https://ijmri.de/index.php/jmsi
volume 4, issue 2, 2025
381
EXPORT POTENTIAL AND EXPORT DIVERSIFICATION STRATEGIES OF
UZBEKISTAN
Otamurodov Anvar Xamidullo ugli
The University of World Economy and Diplomacy
The faculty of international economy and management
Third year student
1.
Introduction
Export of manufactured goods has a significant role in economic growth. Majority of Asian
countries including Uzbekistan have substantial amount of mineral and oil and gas reserves
which are important economic sectors. According to the report written by Dean Belder for
Investing News Network (September 24, 2024) Uzbekistan is the world's tenth-largest gold
producer, mining about 100 metric tons of gold in 2023, and holds some of the largest reserves in
the world at 4500 MT. Uzbekistan’s exporter rank among 138 countries is 73 and importer rank
is 69. China ($1,744,423,820), Russia ($1,703,958,812) and Turkey ($1,638,976,954) are top
three export countries to which Uzbekistan exports precious stones and metals, cotton, copper
et.al.
Considering the fact that Uzbekistan’s economy highly depends on export of precious metals
which are limited, it has become imperative for Uzbekistan government to look at diversifying
its domestic economy. In this way, the International Monetary Fund [IMF] (
considers that economic diversification is an inevitable policy in gold-exporting Asian countries
because it reduces the impact of the external shocks associated with gold markets on the
economy. To ensure the economy’s resistance to unpredictable fluctuations in prices and
limitation in resources export diversification, can be considered as the new driving engine of
economic growth in Uzbekistan. According to the report by Central Bank of Uzbekistan (April,
2024) during March Uzbekistan’s foreign exchange reserves grew to $34.2 billion due to the
record-breaking increases in gold prices. Meanwhile, the physical volume of precious metal
in the reserves fell by 10.9 tons, marking its lowest level since May 2022.
To our knowledge there has been limited research on export diversification strategies of
Uzbekistan. Therefore, our research will contribute to the existing literature in a couple of ways.
Firstly, diversifying the range of export goods helps to achieve the transition of economy away
from the gold-sector. Secondly, we seek to identify additional countries to the existing
counterparts through mentioning transport corridors of Uzbekistan.
2.
Literature review
There have been conducted several studies on export diversification strategies of different
countries and their role not only in global practice but also their national economy which
depends on several factors such as GDP, inflation, unemployment, export, import et.al. In one of
the recent studies (Abdessalem Gouider & Hedi Ben Haddad, 2020) the potential spatial
https://ijmri.de/index.php/jmsi
volume 4, issue 2, 2025
382
diversification of manufactured goods’ exports in Saudi Arabia was explored. To account for the
spatial interactions of Saudi’s manufactured goods’ exports, they used a panel Spatial
Autoregressive (SAR) model for 77 trade partners over the period 2000–2016. The empirical
results suggest, firstly, the existence of spatial interdependence among Saudi’s manufactured
goods’ exports. Secondly, they found that the exogenous variables including GDP, GDP per
capita, trade freedom, bilateral exchange rate, and trade intensity index exert strong spillover
effects on bilateral Saudi’s manufactured goods’ exports. Finally, the study demonstrated
evidence of the highest potential with 34 out of 77 partners. This finding had important
implications for policymakers, mainly in terms of development of the domestic manufacturing
sector and geographic reallocation of Saudi’s manufactured goods’ exports.
On the other hand, Elodie Mania and Arsene Rieber (2019) also mentioned their paper the
purpose of which was to revisit that relationship by questioning the sustainability of such a
strategy. Drawing on a balance of payments constrained growth model, they compared the re-
composition of productive capacities that follows export diversification with the evolution of
countries’ external constraints. Based on econometric estimates of panel data, the lessons of the
model allow them to analyze and compare, over the period 1995–2015, export diversification in
three samples of developing countries, namely: Latin America, Sub-Saharan Africa and
Developing Asia.
Another research conducted by Federico Bonaglia and Kiichiro Fukasaku (2003) discussed
major policy issues related to commodity dependence and export diversification in low-income
countries. Contrary to some widely-held view, it argued that natural resources were not
necessarily a curse - that they do not condemn low-income countries to underdevelopment but
can provide rather a basis for sustained export-led growth. Natural resource-based sectors have
potential for export diversification. The OECD mirror trade data suggest that many different
routes to diversification exist, including resource-based manufacturing and processing of primary
products. However, these opportunities were not being exploited in many low-income countries.
This is because export diversification is typically a slow process, and this process needs to be
sustained by an appropriate and coherent strategy, characterised by a combination of vision, co-
ordination and management of conflicting interests.
Above mentioned researches clearly described how export diversification influenced on
countries’ economy ranging from low income countries such as Africa to high income countries
which is Saudi Arabia analyzing its effects and strategies used by those countries and proved
their suggestions with data and statistics. However there is a few research on export
diversification strategies of middle Asia countries, particularly Uzbekistan and its global effects.
Current research tries to fulfil this gap by providing official statistics by World Trade
Organization (WTO), World bank (WB) and other international organizations.
3.
Methodology
3.1.
Theoretical framework
The Ordinary Least Squares (OLS) method is a fundamental statistical technique used to
estimate linear regression models. Initially introduced by Carl Friedrich Gauss in the early 19th
century as part of his work on error theory, its primary objective is to minimize the sum of the
squared differences between the observed values of the dependent variable and the values
predicted by the model. OLS delivers unbiased and efficient estimates of regression coefficients,
assuming that certain standard conditions are met, including linearity of the model, error
independence, normality of errors, and homoscedasticity (constant variance of errors).
The widespread adoption of OLS can be attributed to its flexibility, ease of interpreting results,
and effectiveness in analyzing data from fields such as economics, social sciences, and finance.
https://ijmri.de/index.php/jmsi
volume 4, issue 2, 2025
383
This method enables researchers to identify and measure relationships between variables, making
it an essential tool in empirical economics, particularly for evaluating the export potential of
nations.
OLS is commonly applied in studies focused on international trade, export capacity, and the
influence of economic factors on trade flows. For example, Felipe, L (2012) used OLS model in
one of his studies to explore the link between export diversification and economic growth in
developing countries. It showed that diversification into a broader range of sectors can support
sustained economic growth.
In addition, Ciccone, A., & Papaioannou, E. (2009) applied OLS to investigate how the structure
of a country's economic sector and human capital contribute to economic growth, with a focus on
education and technological advancement.
In his paper, Rodrik (2006) used OLS models to analyze the role of industrial policy and export
diversification in developing countries’ growth, focusing on the importance of adding new
productive capacities and diversifying export structures.
Kim, M. (2008) used OLS regression to analyze the factors influencing export performance in
emerging market economies, such as exchange rate stability, trade openness, and market
diversification.
We will also use OLS to study export potential of Uzbekistan by analyzing relationship between
export performance and key economic some factors. This approach will allow us to identify
statistically significant factors influencing export performance and develop recommendations for
increasing the country’s export potential.
3.2. Empirical framework
Multivariate time series model is used in this research to determine the relationship between
Uzbekistan's exports and other indicators, such as: GDP per capita, exports by industry, inflation,
exchange rate (US dollar), foreign direct investment, and export growth. Here, Uzbekistan’s
export volume is dependent variable, whereas others are independent variables. We put H
0
and
H
1
hypothesis:
H
0
means there is no relationship between export volume and other independent variables.
H
1
means there is relationship between export volume and other independent variables.
We use linear model to indicate a linear relationship between dependent variable and
independent variables and the relationship can be expressed as follows:
y=β
0
+β
1
x
1
+β
2
x
2
+
⋯
+β
n
x
n
+ ϵ
Here:
y is dependent variable which is export volume
x
1
, x
2
… x
n
are independent variables (exports by industry, inflation, exchange rate (US dollar),
foreign direct investment, and export growth)
β
0
is the intercept (constant term)
β
1
, β
2
, …, β
n
are the coefficients of the independent variables
ϵ is the error term
We also use VAR model to capture the linear interdependencies among multiple time series data.
It is an extension of univariate autoregressive models (AR models) to multivariate time series.
The key idea behind the VAR model is that each variable in the system is modeled as a linear
function of its own past values and the past values of all other variables in the system and it can
be written as:
https://ijmri.de/index.php/jmsi
volume 4, issue 2, 2025
384
We use STATA 18 to create VAR model depending on our collected data and tested the
variables based on augmented Dickey-Fuller test and Johansen cointegration test. The ADF test
is used to determine whether a time series is stationary or has a unit root and the regression can
be expressed as:
The Johansen cointegration test is used to determine the presence and number of cointegration
relationships among multiple time series. The model form is:
This model has significant role in time series econometrics, particularly when analyzing the
long-run equilibrium relationships between non-stationary variables.
4.
Results
As mentioned before, Uzbekistan’s export volume is chosen as a dependent indicator which was
$3,393,714,865 in 2000 and this figure reached $24,066,920,244.36 in 2023. Results show
whether independent variables affected to this growth and if yes how were their significancy.
For the beginning we implement stationary test:
Variables
Test
statistics
value
1
%
critical
value
5
%
critical
value
1 0 %
critical
value
Number of
observations
Mac
Kinnon's
p-value
Degree
of
differentiation
(0, 1, 2)
Export
-4.078
-3.750 -3.000 -2.630 20
0.0011
1
GDPpc
-5.165
-3.750 -3.000 -2.630 22
0.0000
2
Inflation
-5.784
-3.750 -3.000 -2.630 21
0.0000
1
Ex_rate
-4.487
-3.750 -3.000 -2.630 21
0.0001
1
FDI
-7.547
-3.750 -3.000 -2.630 22
0.0000
1
Unemployment - 4.897
-3.750 -3.000 -2.630 23
0.0000
0
Export_growth -4.212
-3.750 -3.000 -2.630 23
0.0019
0
Table 1. Results of ADF test
According to the table, unemployment and export growth are stationary. As they have
differentiation degree of 0, they can be directly used without differencing. Variables which are
export, inflation, exchange rate and FDI are non-stationary but can be stationary after first
differencing. GDP per capita requires two differencing steps to become stationary and exhibits a
more pronounced trend or structural pattern, which necessitates higher-order differencing to
eliminate. The table given above not only shows whether the statistical data is stationary or non-
https://ijmri.de/index.php/jmsi
volume 4, issue 2, 2025
385
stationary, but also gives an opportunity to change non-stationary data to stationary data by
differencing means and variances. The test statistic should be smaller than all the critical values
(1%, 5%, 10%). In addition, the McKenney p-value must be below 0.05. Based on the results in
Table 1, the variables Export, GDP per capita, FDI, Inflation, and Exchange rate were identified
as non-stationary by this test. However, all these variables became stationary after applying first
and second differences. The test further confirms that the independent variables, such as
Unemployment and Export growth, are stationary.
Variables
(1)
(2)
(3)
(4)
(5)
(6)
(1) Export
1.000
(2) GDP per capita 0.706* 1.000
(0.000)
(3) Inflation
-
0.618*
-
0.655*
-
0.598*
1.000
(0.003) (0.000) (0.003)
(4) Exchange rate
0.756* 0.296
0.874* -0.376 1.000
(0.000) (0.240) (0.000) (0.241)
(5) FDI
0.901* 0.758* 0.857* -
0.542*
0.667* 1.000
(0.000) (0.000) (0.000) (0.009) (0.000)
(6) Unemployment -
0.771*
-
0.757*
-
0.687*
0.841* -0.298 -
0.589*
(0.000) (0.000) (0.001) (0.000) (0.056) (0.001)
*** p<0.01, ** p<0.05, * p<0.1
Table 2. Correlation test
Correlation test shows whether there is relationship between the factors: if the result is positive,
it means there is a direct relationship, if the result is negative, then there is an indirect
relationship between factors. For example, there is a direct relationship between export volume
and GDP per capita. The correlation between export volume and inflation is negative which
means that they are indirectly related to each other. Exchange rate and FDI are directly related to
export volume, while the relationship between export volume and employment is negative.
Ex
Coef.
St.Err.
t-
value
p-
value
[95%
Conf
[Interval] Sig
GDPpc
-
1808756.6
798465.98 -1.88
.058
-
3879521.8
128745.89 *
Inflation
29785495 28974562 1.56
.108
-15874258 85874216
Ex_rate
-
612874.89
208547.78 -3.41
.007
-
1135794.2
-
198746.78
***
FDI
.874
.57
2.19
.044
-.014
1.897
*
Unemployment -
8.412e+09
1.316e+04 -6.41
0
-
1.074e+08
-
5.554e+07
***
Export_growth -11235478 15789124 -0.66
.444
-43687154 19214977
Constant
1.188e+10 1.960e+09 6.41
0
7.741e+09 1.657e+12 ***
Mean dependent var
11478214975.27
5
SD dependent var
5874680347.541
R-squared
0.971
Number of obs
24
F-test
241.066
Prob > F
0.000
Akaike crit. (AIC)
1124.972
Bayesian crit. (BIC)
1125.441
*** p<.01, ** p<.05, * p<.1
Table 3. Multiple regression model
https://ijmri.de/index.php/jmsi
volume 4, issue 2, 2025
386
We can start analyzing Table 3 by analyzing Fisher’s test. According to f-criterion, if the p-value
is less than 0.05 hypothesis null should be rejected hypothesis H
1
is accepted. In our condition,
independent variables which are Exchange rate, FDI, unemployment and export growth are
significant as their p value is less than 0.05, while others are statistically insignificant.
Coefficients show how dependent variable can be affected, if we increase independent variables
by one unit. For example, if we increase FDI by one unit, export volume will be increased by
0.874. An increase in unemployment by one unit leads to decrease in export volume by -8.412
units. Similarly, if exchange rate grows by one unit it will result in decreased export volume by
612874.89 units. Additionally, if GDP per capita increases by one unit, it also leads to 1808756.6
units decrease in export volume.
Sample: 2000 to 2023
Log likelihood = -1321.547
FPE = 3.24e+35
Det(Sigma_ml) = 2.62e+58
Number of obs = 24
AIC = 122.3478
HQIC=118.4795
SBIC = 128.0248
Equation
Parms
RMSE
R-sq
chi2
P>chi2
Ex
15
1.6e+08
0.9874
978.4512
0.0000
GDPpc
15
156.21
0.8745
2124.457
0.0000
Inflation
15
8.08743
0.7214
34.57469
0.0000
Ex_rate
15
754.578
0.9616
1023.571
0.0000
Unemployment 15
0.328743
0.8874
378.3254
0.0000
Export_growth 15
12.7135
0.7145
79.87431
0.0000
Coefficie
nt
Std.
err.
Z
P>z
[95%
conf.
interval]
Ex
Ex
L1.
0.851
0.423
1.774
0.054
-0.050
1.498
L2.
1.316
0.874
2.254
0.036
0.097
2.421
GDPpc
L1.
-1612457 1974214
-0.812
0.445
-5127855
2214795
L2.
1798412
1145874
1.354
0.156
-8.28e+06
4654721
Inflation
L1.
2145879
4.02e+08 0.062
0.898
-8.17e+06
8.48e+07
L2.
-1.12e+09 3.87e+04 -3.287
0.002
-1.87e+06
-4.87e+04
Ex_rate
L1.
8.21e+04 312487
2.690
0.007
2.26e+05
1430155
L2.
2.75e+03 2.68e+03 0.887
0.254
-3.23e+05
8.54e+05
Unemployment
L1.
-1.08e+01 8.87e+08 -1.120
0.217
-2.80e+08
6.78e+08
L2.
1.77e+09 1.88e+09 1.212
0.287
-8.78e+08
3.89e+09
https://ijmri.de/index.php/jmsi
volume 4, issue 2, 2025
387
Export_growth
L1.
2.21e+04 3.96e+05 0.578
0.591
-5.78e+07
1.78e+08
L2.
-6.44e+07 2.61e+07 -2.470
0.014
-1.55e+08
-1.73e+07
_cons
-3.74e+09 6.76e+09 -0.512
0.621
-1.65e+10
9.88e+09
Table 4. VAR model
In the VAR model, we observe that all variables have statistically significant coefficients
(P<chi2 = 0.0000). The variables Ex, GDPpc, Ex_rate, and Unemployment exhibit high R-
squared values, indicating that the model explains their variation well. However, the variables
Inflation (R-sq = 0.7214) and Export_growth (R-sq = 0.7145) show lower R-squared values,
suggesting that the model is less accurate for these variables.
For the dependent variable Ex, the coefficient at lag L1 is not significant (P = 0.054), but it
becomes significant at lag L2 (P = 0.036), showing evidence of autocorrelation. The variable
GDP per capita shows no significant coefficients at lags L1 (P = 0.445) and L2 (P = 0.156).
Similarly, for Inflation, the coefficient at L1 is not significant (P = 0.898), but L2 is significant
(P = 0.002), indicating an inverse relationship. The Ex_rate variable is significant at lag L1 (P =
0.007) but not at L2 (P = 0.254).
In the case of Unemployment, both lags (L1 = 0.217; L2 = 0.287) are statistically insignificant.
Finally, for Export_growth, the coefficient at L1 is insignificant (P = 0.591), but it is significant
at L2 (P = 0.014). Based on these results, we can summarize the VAR model as follows:
Y
t
= -3.74e+09 + 1.316L2 Ex
−2
–1.12L2Inflation
−2
+ 8.21L1Ex_rate
−1
–
6.44L2Export_growth
−2
+ ε
During our study we achieved particular results analyzed above, but the analysis has several
limitations that may affect the reliability of the results. Therefore it would be better to explore
this area icluding potential other factors for more extended period, but we were only able to
explore short-term perspectives, not long-term ones. For this reason, we believe that each model
and approach may have limitations. We are going to use the latter approach to include other
external factors in our future research.
5.
Conclusion
Stemming from the findings of our study, we can conclude that Uzbekistan's exports have
experienced positive growth in recent years, playing a significant role in both the short and long-
term economic development of the country. This growth has proven valuable for foreign
economic activity, as it helps increase GDP per capita, reduce unemployment, and boost the flow
of foreign direct investment and foreign currency, which are key indicators of economic
development according to the World Bank methodology. Consequently, one of the most crucial
policy measures for unlocking Uzbekistan's export potential is the expansion of free economic
zones and the establishment of infrastructure to attract foreign investors. Building on the positive
results of this study, further research is ongoing to address unresolved issues, which will be
explored in future studies.
References:
1.
World
Bank.
(n.d.).
World
Development
Indicators
.
Retrieved
from
https://data.worldbank.org/indicator
2.
United Nations Conference on Trade and Development (UNCTAD). (2020).
Foreign
Direct
Investment
Statistics:
The
World
Investment
Report
.
Retrieved
from
https://unctad.org/topic/investment/world-investment-report
https://ijmri.de/index.php/jmsi
volume 4, issue 2, 2025
388
3.
State Committee of the Republic of Uzbekistan on Statistics. (2024).
Economic and
Statistical Overview of Uzbekistan
. Retrieved from https://stat.uz/en
4.
International Monetary Fund (IMF). (2023).
Uzbekistan: Economic Outlook
. IMF
Country Report No. 23/236. Retrieved from https://www.imf.org/en/Countries/UZB
5.
Turaev, A., & Jorayev, R. (2022).
Export Potential and Economic Growth: The Case of
Uzbekistan
.
Journal of International Trade and Economic Development
, 31(4), 320-337.
https://doi.org/10.1080/09638199.2021.1914346
6.
Mavlonov, U. (2021).
Economic Zones and Infrastructure Development in Uzbekistan
.
Asian Economic Policy Review
, 16(2), 214-231. https://doi.org/10.1111/aepr.12359
7.
Kamolov, A. (2023).
The Role of Export Growth in Uzbekistan's Economic Development:
Analysis and Prospects
.
Economics and Business Review for Central Asia
, 15(1), 12-25.
https://doi.org/10.1037/ebr.2023.15.1.12
8.
Usmanov, B. (2020).
Impact of Export and Foreign Investment on Uzbekistan’s GDP
Growth
.
Uzbekistan
Economic
Journal
,
10(3),
45-59.
Retrieved
from
https://www.uzbekeconjournal.org
9.
Zaynab, S. (2022).
Free Economic Zones and Their Role in Attracting Foreign
Investment: A Case Study of Uzbekistan
.
Central Asia Development Review
, 19(2), 81-98.
https://doi.org/10.1080/23178835.2022.1294456
10.
Tashkent Institute of Finance. (2021).
Annual Report on Foreign Direct Investment in
Uzbekistan: Trends and Policies
. Retrieved from
11.
National Bank of Uzbekistan. (2020).
Annual Economic Report: Growth and Investment
Climate
. Retrieved from
12.
International Trade Centre (ITC). (2023).
Uzbekistan Export Performance and Trends
.
Retrieved from
13.
Asadov, A., & Shodiev, M. (2024).
The Development of Uzbekistan’s Export Sector:
Challenges and Strategic Directions
.
Journal of International Trade and Global Economics
,
19(2), 119-133. https://doi.org/10.1504/JITGE.2024.100328
14.
OECD. (2021).
Economic Survey of Uzbekistan 2021: Enhancing Growth through
Structural Reforms
. OECD Publishing. Retrieved from
15.
Murodov, D., & Sattorov, N. (2022).
Improving the Investment Climate in Uzbekistan
through Export-Oriented Economic Zones
.
Economic Development and Policy Review
, 28(3),
76-89. https://doi.org/10.1016/j.edpr.2022.03.004
