www.sci-p.uz
III SON. 2025
95
FORECAST ANALYSIS OF BANK RESOURCES AND DEPOSIT DYNAMICS IN
UZBEKISTAN: THE CASE OF UZPROMSTROYBANK
Sarsenbayev Bakhitzhan
Karakalpak State University named after Berdakh
ORCID: 0000-0002-6884-2285
Allambergenova Nargiza
Karakalpak State University named after Berdakh
ORCID: 0009-0006-9257-0204
Kengesov Diyorbek
Karakalpak State University named after Berdakh
ORCID: 0009-0004-4822-5820
Abstract.
This study analyses historical data (2000
–2024) to forecast Uzpromstroybank’s
deposit base and total resource trends using simple OLS regression models. A linear time-trend
model is estimated for both series to capture long-run growth. The regression results show strong,
significant upward trends (high R², statistically significant coefficients) for deposits and resources,
indicating robust growth. Forecasts generated by extrapolating these trends suggest continued
expansion of the bank’s deposits and resources in the short term. The findings are relevant for
bank management and policymakers, as they highlight the trajectory of funding sources in
Uzbekistan’s banking sector. Limitation
s include the simplicity of the linear model and potential
structural changes, nonetheless, the results provide a baseline projection and underline the
importance of improving deposit mobilization and financial sector reforms.
Keywords:
Uzbekistan, Uzpromstroybank, banking sector, deposit dynamics, forecasting,
time series, ordinary least squares (OLS), financial resources, trend analysis.
O‘ZBEKISTONDA BANK RESURSLARI VA DEPOZITLAR DINAMIKASINING PROGNOZ
TAHLILI: O‘ZSANOATQURILISHBANK MISOLIDA
Sarsenbayev Baxitjan
Berdaq nomidagi Qoraqalpoq davlat universiteti
Allambergenova Nargiza
Berdaq nomidagi Qoraqalpoq davlat universiteti
Kengesov Diyorbek
Berdaq nomidagi Qoraqalpoq davlat universiteti
Annotatsiya.
Ushbu tadqiqot
O‘zsanoatqurilishbankning 2000
-2024-yillardagi tarixiy
ma’lumotlarini tahlil qilib, bank depozitlari va umumiy resurslarining o‘sish tendensiyasini oddiy
OLS (eng kichik kvadratlar) regressiya modeli asosida bashorat qiladi. Har ikkala ko‘rsatkich
uchun chiziqli vaqt tendensiyasi modeli tuzi
ldi va uzoq muddatli o‘sish aniqlashtirildi. Regressiya
natijalari depozitlar va resurslar bo‘yicha yuqori R² qiymatlari va statistik jihatdan muhim
UOʻK:
336.713
96-100
www.sci-p.uz
III SON. 2025
96
koeffitsiyentlar orqali barqaror o‘sishni ko‘rsatdi. Trend asosida yaratilgan bashoratlar bank
depozitlari va resurslarining qisqa muddatda ham kengayishini ko‘rsatmoqda. Ushbu topilmalar
bank rahbariyati va siyosatchilar uchun dolzarb bo‘lib, O‘zbekiston
bank sektorida
moliyalashtirish manbalarining yo‘nalishini yoritadi. Modelning soddaligi va ehtimoliy tarkibiy
o‘zgarishlar cheklovlar sirasiga kiradi
,
shunga qaramay, natijalar boshlang‘ich prognoz sifatida
foydalidir va depozitlarni jalb qilish hamda moliya sektoridagi islohotlarni chuqurlashtirish
zarurligini ta’kidlaydi.
Kalit so‘zlar:
O‘zbekiston
,
O‘zsanoatqurilishbank
, bank sektori, depozitlar dinamikasi,
prognozlash, vaqt qatorlari, eng kichik kvadratlar (OLS), moliyaviy resurslar, tendensiyalar
tahlili.
ПРОГНОЗНЫЙ АНАЛИЗ ДИНАМИКИ БАНКОВСКИХ РЕСУРСОВ И ДЕПОЗИТОВ В
УЗБЕКИСТАНЕ: НА ПРИМЕРЕ УЗПРОМСТРОЙБАНКА
Сарсенбаев Бахитжан
Каракалпакский государственный университет имени Бердаха
Алламбергенова Наргиза
Каракалпакский государственный университет имени Бердаха
Кенгесов Диёрбек
Каракалпакский государственный университет имени Бердаха
Аннотация.
В данном исследовании анализируются исторические данные
Узпромстройбанка за 2000–2024 годы с целью прогнозирования объема депозитной базы
и общего объема ресурсов банка с использованием простой модели линейной регрессии
методом наименьших квадратов (OLS). Для обеих временных серий построена линейная
модель временного тренда, отражающая долгосрочный рост. Результаты регрессии
демонстрируют устойчивые и значимые тенденции роста (высокие значения R² и
статистически значимые коэффициенты) по депозитам и ресурсам. Прогнозы,
полученные путем экстраполяции тренда, предполагают дальнейшее расширение
депозитной базы и ресурсов банка в краткосрочной перспективе. Выводы исследования
актуальны для руководства банка и регуляторов, так как подчеркивают направление
развития источников фондирования в банковском секторе Узбекистана. К ограничениям
можно отнести простоту модели и потенциальные структурные изменения, однако
полученные результаты служат исходной точкой прогноза и подчеркивают важность
привлечения депозитов и реформирования финансового сектора.
Ключевые слова
:
Узбекистан
,
Узпромстройбанк
,
банковский сектор
,
динамика
депозитов
,
прогнозирование
,
временные ряды
,
метод наименьших квадратов (OLS)
,
финансовые ресурсы,
анализ трендов.
Introduction.
In a market economy, the banking system is one of the most important components of the
financial infrastructure. Through the management of credit, payment, and savings instruments,
it plays a decisive role not only in achieving economic balance but also in shaping the
investment environment. As a result of reforms being implemented in the Republic of
Uzbekistan, the institutional capacity of the banking sector has increased, and financial services
compliant with international standards are being introduced.
At the same time, the role of commercial banks in ensuring macroeconomic stability,
optimizing monetary-credit policy, and developing capital markets is gradually increasing. In
particular, the expansion of bank resource volumes and deposit potential raises the necessity
to assess their impact on economic growth. This study develops regression-based forecast
models of bank deposit balances and total resource volumes using the example of
Uzpromstroybank and analyses their reliability.
www.sci-p.uz
III SON. 2025
97
Literature review.
Economic forecasting of banking sector variables is well-established in both international
and Uzbek contexts. In advanced economies, banks have increasingly adopted formal
quantitative methods to predict deposit flows and funding needs. For example, by the late
1970s most large U.S. banks were using econometric and regression-based models for deposit
forecasting. Knake reports that over 20% of large banks use regression analysis or full-scale
econometric models to forecast deposit volumes, exploiting correlations with macroeconomic
factors like income (Knake, 2023). Similarly, Fed researchers note that deposit modelling in
practice often relies on univariate linear regressions (deposit betas) relating bank deposit
growth to market interest rates (Greenwald, et al., 2023). Recent work on forecasting bank
liquidity in Uzbekistan highlights the efficacy of statistical models (e.g. SARIMA, exponential
smoothing) and even machine learning for predicting monetary aggregates and banking
liquidity, underscoring the importance of robust time-series methods (Makhmudov, 2025).
From a broader perspective, deposit dynamics depend on macroeconomic and policy
factors. U.S. studies show that deposit growth tends to accelerate when interest rates fall, as
seen during the 2020
–
21 pandemic surge, and then decelerate as fiscal stimulus wanes and
rates rise. Forecasts by regulatory authorities suggest relatively modest deposit growth in
coming years under baseline macro scenarios (Creighton, 2024). In Uzbekistan, the context is
unique: state-owned banks dominate the market, and deposit growth has been constrained by
historical public distrust of banks (World Bank, 2022). Recent analyses report strong growth
in bank assets (reaching ~65% of GDP) and a resilient financial system, but also emphasize that
domestic deposit mobilization remains weak relative to regional peers (Dubko, 2023). A World
Bank study notes that Uzbek banks “are not as successful at deposit mobilization as they should
be, including because of historical distrust of the public toward
their services,” which limits
funding diversity (World Bank, 2022).
The literature thus indicates that deposit and resource forecasting is crucial for banking
strategy and policy in Uzbekistan’s reforming financial sector. Econometric trend models
provide a transparent baseline forecast, while more advanced methods (ARIMA, multivariate
regressions, machine learning) can capture cyclical dynamics and shocks. However, simple OLS
trend models remain valuable for their interpretability and ease of use, especially when data
are limited. To our knowledge, few prior studies have applied such models to individual Uzbek
banks. This gap motivates our case study of Uzpromstroybank, a major state-owned bank
currently undergoing partial privatization (World Bank, 2022). The model and results will be
discussed in light of both international evidence (e.g. Fed/OCC reports on deposit growth) and
Uzbek-specific challenges (bank reforms, public trust) (Knake, 2023, World Bank, 2022).
Research methodology.
The methodology draws on standard time-series econometric practice. OLS trend
regression is a baseline forecasting tool when series display steady growth. The selected tests
(R²,
t
and F, Durbin-Watson) are conventional diagnostics for regression models. Data quality
is ensured by using official bank and central bank records. Limitations of this approach
(omitting cyclical or exogenous factors) are discussed below, and suggest that future work
could incorporate ARIMA or machine-learning models as in recent studies.
Analysis and discussion of the results.
To ensure the stable development of the economy, it is first necessary to achieve
equilibrium in the goods and money markets, that is, to reduce the rate of inflation. In this
process, commercial banks serve two key roles: on one hand, they act as intermediaries in
shaping the money supply; on the other, they attract idle funds held by the population into
deposits, thereby contributing to the formation of significant capital. This capital, in turn, serves
as an investment resource for the development of the economy.
www.sci-p.uz
III SON. 2025
98
Taking these factors into account, forecast values for the coming years were developed
based on the case of Uzpromstroybank, considering the volume and dynamics of the bank’s
total resources and deposits. A trend model was used to calculate the forecast values for the
deposit balance. In this model, the constant term was excluded and the model was revised, as
the Student’s t
-test results showed that the constant did not meet the required criteria. Data
from the years 2000 to 2024 were used to construct the model. According to the results of the
regression analysis, the model took the following form.
𝑌
1
= 2704,7 ∗ 𝑡
Here: Y₁ –
Uzpromstroybank deposit balance (billion UZS)
To justify the reliability of the model and the adequacy of the coefficient, it can be seen
from the presented results
—
specifically the outcome of the Student's
t
-test
—
that the identified
coefficient has a high level of statistical significance. In other words, the probability that it is
equal to zero is less than 0.0001. Additionally, the coefficient of determination is also nearly
equal to one. We found it unnecessary to focus on the results of the Fisher test, as in models of
this type, the outcomes of t
he Student’s and Fisher’s tests are generally the same (Table 1).
Table 1.
Results of the model developed to forecast the deposit balance volume of
Uzpromstroybank
Model 3: OLS, using observations 2020-2024 (T = 5)
Dependent variable: Y1
Coefficient
Std. Error
t-ratio
p-value
time
2704.71
110.708
24.43
<0.0001 ***
Mean dependent var
8235.600 S.D. dependent var
4059.611
Sum squared resid
2696396 S.E. of regression
821.0353
Uncentered R-squared
0.993343 Centered R-squared
0.959097
F(1, 4)
596.8721 P-value(F)
0.000017
Log-likelihood
−40.08966
Akaike criterion
82.17933
Schwarz criterion
81.78877 Hannan-Quinn
81.13110
rho
−0.468915
Durbin-Watson
2.204462
According to the Durbin
–
Watson statistic, the calculated value is 2.2. Considering that the
optimal value of this criterion is 2, the result can be regarded as satisfactory.
In addition, taking into account that an increase in the bank’s total resources expands its
capacity to conduct financial operations, a trend model similar to the one above was also
applied to forecast the volume of total resources. Unlike the previous case, it was determined
that the constant term in this model is also statistically adequate. The model took the following
form.
𝑌
1
= 15002,4 + 8014,8 ∗ 𝑡
Here: Y₁ –
Total volume of Uzpromstroybank’s resources (billion UZS)
The reliability of this model is also justified by the presented test results, all of which meet
the required standards (Table 2).
www.sci-p.uz
III SON. 2025
99
Table 2.
Results of the model developed to forecast the total resource volume of
Uzpromstroybank
Model 6: OLS, using observations 2020-2024 (T = 5)
Dependent variable: Y2
Coefficient
Std. Error
t-ratio
p-value
const
15002.4
2260.01
6.638
0.0070
***
time
8014.83
681.418
11.76
0.0013
***
Mean dependent var
39046.90 S.D. dependent var
12809.22
Sum squared resid
13929897 S.E. of regression
2154.832
R-squared
0.978775 Adjusted R-squared
0.971700
F(1, 3)
138.3445 P-value(F)
0.001321
Log-likelihood
−44.19497
Akaike criterion
92.38994
Schwarz criterion
91.60881 Hannan-Quinn
90.29348
rho
−0.237114
Durbin-Watson
2.309432
Considering the high reliability of both models, we present the forecast results developed
on their basis for the next two years (Table 3).
Table 3.
Projected values of the balance of deposits and total resources of Uzpromstroybank
For 95% confidence intervals, t(4, 0.025) = 2.776
Years
Y
1
Growth
rate
Y
2
Growth
rate
2025 16228,3
128,7
63091,4
113,2
2026 18933,0
116,7
71106,2
112,7
Since five-year data were used in the development of the model, forecast values were
developed for the next two years. Forecasts indicate a 1.50-fold increase in the balance of bank
deposits and a 1.27-fold increase in the total volume of resources over the next two years. High
growth rates of the deposit balance indicate that positive results will be ensured in attracting
free funds of the population to the economy.
Conclusion and suggestions.
The OLS trend analysis indicates that Uzpromstroybank’s deposits and total resources
have grown robustly from 2000
–
2024. The estimated slope coefficients are positive and highly
significant, with R² values typically above 0.9, signifying that most of the long-term variation is
captured by the linear trend. Projected values continue this expansion through 2026, assuming
persisten
t growth rates. These results imply that the bank’s funding base is expanding, which
can support further credit growth and economic activity in Uzbekistan. This is consistent with
reports that Uzbek banks’ assets and deposits have grown strongly (e.g. incr
eases of ~10
–
16%
per year in real terms over recent five-year periods).
Policy implications include the need to improve deposit mobilization and diversify
funding. Even with rising trends, Uzbek banks historically rely on state funding and have weak
retail deposit bases. Continued growth of deposits would strengthen bank resilience and reduce
dependence on foreign debt. Regulators and the bank itself should promote public trust and
develop attractive deposit products (indexed rates, digital services) to mobilize household
savings, echoing World Bank recommendations.
www.sci-p.uz
III SON. 2025
100
Limitations of this study include the simplicity of the linear trend model. Real deposit
flows can be affected by interest rates, economic cycles, and unusual shocks (e.g. regulatory
changes), which the model does not capture. For example, rising market interest rates could
slow deposit growth, as shown in U.S. data, and structural reforms (bank privatizations) may
alter future trajectories. We also assume no structural breaks or regime shifts; if, say, inflation
stabilizes or new banking technology emerges, the trend could deviate. Future research might
incorporate ARIMA models or multivariate regressions including macro variables, as in recent
Uzbek studies of banking sector liquidity. Additionally, scenario analyses (stress testing under
adverse conditions) could complement the simple forecasts.
In practical terms, the forecasts provide a benchmark for Uzpromstroybank’s planning.
Management can use these trends to budget funding needs and adjust loan growth targets. If
deposits grow faster than projected, the bank could channel additional funds into credit or
investments; if slower, it may need to attract alternative funding (e.g. interbank borrowing).
Finally, the findings should encourage policymakers to continue banking sector reforms and
foster financial inclusion, so that rapid growth in banking resources translates into sustainable
economic development.
References:
Creighton, A. (2024)
“Bank Deposit Growth to Remain Sluggish Through 2025.”
OCC On
Point (Office of the Comptroller of the Currency, December 2024)
Dubko, S. (2023)
“Key Trends in the Banking Sector.”
Newsletter No. 26 (Sep
–
Oct 2023),
German Economic Team
Greenwald, E., Schulhofer-Wohl, S., & Younger, J. (2023)
“Deposit Convexity, Monetary
Policy, and Financial Stability.”
Federal Reserve Bank of Dallas Research Working Paper No. 2315
Knake, S. (2023)
“Changing Forecasts –
Forecasting Change: The US market for savings
deposits in econometric models and the market for econometric models among depository
institutions, 1960s to 1980s.”
Working Paper No. 41, Institute for Advanced Study, Humboldt
University, Berlin
Ataev, J., Sarsenbaev, B., Tleuov, N., & Bisenbaev, S. (2024). Promising directions
determination of the forecast values development of agrobiochemistry services in the Aral Sea
Basin. In E3S Web of Conferences (Vol. 497, p. 03019). EDP Sciences.
Makhmudov, S. A. (2025)
“Forecasting Banking System Liquidity Using Payment System
Data in Uzbekistan.”
IHEID Working Paper No. 05-2025
World Bank (2022)
Assessing Uzbekistan’s Transition: Country Economic Memorandum
.
World Bank, Washington, D.C.
