (2019) show that the number of educated
unemployed can be increased by increasing the minimum wage. However, Yunitasari
et al. (2021) and Adesola
et al. (2017) prove that the minimum wage does
not affect educated unemployment.
Another factor that influences educated unemployment is HDI. The
Human Development Index (HDI) measures how people obtain health, income and education
(Assa, 2021). The higher the HDI in a region,
the more human resources that have graduated with superior competencies and
economic expansion so that it can reduce the number of educated unemployed.
Previous research results prove that HDI can reduce the number of unemployed (Auliya
& Agusalim, 2022;
Soekapdjo & Oktavia, 2021 ). However, contrary to the
research results by Widowati
et al. (2022), HDI does not affect educated
unemployment.
Referring to
the previous explanation regarding the problem of educated unemployment in West
Java Province, which is increasing and difficult to avoid because it can
increase socio-economic chaos in the province. Educated unemployment, in particular, refers to a situation
where highly educated and qualified individuals cannot find suitable employment
opportunities. Educated unemployment is a problem in many countries worldwide
and has economic and social implications. Therefore, exploring the factors
influencing educated unemployment to develop policies and measures to address
this problem effectively is very important. Therefore, this
research aims to analyze economic growth, foreign investment, wages, and the human
development index on educated unemployment in West Java province for 2017-2021.
The benefit of this research is that it is hoped that it will become reference
material and references for readers to carry out further research, especially
related to the factors that influence educated unemployment, and it is also
hoped that the research results can contribute ideas to the government,
especially the West Java Provincial Government in determining employment
policies. , as well as being a consideration for the central and regional
governments, especially the West Java Provincial governments, in formulating
policies to reduce educated unemployment.
METHOD
This research uses secondary data, panel data with a cross-section
of 27 districts/cities and a time series for 5 years from 2017 to 2021. Panel
data is a combination of time series data and cross-section data (Widarjono,
2013). The approach in this research uses quantitative research. According to Cooper & Schindler (2014), quantitative data is a type of research that is measured through a
set of numbers to obtain information; then, these numbers can be processed
statistically through the Eviews Program version 12. The data used in this
research is educated unemployment data, processed from a survey on the West
Java Province Central Statistics Agency website based on data on the State of
the Labor Force in West Java Province from 2017 to 2021. This research was
conducted in West Java Province, involving 18 Regencies and 9 Cities in the
province. The research period was carried out from December 15 2022, to March
30 2023, from preliminary research until data collection for processing. The
data analysis technique used is model selection, such as the Chow test, Hausman
test, Lagrange-Multiplier (LM) test, and classical assumption test and
hypothesis test.
RESULTS AND DISCUSSION
Panel Data Regression Test
The initial step taken
in this research was to collect data from 18 districts and 9 cities obtained
from BPS West Java Province for the 2017-2021 period with the variables Economic
Growth, Foreign Investment, Wages, HDI and educated unemployment. Then, after
collecting the data, proceed with describing the research variables with
descriptive statistics consisting of average, standard deviation, minimum and
maximum values.
Next, panel data
regression testing was carried out with three procedures for selecting the best
model, namely the Chow test to choose between CEM or FEM, the Hausman test to
choose between FEM or REM, and the Lagrange multiplier (LM) test, which was used
to choose between CEM or REM. The results obtained were that the Chow test's
F-calculated probability value was 0.0000, where < sig. 0.05. This means the
Fixed Effect Model is more suitable than the Common Effect Model. Then,
continuing by carrying out the Hausman test, the random cross-section
probability value was 0.1810, where > sig.0.05, which means the Random
Effect model is more suitable than the Fixed Effect Model. So, it is necessary
to carry out the next test, namely the LM test; based on the results of the LM
test, the Breusch-Pagan Cross-section probability value is 0.000, where <
sig. 0.05 means the Random Effect Model is more suitable than the Common Effect
Model. This results in the conclusion that the more appropriate model is the Random
Effect Model.
The next step is to
carry out the classical assumption test, which consists of four tests:
normality, multicollinearity, heteroscedasticity, and autocorrelation. The
first classical assumption test is the normality test, and it is known that the
probability value is 0.120891 so that the residuals are normally distributed.
The second classical assumption test, namely the multicollinearity test,
concluded that there was no high correlation between independent variables
(above 10). This means that multicollinearity does not occur. Then, the next
test, namely heteroscedasticity, obtained a probability value for each
independent variable above 0.05, meaning there were no cases of
heteroscedasticity. Furthermore, the autocorrelation test obtained a value of
dw=2.198 so that dU<dw<4-dU (1.7802<2.198<2.220), so it can be
concluded that there is no case of autocorrelation.
After carrying out the
classical assumption test, the next step is hypothesis testing consisting of
the t-test, F-test, and R2 determination. In the t-test, only the Foreign
Investment variable (X2) does not have a partial positive effect on
Educated Unemployment (Y), with a coefficient value of Foreign Investment (X2)
of 0.376599. So, hypothesis 2 (H2) is rejected. Meanwhile, based on
the results, the variables Economic Growth (X1), Wages (X3),
and HDI (X4) can be concluded that they have a partial effect on
Educated Unemployment.
The F test results show
that the variables Economic Growth (X1), Foreign Investment (X2),
Wages (X3), and HDI (X4) simultaneously influence
Educated Unemployment (Y) because of the prob value. The result obtained is
0.000, meaning it is smaller than 0.05. The coefficient of the determination
test obtained a figure of 0.945403. This means that the contribution of all
independent variables (Economic Growth (X1), Foreign Investment (X2),
Wages (X3), and HDI (X4) in explaining the dependent
variable (Educated Unemployment (Y) is 94.5%. Moreover, the remaining 5.5% is
explained by other variables outside this research model.
This analysis design is carried out by processing the
data that has been collected and then analyzing it using statistical tools.
Panel data regression has a development of linear regression with the Ordinary
Least Square (OLS) method, which has specificities in data type and analysis
objectives. Regarding data type, panel data regression has cross-section and
time series data types. The cross-section nature of data is shown by data that
consists of more than 1 entity, and then the time series nature is shown by
more than 1-time observation.
Selection of Panel Data Estimation Model
Techniques
There are three procedures in testing the
selection or suitability of modelling that will be used to select a panel data
regression model, namely:
a)
A statistical test is used to choose between CEM or FEM or Chow test.
b)
The Hausman test is used to choose between FEM or REM.
c)
The Lagrange multiplier (LM) test is used to choose between CEM or REM.
1.
Test Chow
The Chow test chooses the appropriate method between Common Effect Mode or Fixed Effect Mode. This
test follows the F-statistic distribution.
The hypothesis used in this research is as
follows:
H0 : Then the model is used. Common Effects
Ha: Then use Fixed Effect mode and continue testing Hausman
If the F-calculated value is greater than the F-table value, it is considered significant and rejects
H0. In other words, accepting Ha, which states that the estimate with Fixed
Effect Mode I is better than the estimate with Common
Effect Mode I.
The guidelines used in drawing Chow test conclusions are:
a.
If the value of probability F ≥ 0.05 means H0 is accepted, then the model is used Common Effects.
b.
If the probability value F < 0.05 means H0 is rejected, then model is used Fixed Effect and continued with the Hausman test to choose whether to use model Fixed Effect or Random Effect.
Table 1. Chow test
|
Redundant
Fixed Effects Tests
|
|
|
Equation:
Untitled
|
|
|
Cross-section
fixed effects test
|
|
|
Effects Test
|
Statistics
|
df
|
Prob.
|
|
Cross-section F
|
16.160240
|
(26,104)
|
0.0000
|
|
Chi-square cross-section
|
218.351431
|
26
|
0.0000
|
Source: 2023 data processing
Based on the results of the Chow
test, the F-calculated probability value is 0.0000, where < sig. 0.05. This
means the Fixed Effect Model is more suitable than the Common Effect Model.
2.
Hausman test
The Hausman test is used to choose the best approach between the
Random Effect Model (REM) and Fixed Effect Model (FEM) approaches in estimating
panel data. The basis for decision-making is as follows:
a.
Suppose the
probability value for a random cross-section is > 0.05 significant value. H0
is accepted in that case, so the most appropriate model to use is the
Random Effect Model (REM).
b.
If the
probability value for a random cross-section is <0.05 significant, then H0
is rejected, so the most appropriate model to use is the Fixed Effect Model
(FEM).
Table 2. Hausman test
|
Correlated
Random Effects - Hausman Test
|
|
|
Equation:
Untitled
|
|
|
|
Cross-section
random effects test
|
|
|
Test Summary
|
Chi-Sq.
Statistics
|
Chi-Sq.
df
|
Prob.
|
|
Random cross-section
|
6.253750
|
4
|
0.1810
|
Source: Data Processing 2023
Based on the Hausman test results, the random cross-section
probability value is 0.1810,> 0.05. This means the Random Effect Model
is more appropriate than the Fixed Effect Model.
3.
LM Test
the test aims to determine the
best estimation model between the CEM and REM models. LM test calculations were carried out using the
Breusch-Pagan method. Hypotheses in the LM
test include :
a.
If the Breusch-Pagan Cross-section value ≤ significance 0.05, reject H0, meaning the Random Effect Model (REM)
is selected.
b.
If the Breusch-Pagan Cross-section value ≥ significance 0.05, then accept H0, that is, the Common Effects Model (CEM)
selected.
Table 3. LM Test
|
Lagrange
Multiplier Tests for Random Effects
|
|
Null
hypothesis: No effects
|
|
|
Alternative
hypotheses: Two-sided (Breusch-Pagan) and one-sided (all others) alternatives
|
|
|
Test Hypothesis
|
|
|
Cross-section
|
Time
|
Both
|
|
Breusch-Pagan
|
137.4160
|
1.344587
|
138.7606
|
|
|
(0.0000)
|
(0.2462)
|
(0.0000)
|
Source: Data Processing 2023
Based on the LM test results, the Breusch-Pagan Cross-section probability
value was equal to 0.000 where < sig.
0.05. This means the Random Effect Model is more appropriate than the
Common Effect Model.
4.
Results of
Panel Data Regression Model Selection
After looking at the three test results for selecting the panel
data regression model that had been tested, the researchers concluded that the
more appropriate model was the Random Effect Model. The regression equation
model and summary of research results are as follows.
Table 4. Random Effect Model
Results
|
Variables
|
Coefficient
|
Std. Error
|
t-Statistics
|
Prob.
|
|
C
|
27434.91
|
20529.01
|
1.336397
|
0.1838
|
|
Economic
growth
|
-4.688526
|
1.229111
|
-3.814565
|
0.0002
|
|
Foreign
Investment
|
-0.080902
|
0.540808
|
-0.149595
|
0.8813
|
|
Wages
|
22.58810
|
2.888918
|
7.818878
|
0.0000
|
|
HD
|
-852.2376
|
298.9419
|
-2.850847
|
0.0051
|
Source: 2023 data processing
PT = 27434.91 + 4.688526 Economic Growthi . t - 0.080902 Foreign Investmenti . t + 22.58810 Wagesi . t - 852.2376 HDIi . t
Classic assumption test
1.
Normality
test
The classic normality test is carried out
to test whether the regression model's residual values are normally
distributed. If the probability value is above 0.05, the data in the equation
is normally distributed. Test results
can be seen:
Figure 8. Normality Test
Source: 2023 data processing via Eviews version 12
Based on the normality test in Eviews 12 in Figure 4.1, it can be
seen that the normality test produces a probability value of 0.120891 so that
the residuals are normally distributed.
a.
Multicollinearity
Test
The
multicollinearity test was carried out to determine whether or not there was a
correlation between independent variables. A regression model that is said to
be good should not have any correlation between independent variables. If a VIF
value above 10 is found, it is suspected that there is multicollinearity in the
equation model. Test results can be seen:
Table 5. Multicollinearity Test
|
|
Coefficient
|
Uncentered
|
Centred
|
|
Variables
|
Variance
|
VIF
|
VIF
|
|
C
|
5.34E+08
|
159.0174
|
NA
|
|
Economic
growth
|
3.644991
|
4.953995
|
2.648287
|
|
Foreign
Investment
|
0.172601
|
2.550322
|
2.150002
|
|
Wages
|
8.046426
|
19.75825
|
1.982520
|
|
HDI
|
129902.2
|
189.2771
|
1.239792
|
Source:
2023 data processing via Eviews version 12
Based on the results from Table 5, the results of the
multicollinearity test state that if there is no high correlation between
independent variables (Economic Growth, Foreign Investment, Wages, and Human
Development Index) where the VIF value is less than 10, then H0 is accepted.
Thus, there is no multicollinearity problem between the independent variables
in the regression model.
b.
Heteroscedasticity
Test
The heteroscedasticity test violates classical
assumptions where disturbance is detected in the regression equation model.
Test heteroscedasticity can be done using the Glejser test, which regresses the
absolute value of the residual on the independent variable.
Based on decision-making:
1) If the probability value is > 0.05,
then Ha is rejected, and Ho is accepted, which means there is no
heteroscedasticity problem.
2) If the probability value is <0.05, Ho
is rejected, and Ha is accepted, meaning there is a heteroscedasticity problem.
The following are the results of the
heteroscedasticity test which can be seen:
Table 6. Heteroscedasticity
Test
|
Dependent Variable: ABSRESSION
|
|
|
|
Method: Panel EGLS (Cross-section random effects)
|
|
Date: 08/17/23 Time: 01:38
|
|
|
|
Sample: 2017 2021
|
|
|
|
Periods included: 5
|
|
|
|
Cross-sections included: 27
|
|
|
|
Total panel (balanced) observations: 135
|
|
|
Swamy and Arora estimator of component variances
|
|
Variables
|
Coefficient
|
Std.
Error
|
t-Statistics
|
Prob.
|
|
C
|
7163,798
|
15462.98
|
0.463287
|
0.6439
|
|
Economic
growth
|
3.222665
|
423.2471
|
0.007614
|
0.6790
|
|
Foreign
Investment
|
-0.044732
|
0.366120
|
-0.122178
|
0.9029
|
|
Wages
|
6.921003
|
2.122270
|
3.261132
|
0.1400
|
|
HDI
|
-161.9488
|
232.1549
|
-0.697590
|
0.4867
|
Source: 2023 data processing via Eviews
version 12
Based on these tests, the prob. of each independent variable was
obtained above 0.05. This means that there are no cases of heteroscedasticity.
Table 7. Interpretation of the
Glejser Test
|
Independent Variable
|
Prob.
|
Decision
|
|
Economic
growth
|
0.5790
|
Heteroscedasticity does not
occur
|
|
Foreign
Investment
|
0.9029
|
Heteroscedasticity does not
occur
|
|
Wages
(UMK)
|
0.1400
|
Heteroscedasticity does not
occur
|
|
Human Development Index (HDI)
|
0.4867
|
Heteroscedasticity does not
occur
|
Source: 2023 data processing via Eviews
version 12
c.
Autocorrelation
Test
The autocorrelation test aims to determine whether, in a model,
there is a correlation between disturbing errors in periods t and t-1. The
results of the autocorrelation test can be seen as follows:
Table 8. Autocorrelation Test
|
Durbin-Watson
stat
|
1.890993
|
Based on Table
8, it can be seen that the Watson Durbin value is between dU and 4-dU with
n=135 and K=4. The value of dw=1.891 so that dU < dw < 4-dU (1.7802 <
1.891 < 2.220) means it can be concluded that there is no case of
autocorrelation.
Hypothesis testing consists of a partial
test (t test), simultaneous test (F test), and Adjusted coefficient of
determination test (R2) with estimates for panel data linear regression
using the Fixed Effect Model (FEM) as follows:
1. Partial test
(t-Test)
According to Ghozali (2018), the t-test
is a test to see the correlation between independent and dependent variables,
carried out individually (partially). The t-test was used with a significance
level of 0.05 in this study.
Table 9 .
Partial Test (t-Test)
|
Variables
|
Coefficient
|
Std.
Error
|
t-Statistics
|
Prob.
|
|
C
|
27434.91
|
20529.01
|
1.336397
|
0.1838
|
|
Economic growth
|
-4.688526
|
1229.111
|
-3.814565
|
0.0002
|
|
Foreign Investment
|
-0.080902
|
0.540808
|
-0.149595
|
0.8813
|
|
Wages
|
22.58810
|
2.888918
|
7.818878
|
0.0000
|
|
HDI
|
-852.2376
|
298.9419
|
-2.850847
|
0.0051
|
Source: 2023 data processing via Eviews
version 12
Based on Table 9 above, it can be seen from the results of the
t-test the influence of the variables Economic Growth (X1), Foreign
Investment (X2), Wages (X3), and HDI (X4) on
the Educated Unemployment variable (Y). Based on the table above, the prob value can be known. The Economic
Growth variable (X1) is 0.0002 with a coefficient value -4.689. This
means that the Economic Growth variable (X1) partially has a
negative and significant effect on Educated Unemployment (Y). If the Economic
Growth variable (X1) increases by one per cent, Educated
Unemployment (Y) will decrease by 4.689%.
Meanwhile, the value of prob. the Foreign
Investment variable (X2) is 0.8813. This means that the Foreign
Investment variable (X2) does not have a positive effect on Educated
Unemployment (Y) partially with a coefficient value of Foreign Investment (X2)
of -0.080902. So, in this way, hypothesis 2 (H2) is rejected.
Prob value. On the Wage variable (X3)
of 0.000. This means that the Wage variable (X3) has a positive
effect on Educated Unemployment (Y) partially with a Wage coefficient value (X3)
of 22.58810. So, hypothesis 3 (H3) is accepted.
Prob value. The HDI variable (X4)
is 0.0051. This means that the HDI variable (X4) harms Educated
Unemployment (Y) partially with an HDI coefficient value (X4) of
-852.2376. Thus, hypothesis 4 (H4)
is accepted.
2. Simultaneous
Test (F Test)
The F test determines the magnitude of all independent variables'
influence on the dependent variable. Alternatively, in other words, the F test
(simultaneous test) is used to determine whether all independent variables
influence the dependent variable together.
Table 10. Model
Test (F Test)
|
F-statistic
|
29.79688
|
|
Prob(F-statistic)
|
0.000000
|
Source: 2023 data processing via Eviews
version 12
The F test is useful in simultaneously
measuring the influence of independent variables on the dependent variable. The
F test results show that the variables Economic Growth (X1), Foreign
Investment (X2), Wages (X3), and HDI (X4)
simultaneously influence Educated Unemployment (Y) because the prob value
obtained is 0.000 meaning it is smaller from 0.05.
3. Determination Coefficient Test (R2 Test )
2 test measures the size of the dependent variable that
the independent variable can explain. The criteria used for the R2
value are in the form of a percentage:
Table 11.
Determination Coefficient Test (R2 Test)
Source: Data Processing 2023
The coefficient of determination test
obtained a figure of 0.678305. This means that the contribution of all
independent variables (Economic Growth (X1), Foreign Investment (X2),
Wages (X3), and HDI (X4) in explaining the dependent
variable Educated Unemployment (Y) is 67.8% and other variables outside this
research model explain the remaining 32.2%.
The Effect of Economic Growth on Educated Unemployment
This research analyzes the effect of Economic Growth on educated
unemployment in West Java Province from 2017 to 2021. The research results
prove that Economic Growth has a probability value. 0.0002 < 0.05
means that economic growth statistically affects educated unemployment. The
results of this research proved the proposed hypothesis so that H1
could be accepted. Several previous studies support the results of this
research, such as research by Yunitasari et al. (2021), which analyzed the relationship between the level of Economic Growth and
educated unemployment in Indonesia, which has proven that statistically,
Economic Growth can also reduce educated unemployment. Another study was
conducted by Chris (2015) in his research in Nigeria, which proved that the two have a
significant relationship.
Economic growth in West Java Province can also trigger increased
consumption in society, which will encourage some industrial sectors in the
region to increase production. Increasing production to meet people's needs can
create new job opportunities and can also increase demand for high-skilled
workers, which leads to a decrease in the level of educated unemployment. This
condition can reduce the mismatch between the skills possessed by workers and
the skills required by employers. The research results by Galbraith Hale (2007) found that economic growth was associated with a decrease in the
overall unemployment rate, including reducing educated unemployment.
The Effect of
Foreign Investment on Educated Unemployment
The research results prove that foreign investment has prob
value. 0.8813 > 0.05 means that foreign investment statistically has no
effect on educated unemployment, so H2 is rejected. This study's results align with several previous studies,
such as those conducted by Eneji et al. (2013), that foreign investment has no significant effect in reducing
educated unemployment. Widowati et al. (2022) also conducted research in Central Java Province, which failed to
prove the effect of foreign investment on educated unemployment. This means
that the high or low amount of foreign investment entering West Java Province
has no impact on reducing the number of educated unemployed.
The influx of foreign capital can fund new projects, so this
requires labour (Irpan
et al., 2016). However, in this research, it is
proven that the use of foreign capital allows investors to bring in their
workforce rather than employing local workers. According to Open Data, West
Java (2023) reports that the number of foreign workers from foreign capital
absorption in 2021 will be 72,554. This results in minimal opportunities for
local workers who have high skills. Besides that, foreign investors may focus
on industries that require more low-skilled workers compared to high-skilled
workers so that foreign investment does not impact reducing educated
unemployment. These results contrast with research by Ma'in
et al. (2021), which proves that foreign
investment can reduce educated unemployment.
The Effect of Wages on Educated Unemployment
The relationship between wages and educated unemployment has a prob
value. 0.000 < 0.05. These results state that MSEs can significantly
influence educated unemployment, so H3 is accepted. Lativa & Susilastuti (2022), in their research in the province of Indonesia, prove that
increasing wages can increase the number of educated unemployed. Anjarwati Juliprijanto (2021) confirm that higher wages can significantly increase educated
unemployment.
This research proves that wages that are too high can cause job
switching and a mismatch between skills and job requirements. If wages are too
high, employers may not hire workers with the necessary skills and expertise,
leading to a lack of job opportunities for highly skilled workers. This, in
turn, may lead to an increase in educated unemployment. In addition, high wages
can cause the movement of skilled workers to be replaced by automation (Widowati et al., 2022). This, in turn, may lead to a decline in the rate of job
creation, which may exacerbate the problem of educated unemployment.
The Influence of HDI on Educated Unemployment
The research results between HDI and educated unemployment have a prob
value. Equal to 0.0051 < 0.05. These results prove that HDI has a
significant effect on educated unemployment. Previous studies support that a
higher HDI can reduce the number of educated unemployed. For example, research
by Auliya Agusalim (2022) in provinces in Indonesia has proven that HDI can have a
significant effect on the educated unemployment variable. Research by Soekapdjo
Oktavia (2021) confirms that HDI can reduce the
number of educated unemployed.
HDI measures a country's social and economic development and
considers life expectancy, education and per capita income (BPS, 2021). HDI can significantly impact educated unemployment (Auliya & Agusalim, 2022) because a higher level of development can increase access to
education and training, better job opportunities, and reduce the level of
educated unemployment. The method used by the local government is to provide
appropriate education and training. Various types of public and private high
schools and universities, as well as training institutions, can help reduce the
mismatch between the skills possessed by workers and the skills required by
employers.
CONCLUSION
In
this study analyzing the influence of Economic Growth, Foreign Direct
Investment, Wages, and Human Development Index (HDI) on educated unemployment
in West Java Province, several conclusions can be drawn: 1) Economic Growth:
The research affirms a significant negative impact of Economic Growth on
educated unemployment (with a probability value of 0.0002 < 0.05),
supporting the hypothesis (H1). Regions with higher economic growth tend to
have lower levels of educated unemployment due to increased job opportunities,
particularly in sectors requiring high-skilled labor. This underscores the
importance of policies promoting economic growth, business development,
entrepreneurship, and investment in education and training. 2) Foreign Direct
Investment (FDI): FDI does not significantly affect educated unemployment, as
indicated by the probability value of 0.8813 > 0.05, rejecting the
hypothesis (H2). This finding aligns with previous studies, suggesting that the
influx of foreign investment does not necessarily lead to a decrease in
educated unemployment. Policymakers should focus on other factors beyond FDI to
address educated unemployment effectively. 3) Wages: The study demonstrates a
significant positive impact of Minimum District Wage on educated unemployment
(with a probability value of 0.000 < 0.05), confirming hypothesis H3. Higher
wages can increase educated unemployment, especially in sectors relying on
low-skilled labor. Policymakers should encourage policies supporting
high-skilled labor, such as investments in high-tech industries and tailored
education and training programs. 4) Human Development Index (HDI): The research
proves the significant negative influence of HDI on educated unemployment (with
a probability value of 0.0051 < 0.05), supporting hypothesis H4. Higher HDI
correlates with better-educated workforce, enhancing adaptability to labor
market changes. Policymakers should prioritize investments in education,
healthcare, and policies promoting social mobility and economic growth to
reduce educated unemployment. In summary, the study underscores the importance
of economic growth, improvement in human development indicators, and strategic
wage policies to address educated unemployment effectively. Policymakers should
focus on creating a conducive environment for business development, promoting
high-skilled labor, and investing in education and technology to tackle the
challenges of educated unemployment in the region.
REFERENCES
Adesola, O., Oluseye, A., Omonike, A.,
Omowunmi, O., & Oluwaseunfunmi, O. (2017). Causes and Effects of Graduate Unemployment on the Nigerian
Economy (The Case Study of Lagos State). Asian Research Journal of Arts
& Social Sciences, 2(1), 110.
https://doi.org/10.9734/arjass/2017/30722
Aktar, G., Miah, T.,
Islam, A., & Hasan, N. (2021). Uncertainty After Graduation: A Deep Dive
into The Present Scenario of Graduate Unemployment Problem in Bangladesh. Pacific
International Journal, 4(2), 4953.
https://doi.org/10.55014/pij.v5i1.123
Alam, J., Alam, Q.
N., & Hoque, M. T. (2020). Impact of GDP, Inflation, Population Growth and
FDI on Unemployment: A study on Bangladesh Economy. African Journal of
Economics and Sustainable Development, 3(3), 6779.
https://doi.org/10.52589/ajesd/cah2iyqj
Anand, S., & Sen,
A. (1997). Concepts of Human Development and Poverty: A Multidimensional
Perspective. In Poverty and Human Development: Human Development Papers
(pp. 120).
Anjarwati, L., &
Juliprijanto, W. (2021). Determinan Pengangguran Terdidik Lulusan Universitas
di Pulau Jawa. Jurnal Ekonomi Pembangunan, 10(3), 178187.
https://doi.org/10.23960/jep.v10i3.280
Assa, J. (2021). Less
is more: The implicit sustainability content of the human development index. Ecological
Economics, 185(March), 107045.
https://doi.org/10.1016/j.ecolecon.2021.107045
Auliya, L., &
Agusalim, L. (2022). Pengaruh Covid-19 dan Upah terhadap Pengangguran
Terdidik di Indonesia : Analisis Data Panel 33 Provinsi Tahun 2010-2021.
17(3). https://doi.org/10.47198/naker.v17i3.148
Chris, O. I. (2015).
Graduate Unemployment and Economic Growth : Nigerian Experience Under
Democratic Government. International Journal of Engineering Research and
General Science, 3(1), 13891393.
Cooper, D. R., &
Schindler, P. S. (2014). Business Research Methods 12th Edition. In Business
Research Methods.
Dachito, A., Alemu,
M., & Alemu, B. (2021). The Impact of Public Education Expenditures on
Graduate Unemployment : Cointegration Analysis to Ethiopia. International
Journal of Commerce and Finance, 7(1), 117.
Eneji, M. A.,
Mai-Lafia, D., & Weiping, S. (2013). Socio-economic impact of graduate
unemployment on Nigeria and the vision 20:2020. International Journal of
Development and Sustainability, 2(1), 148176.
Galbraith, J. K.,
& Hale, T. (2007). Between-Sector Earnings Inequality in the United States.
UTIP Working Paper, 43., 127.
Harahap, E. F.
(2018). Study of Minimum Wage, Level of Education, Employment Opportunity, and
Unemployment Educated: Empirical Study in Padang. European Journal of
Business and Management, 10(3), 3843.
Irpan, H. M., Saad,
R. M., Nor, A. H. S. M., Noor, A. H. M., & Ibrahim, N. (2016). Impact of
Foreign Direct Investment on the Unemployment Rate in Malaysia. Journal of
Physics: Conference Series, 710, 110.
https://doi.org/10.1088/1742-6596/710/1/012028
Kinasih, putri suri, & Nihaya, dyah maya. (2022). Determinants of The
Unemployment University Graduates in Java Island. Indonesian Journal of
Development Economics, 5(1), 15051519.
Lativa, S., &
Susilastuti, D. (2022). Analysis of the Impact of Economic Growth and Wages
on Educated Unemployment in Indonesian Provincial Area.
https://doi.org/10.4108/eai.30-10-2021.2315852
Main, M., Othman, S.
M., Isa, S. S. M., & Junos, S. (2021). Unemployment and macroeconomic
indicators in Malaysia. Global Business and Management Research: An
International Journal, 13(4), 3845.
Magazzino, C. (2014).
Electricity demand, GDP and employment: Evidence from Italy. Frontiers in
Energy, 8(1), 3140. https://doi.org/10.1007/s11708-014-0296-8
Manaa, A., & ul
Haq, M. A. (2020). The Effects of SMEs, Population and Education level on
Unemployment in Kingdom of Bahrain. IKSP Journal of Business and Economics,
1(2), 2333.
Maqbool, M. S.,
Mahmood, T., Sattar, A., & Bhalli, M. N. (2013). Empirical Evidences from
Pakistan. Pakistan Economic and Social Review, 51(2), 191208.
Reimeingam, M.
(2014). Educated Unemployment in Sikkim: An Outcome of Educational Development.
Journal of North East India Studies, 4(1), 5566.
Sadikova, M., Faisal,
F., & Resatoglu, N. G. (2017). Influence of energy use, foreign direct
investment and population growth on unemployment for Russian Federation. Procedia Computer Science, 120, 706711.
https://doi.org/10.1016/j.procs.2017.11.299
Setyadi,
D., Sudarso, Y., Nahar, M., & Sugiyanta, S. (2019). Duration of educated unemployment. Economic Journal of
Emerging Markets, 11(2), 213222.
https://doi.org/10.20885/ejem.vol11.iss2.art8
Soekapdjo,
S., & Oktavia, M. R. (2021). Pengaruh Inflasi, Indeks Pembangunan Manusia,
Dan Upah Minimum Provinsi Terhadap Pengangguran Di Indonesia. Jurnal
Ecodemica Jurnal Ekonomi Manajemen Dan Bisnis, 5(2), 94102.
https://doi.org/10.31294/eco.v5i2.10070
Wang, F., & Liao,
H. (2022). Unexpected economic growth and oil price shocks. Energy Economics,
116, 112. https://doi.org/10.1016/j.eneco.2022.106430
Widowati, S., Putro,
T. R., & Mulyaningsih, T. (2022). Macro-Economic Impact on Educational
Unemployment in Central Java , Indonesia. The International Journal of
Social Sciences World, 4(1), 272280.
Yόksel, S., &
Adalı, Z. (2017). Determining Influencing Factors of Unemployment in
Turkey With Mars Method. International Journal of Commerce and Finance, 3(2),
2536.
Yunitasari, D.,
Khotimah, K., & Fathorrazi, M. (2021). the Implication of Brain Gain on
Brain Drain Phenomenon in Overcoming the Problem of Educated Unemployment in
Indonesia. Sosiohumaniora, 23(1), 133.
https://doi.org/10.24198/sosiohumaniora.v23i1.26749
