https://ijmri.de/index.php/jmsi
volume 4, issue 6, 2025
41
KEY MACROECONOMIC FACTORS AFFECTING THE UNEMPLOYMENT RATE
Tolagan Botirovich Sodiqov
Associate Professor, Faculty of Economics and Tourism,
Oriental University
Annotation:
This study investigates the macroeconomic factors affecting unemployment in the
Shanghai Cooperation Organization (SCO) member countries over the period 2004–2023, based
on data extracted from the World Bank database. We constructed a model and conducted a
regression analysis using the Ordinary Least Squares (OLS) method to explore the relationships
among the variables. The findings are presented below.
Keywords:
Unemployment, Inflation, Foreign Direct Investment (FDI), STATA-14 software,
OLS method, VIF indicators, Breusch-Pagan test, Breusch-Godfrey test.
1. Introduction.
In today’s context, unemployment has become one of the most frequently
debated issues. The global rise in unemployment has led many people in various countries and
cities to seek employment abroad rather than in their own countries. Moreover, in developed
nations, unemployment has been associated with increased crime rates and heightened social
tensions.
Let us first define the term “unemployment.” Unemployment is a socio-economic phenomenon
that reflects the underutilization of a portion of the available labor force within a country.
Individuals who are able to work and willing to be employed but are not currently working are
considered unemployed.
Unemployment can be classified into several types, including structural, frictional, institutional,
technological, regional, hidden, stagnant, and cyclical unemployment.
Among the primary causes of unemployment are the transitional periods during which workers
leave one job for another or require time to acquire new skills and adapt to a new workplace.
These factors mainly account for structural and frictional unemployment, which together
constitute natural unemployment.
In addition, factors such as a decline in aggregate demand, reduction in production,
underdevelopment or inefficiency of enterprises that generate jobs, and the automation of labor,
including the adoption of robotics and information technologies, also contribute to rising
unemployment levels.
Furthermore, enterprise managers may not be able to lay off all workers at once. As a result,
some employees may not be fully employed during the workday or workweek, leading to hidden
unemployment.
In some countries, there are individuals who are physically able to work but have lost their jobs
and have no interest in seeking employment or receiving unemployment benefits. Such
conditions give rise to stagnant unemployment.
Among the key macroeconomic factors influencing unemployment, the most prominent are
inflation and investment, along with other economic variables.
Inflation refers to the depreciation of a currency due to the supply of money exceeding the
demand for goods and services, leading to a general increase in prices.
Investment, on the other hand, is the allocation of financial resources for the restoration and
expansion of fixed and working capital, as well as for the enhancement of production capacity.
Investments can take the form of cash, bank loans, shares, and other securities. Nominal
https://ijmri.de/index.php/jmsi
volume 4, issue 6, 2025
42
investment refers to investment in monetary terms, while real investment refers to the actual
acquisition of investment resources with that money.
Based on this data and the World Bank Databank, we constructed a model to study the influence
of investment and inflation on unemployment using STATA-14 software.
Methodology .
As mentioned above, we collected relevant data from the World Bank Databank.
Specifically, we used the following indicators:
“Inflation, consumer prices (annual %)”
“Unemployment, total (% of total labor force) (modeled ILO estimate)”
“Foreign direct investment, net inflows (% of GDP)”
From these, we developed an econometric model.
Our dependent variable (Y) is:
Unemployment, total (% of total labor force) (modeled ILO estimate)
The independent variables (X1 and X2) are:
X1 = Inflation
X2 = Investment
Thus, we briefly labeled our variables:
Y = Unemployment
X1 = Inflation
X2 = Investment
Unemployment it = B0 + B1*Inflation it + B2* Investment it + u it
These statistical data were obtained from the World Bank’s World Development Indicators
(WDI) database. The data were collected for the eight member countries of the Shanghai
Cooperation Organization (SCO). These countries include Azerbaijan, Uzbekistan, the Russian
Federation, the Islamic Republic of Iran, China, Kyrgyzstan, Tajikistan, and Kazakhstan. The
dataset covers the period from 2004 to 2023.
After inputting our model into STATA, we tested for the presence of multicollinearity,
heteroskedasticity, and autocorrelation issues. The results revealed the presence of
heteroskedasticity in the model. Therefore, we transformed the variables by applying logarithmic
conversion.
Following that, we began our analysis by conducting descriptive statistics on the model.
To do this, we used the sum command in STATA-14, which generated the following statistical
indicators:
Variable
Obs Mean
Std. Dev. Min
Max
lnunemployment 160 1.770793 0.3999066 0.8175747 2.615935
lninflation
136 1.939353 0.8446514 -0.9445917 3.772496
lninvestment
147 1.003031 0.9917011 -1.624575
3.995722
The values in the Mean column represent the arithmetic average of each variable, while the Std.
Dev. column displays the standard deviation, reflecting the average dispersion of the
observations around the mean.
Next, we will examine the distribution series and normality characteristics of the dataset.
To accomplish this, we will generate histograms of the variables in STATA-14 and compare
them against the graphical normal distribution curve.
https://ijmri.de/index.php/jmsi
volume 4, issue 6, 2025
43
0
.5
1
1.
5
2
D
en
si
ty
1
1.5
2
2.5
lnishsizlik
0
.2
.4
.6
D
en
si
ty
-1
0
1
2
3
4
lninflatsiya
0
.1
.2
.3
.4
.5
D
en
si
ty
-2
0
2
4
lninvestitsiya
As can be seen, Diagram 1 indicates that unemployment approximately follows a normal
distribution. Similarly, Diagram 2 also suggests that inflation aligns well with a normal
distribution pattern. Compared to Diagram 1, Diagrams 2 and 3 exhibit a higher degree of
normality, showing a closer fit to the theoretical normal curve.
To assess normality more precisely, we used the swilk command in STATA-14, which generated
a statistical table for formal analysis of the normal distribution.
According to the results in this table, the unemployment and inflation variables appear to
conform to the assumptions of normal distribution. However, the investment variable does not
satisfy the conditions of normality.
https://ijmri.de/index.php/jmsi
volume 4, issue 6, 2025
44
To understand the relationship among these three economic indicators, we first turned to the use
of a scatter plot for visual inspection.
1
1.
5
2
2.
5
ln
is
hs
iz
lik
-1
0
1
2
3
4
lninflatsiya
1
1.
5
2
2.
5
ln
is
hs
iz
lik
-2
0
2
4
lninvestitsiya
It is clearly observed that inflation has an impact on unemployment, and this effect appears to be
statistically significant. However, we can conclude that investment does not have a significant
impact on unemployment.
In the first scatter plot, although the data points are not perfectly aligned, they generally trend
upward and to the right, indicating a positive relationship. From this, we can infer that as
inflation increases, unemployment also tends to increase — suggesting a direct correlation in
Scatter Plot 1.
In Scatter Plot 2, however, the data points appear randomly scattered, showing no clear pattern.
This leads us to conclude that there is no meaningful relationship between investment and
unemployment in this case.
To further investigate the interaction among these three economic indicators, we must calculate
and analyze their correlation coefficients.
It is evident that there is a relationship between inflation and unemployment, and we can
describe this as a positive correlation.
As for the relationship between unemployment and investment, the correlation coefficients are
https://ijmri.de/index.php/jmsi
volume 4, issue 6, 2025
45
close to zero, suggesting a weaker relationship. However, the direction of this weak relationship
appears to be slightly negative, allowing us to cautiously conclude the presence of a mild
negative correlation.
Now, in order to fulfill the main objective of our study, we proceed to run a regression analysis
using our specified model. The outcomes of this regression will be presented and interpreted in
the next section.
Analysis of Results. Finally, we arrive at the regression analysis stage. To begin our
interpretation, we must first become familiar with the initial regression output generated from
our model.
As observed in the regression output table, we obtained the coefficients B₀, B₁, and B₂. All three
coefficients are statistically significant, as their p-values are less than 0.05, indicating a high
level of reliability.
This allows us to conclude the following interpretations:
If unemployment increases by one unit, inflation increases by approximately 0.2 units.
Conversely, if unemployment rises by one unit, investment decreases by about 0.081 units.
To evaluate the overall quality of the model, it is essential to test for potential problems such as
multicollinearity and heteroskedasticity.
Therefore, we conducted the Variance Inflation Factor (VIF) test to assess multicollinearity, and
the Breusch-Pagan test to detect the presence of heteroskedasticity.
https://ijmri.de/index.php/jmsi
volume 4, issue 6, 2025
46
It is evident that our model does not suffer from heteroskedasticity, as the p-value obtained from
the Breusch-Pagan test is greater than 0.05. This suggests that the variance of the error terms is
constant across observations — an important assumption in linear regression.
Likewise, the issue of multicollinearity is not present in our model. Had the VIF (Variance
Inflation Factor) values exceeded 10, we would have concluded that multicollinearity exists.
However, since the VIF values are well below this threshold, we can confidently state that
multicollinearity is not a concern.
In addition, it is important to assess whether there is any autocorrelation problem in the model.
Therefore, checking for the presence or absence of autocorrelation is a logical next step to
validate the reliability of our regression model.
As seen in the table, there is clear evidence of autocorrelation in the model, even after the tenth
observation. This indicates a violation of one of the key classical regression assumptions.
To address this issue, we applied a solution by implementing lagging (lag transformation) to
mitigate the effects of autocorrelation in the model.
Now we can confidently state that there is no autocorrelation present in our model.
Conclusion. This research was conducted using data from Shanghai Cooperation Organization
(SCO) member countries over the period 2004–2023. The focus of the study was to examine the
main macroeconomic factors influencing unemployment. To explore these relationships,
regression analysis was performed using STATA-14 software.
https://ijmri.de/index.php/jmsi
volume 4, issue 6, 2025
47
According to the results, all regressors were found to be statistically significant. The analysis
showed that an increase in inflation leads to a rise in unemployment, whereas an increase in
investment results in a decline in unemployment.
References
[1] http://www.lex.uz
[2] http://www.stat.uz
[3] http://www.wordbank.org
[4] http://www.imf.org
[5] O’zbekiston iqtisodiyoti.2016-yil uchun axborot tahliliy byulleten,Toshkent-2016.
