Econometric Analysis of Financial Support for Business Entities

Studies Management and Finance Economics, of Journal 0504-2644 (online): ISSN 0490,-2644 (print): ISSN 5202 December 12 Issue 80 Volume 8.317 Factor: Impact ,09-i12-10.47191/jefms/v8 DOI: Article 7076-7661 No: Page JEFMS, Volume 08 Issue 12 December 2025 www.ijefm.co.in Page 7661 Econometric Analysis of Financial Support for Business Entities Nasriddinov Qobilbek Qurbonbekovich The main doctoral student of the Denau Institute of Entrepreneurship and Pedagogy. https://orcid.org/0009-0006-5106-1403 ABSTRACT: The factors affecting the growth of loans allocated as an instrument of financial support by commercial banks to small business entities in Uzbekistan include the following. In particular, the volume of loans granted by commercial banks to small business entities is a random variable, and the independent variables are inflationary expectations of business entities, the key interest rate of the central bank, the volume of deposits of commercial banks, the volume of capital of commercial banks, nonperforming loans (NPL) of commercial banks, the volume of liquid assets of commercial banks, the volume of profits of commercial banks, and the impact of the interest rate on deposits of commercial banks are analyzed based on an econometric model. KEYWORDS: loans, inflation expectation, interest rate, deposit, capital, NPL, liquid assets, profit, deposit interest rate. INTRODUCTION In our study, we evaluated the impact of these factors on the volume of loans allocated to small businesses from commercial banks, the factors affecting it, such as inflation expectations of business entities, the key interest rate of the Central Bank, the volume of deposits of commercial banks, the volume of capital of commercial banks, non-performing loans of commercial banks, the volume of liquid assets of commercial banks, the volume of profit of commercial banks, and the interest rate on deposits of commercial banks based on an econometric model. Today, financial support to small businesses through the allocation of loans by commercial banks is one of the main goals in Uzbekistan. Therefore, based on quarterly data on the volume of loans granted to small businesses from commercial banks as a dependent variable, inflation expectations of business entities as a dependent variable, the key interest rate of the central bank, the volume of deposits of commercial banks, the volume of capital of commercial banks, non-performing loans (NPL) of commercial banks, the volume of liquid assets of commercial banks, the volume of profit of commercial banks and the interest rate of deposits of commercial banks, the factors affecting the volume of loans granted to small businesses and their influence on the autoregressive distributed lag (ARDL) model [1] based on complex econometric methods are selected from monthly indicators from 2020 to 2024 (the main tables of the econometric analysis are given in the appendix). LITERATURE REVIEW In econometrics, in many cases, special models are important in time series analysis to capture complex dynamic relationships and to solve non-stationary problems [2]. Among them, VECM-vector error correction models, VAR-vector autoregression, and ARDL-autoregressive distributed lag models are common econometric tools, each of which offers its own advantages in addressing research questions. Vector autoregression studies any endogenous variable by calculating its past values, as well as the past values of all other endogenous variables in the system, as a linear function, which is suitable for analyzing and forecasting impulse responses based on a series of time series that are interrelated [3]. However, according to standard VAR models, variables are interpreted as stationary [4]. When non-stationary variables exhibit long-term relationships, VECMs are the preferred modeling approach, combining dynamic adjustment and balancing mechanisms into a single model [5]. Unlike VAR and VECM models, ARDL models provide a flexible alternative that is useful for capturing long-term relationships [6]. The recent advancements in ARDL models to capture asymmetric and nonlinear generalizations, quantile-related approaches, and even spatiotemporal dynamics are considered to be recent advances in ARDL models, increasing their utility in various research contexts [2]. ARDL models can also effectively handle serial correlation, endogeneity, and heteroskedasticity [7] [8]. In addition, another advantage of the ARDL model is that it differs from other cointegration methods in that it can accommodate variables with mixed orders of integration [7] [9]. The ability of the ARDL model to integrate variables in different integration patterns gives it an Econometric Analysis of Financial Support for Business Entities JEFMS, Volume 08 Issue 12 December 2025 www.ijefm.co.in Page 7662 advantage over traditional cointegration methods that require all variables to be combined in the same order [10]. In addition, the ability of the ARDL model to simultaneously estimate shortand long-run coefficients in a single equation increases its effectiveness in economic analysis on a large scale [11]. METHODOLOGY The formula (1) above expresses a simple functional relationship and explains the composition of arbitrary and capricious variables. The data we are analyzing are based on monthly data from 2020-2024, and a dynamic time series econometric analysis was performed on the selected data. The reason why we conducted econometric analysis based on the autoregressive distributed lag (ARDL) model is that this model is more suitable for determining the dependencies and errors of variables than other models, in particular the vector auto-regression (VAR) model and the vector error correction (VECM) model [12], and taking into account the possibility of long-term and short-term forecasting, the autoregressive distributed lag (ARDL) model was chosen for our study. We conducted an analysis in this model to determine the long-term and short-term dependencies of arbitrary and capricious variables through the autoregressive distributed lag (ARDL) model of Pesaran and Shin. The convenience of this model is that it allows you to determine whether the selected variables have a long-term and short-term relationship with each other and which variables will have a greater impact in the future. The autoregressive distributed lag (ARDL) model determines the stationary state of the variables through the unit root test for econometric analysis. We express the influence of factors affecting the volume of loans allocated to small businesses in the form of the following simple mathematical formula. LOANS = (inflation expectation, interest rate, deposit, capital, NPL, liquid assets, profit, deposit interest rate). Here, loans – change in the volume of loans allocated to small business entities; inflation expectation – change in inflation expectations of business entities; interest rate – change in the bank's base interest rate; deposit – change in the volume of deposits of commercial banks; capital – changes in the volume of capital of commercial banks; NPL – change in problem loans of commercial banks; liquid assets – change in the volume of liquid assets of commercial banks; profit – change in the volume of profit of commercial banks; deposit interest rate – change in the interest rate on deposits of commercial banks; DISCUSSION AND RESULTS The founders of the autoregressive distributed lag (ARDL) model, Pesaran, Shin and Smith, explained in their research that the selected variables I(0) and I(1) have cointegration with the zero and first differences, respectively. Based on the research of these scientists, we will perform a unit root test for the variables for econometric analysis. (Table 1). Unit Root Test 1 Variables Extended Dickey-Fuller’s Test (ADF) [13] Phillips-Perron’s Test (PP) [14] Zero Order Constant Trend Constant Trend Lnloans -4.395751 *** -6.101731*** -4.315747*** -6.101731*** lndeposit _interest_rate -3.005748** -2.971732** -3.013204** -3.005748** Lnprofit -4.215587*** -4.545047*** -4.229801*** -4.594571** First Order Dlncapital -8.652921*** -8.948335*** -8.596153*** -8.944021*** dlndeposit -6.679079*** -6.617768*** -6.679079*** -6.617768*** dlninflation_expectation -9.990193*** -9.896753*** -12.09719*** -11.94786*** dlninterest_rate -7.545597*** -7.485072*** -7.575607*** -7.509313*** dlnliquid_asset -7.753537*** -7.740128*** -7.755278*** -7.746153*** Dlnnpl -7.121599*** -7.416520*** -7.237121*** -7.482454*** *** indicates statistical significance at 1%, ** indicates statistical significance at 5%, * indicates statistical significance at 10%. 1 Prepared by the author using Eviews-12.0 software based on data from the Central Bank of the Republic of Uzbekistan. Econometric Analysis of Financial Support for Business Entities JEFMS, Volume 08 Issue 12 December 2025 www.ijefm.co.in Page 7663 The Extended Dickey-Fuller and Phillips-Perron tests based on Table 1 show that the variables dlncapital I(1), dlndeposit I(1), dlninflation_expectation I(1), dlninterest_rate I(1), dlnliquid_asset I(1) and dlnnpl I(1) are stationary in the primary difference, in particular in the first order, while lnloans I(0), lndeposit _interest_rate I(0) and lnprofit I(0) are stationary in the zero order. The reliability level of the results of the extended Dickey-Fuller and Phillips-Perron tests is considered reliable at 95%. Since the analysis results show that the statistical significance of all variables is significant at 5%, the reliability level is therefore considered reliable at 95%. According to the nature of the ARDL model, the variables must be stationary in the zero and first orders. After the unit root test, we perform the ARDL bounds test to determine whether there is cointegration between the variables. Based on it, we perform hypothesis testing. We perform hypothesis testing as follows: null hypothesis (N0): there is no long-run relationship, (N1) there is a long-run relationship. The ARDL bounds test is shown in the table below (Table 2). (ARDL) bounds test. Null hypothesis (N0): no long-run relationship exists 2 *** indicates statistical significance at 1%, ** indicates statistical significance at 5%, * indicates statistical significance at 10%. The test results in Table 2 show that there is a long-run relationship between the variables. The bounds test F-statistic is 3.231052, which means that the statistical significance of the relationship value is higher than 10%, which means that there is a long-run relationship between the arbitrary variable lnloans and the arbitrary variables lndeposit _interest_rate, lnprofit, dlncapital, dlndeposit, dlninflation_expectation, dlninterest_rate, dlnliquid_asset and dlnnpl. Since the hypothesis test showed that (N0): rejects the long-run relationship, (N1) accepts the hypothesis that there is a long-run relationship. It is worth noting that this indicates that there is long-run cointegration between the arbitrary and arbitrary variables. Once the existence of cointegration is proven, we determine the long-run correlation (ARDL) model: 𝑙𝑛𝑙𝑜𝑎𝑛𝑠𝑡= 𝜑0+𝛽1𝑙𝑛𝑑𝑒𝑝𝑜𝑠𝑖𝑡_𝑖𝑛𝑡𝑒𝑟𝑠𝑡_𝑟𝑎𝑡𝑒𝑡−1 +𝛽2𝑙𝑛𝑝𝑟𝑜𝑓𝑖𝑡𝑡−1 +𝛽3𝑑𝑙𝑛𝑐𝑎𝑝𝑖𝑡𝑎𝑙𝑡−1 +𝛽4𝑑𝑙𝑛𝑑𝑒𝑝𝑜𝑠𝑖𝑡𝑡−1 + 𝛽5𝑑𝑙𝑛𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛_𝑒𝑥𝑝𝑒𝑐𝑡𝑖𝑜𝑛𝑡−1 +𝛽6𝑑𝑙𝑛𝑖𝑛𝑡𝑒𝑟𝑒𝑠𝑡_𝑟𝑎𝑡𝑒𝑡−1 +𝛽7𝑑𝑙𝑛𝑙𝑖𝑞𝑢𝑖𝑑𝑖𝑡𝑦_𝑎𝑠𝑠𝑒𝑡𝑡−1 +𝛽8𝑑𝑙𝑛𝑛𝑝𝑙𝑡−1 +𝜀𝑡 (2) Based on the above formula (2), in our econometric analysis, a long-run correlation test was conducted using the ARDL method to determine the degree of influence of factors affecting the volume of loans allocated by commercial banks to small businesses. The results of this test are reflected in the following table (Table 3). Table 3: Estimation in the autoregressive distributed lag (ARDL) model 3 2 Prepared by the author using Eviews-12.0 software based on data from the Central Bank of the Republic of Uzbekistan. 3 Prepared by the author using Eviews-12.0 software based on data from the Central Bank of the Republic of Uzbekistan. Statistical Test Value K F-Statistics 3.231052* 8 The Essential Value Of Connection Importance Level I(0) bound I(1) bound 10% 1.85 2.85 5% 2.11 3.15 2,5% 2.33 3.42 1% 2.62 3.77 Variables (ARDL) (1,2,3,3) Coefficient Standard Error F-statistics Plausibility lnloans (-1) 0.192419 0.092990 2.069252 0.0469 lnloans (-2) -0.000101 0.217620 -0.000466 0.9996 lnloans (-3) 0.541615 0.114412 4.733897 0.0000 lndeposit _interest_rate 0.730321 0.225569 3.237677 0.0029** Lnprofit -0.022161 0.092138 -0.240517 0.8115 Econometric Analysis of Financial Support for Business Entities JEFMS, Volume 08 Issue 12 December 2025 www.ijefm.co.in Page 7664 *** indicates statistical significance at 1%, ** indicates statistical significance at 5%, * indicates statistical significance at 10%. lnprofit (-1) 0.085241 0.077593 1.098564 0.2804 lnprofit (-2) -0.108585 0.066002 -1.645176 0.1100 Dlncapital -0.570674 3.762666 -0.151668 0.8804 dlndeposit -2.778726 3.656167 -0.760011 0.4530 dlndeposit (-1) 4.839602 2.634982 1.836673 0.0759 dlninflation_expectation -0.897396 0.452068 -1.985091 0.0560 dlninflation_expectation (-1) -0.193954 0.545811 -0.355351 0.7247 dlninflation_expectation (-2) -0.708221 0.566969 -1.249136 0.2210 dlninflation_expectation (-3) -1.009430 0.394022 -2.561858 0.0155 dlninterest_rate -0.814477 0.785195 -1.037293 0.3076 dlninterest_rate (-1) -0.159228 0.726860 -0.219062 0.8280 dlninterest_rate (-2) 2.421489 1.121983 2.158223 0.0388 dlninterest_rate (-3) 0.745252 1.332372 0.559343 0.5799 dlninterest_rate (-4) 2.708201 1.015043 2.668064 0.0120 dlnliquid_asset -0.000610 1.006230 -0.000606 0.9995 dlnliquid_asset (-1) -1.862672 0.887418 -2.098980 0.0441 dlnliquid_asset (-2) -1.362469 0.503125 -2.708011 0.0109 Dlnnpl 0.838444 0.315943 2.653783 0.0124** С 0.776507 1.864106 0.416557 0.6799 Model Criteria Residue Square 0.864695 Square Of The Modified Residuals 0.815921 F-Statistics 2.671881 Plausibility (F-statistics) 0.000667** Econometric Analysis of Financial Support for Business Entities JEFMS, Volume 08 Issue 12 December 2025 www.ijefm.co.in Page 7665 The data in Table 3 show that, as can be seen from the ARDL coefficients, the volume of commercial bank loans to small businesses is positively affected by the interest rate on commercial bank deposits and the problem loans of commercial banks. However, the increase in the volume of commercial bank loans to small businesses is mainly caused by the increase in the interest rate on commercial bank deposits and the problem loans of commercial banks. With other factors remaining unchanged (seteris paribus), an increase in the interest rate on commercial bank deposits by one percent will lead to an increase in the volume of commercial bank loans to small businesses by 0.73 percent. Also, an increase in the volume of problem loans of commercial banks by one percent will lead to a decrease in the volume of commercial bank loans to small businesses by 0.83 percent. It is clear from this that the increase in the volume of commercial bank loans to small businesses is directly proportional to the increase in the volume of problem loans of commercial banks. Since the program used automatically selects the most optimal one from the distributed lags based on the Schwarz criterion [15], we consider the Schwarz criterion coefficients for the lag combination (Figure 3). .112 .116 .120 .124 .128 .132 .136 .140 .144 .148 Model748890 Model733265 Model748885 Model748265 Model780140 Model748270 Model742640 Model748765 Model733640 Model733140 Model780765 Model779520 Model733260 Model779515 Model777015 Model748889 Model780015 Model749265 Model702640 Model748760 Akaike Information Criteria (top 20 models) Model748890: ARDL(3, 0, 2, 0, 1, 3, 4, 2, 0) Model733265: ARDL(3, 0, 3, 0, 1, 3, 4, 2, 0) Model748885: ARDL(3, 0, 2, 0, 1, 3, 4, 3, 0) Model748265: ARDL(3, 0, 2, 0, 2, 3, 4, 2, 0) Model780140: ARDL(3, 0, 0, 0, 1, 3, 4, 2, 0) Model748270: ARDL(3, 0, 2, 0, 2, 3, 4, 1, 0) Model742640: ARDL(3, 0, 2, 2, 1, 3, 4, 2, 0) Model748765: ARDL(3, 0, 2, 0, 1, 4, 4, 2, 0) Model733640: ARDL(3, 0, 3, 0, 1, 0, 4, 2, 0) Model733140: ARDL(3, 0, 3, 0, 1, 4, 4, 2, 0) Model780765: ARDL(3, 0, 0, 0, 0, 3, 4, 2, 0) Model779520: ARDL(3, 0, 0, 0, 2, 3, 4, 1, 0) Model733260: ARDL(3, 0, 3, 0, 1, 3, 4, 3, 0) Model779515: ARDL(3, 0, 0, 0, 2, 3, 4, 2, 0) Model777015: ARDL(3, 0, 0, 1, 1, 3, 4, 2, 0) Model748889: ARDL(3, 0, 2, 0, 1, 3, 4, 2, 1) Model780015: ARDL(3, 0, 0, 0, 1, 4, 4, 2, 0) Model749265: ARDL(3, 0, 2, 0, 1, 0, 4, 2, 0) Model702640: ARDL(3, 1, 0, 0, 0, 3, 4, 2, 0) Model748760: ARDL(3, 0, 2, 0, 1, 4, 4, 3, 0) Figure 1. Schwarz criterion (top 20 models) 4 4 Prepared by the author using Eviews-12.0 software based on data from the Central Bank of the Republic of Uzbekistan. Econometric Analysis of Financial Support for Business Entities JEFMS, Volume 08 Issue 12 December 2025 www.ijefm.co.in Page 7666 Like the Akaike and Hanna-Quinn criteria, the Schwarz criterion selects the model with the smallest value. As can be seen from Figure 1, the (ARDL) lag distribution with the smallest value, that is, the most optimal model choice, is the lag with the combination distribution (3, 0, 2, 0, 1, 3, 4, 2, 0), and the result in Table 3 above is based on this distribution. In order to achieve a perfect and detailed result in econometric analysis, it is advisable to check the short-term and longterm dependence in the ARDL model. To conduct a short-term dependence test, we introduce the elements of the error correction mechanism into the econometric model and bring it to the following form: 𝑙𝑛𝑙𝑜𝑎𝑛𝑠𝑡= 𝛼0+∑𝛼1 𝑝 𝑡𝑙𝑛𝑑𝑒𝑝𝑜𝑠𝑖𝑡_𝑖𝑛𝑡𝑒𝑟𝑠𝑡_𝑟𝑎𝑡𝑒𝑡−𝑖 +∑𝛼2𝑙𝑛𝑝𝑟𝑜𝑓𝑖𝑡𝑡−𝑖 𝑞 𝑡+∑𝛼3𝑑𝑙𝑛𝑐𝑎𝑝𝑖𝑡𝑎𝑙𝑡−𝑖 𝑞 𝑡+∑𝛼4𝑑𝑙𝑛𝑑𝑒𝑝𝑜𝑠𝑖𝑡𝑡−𝑖 𝑞 𝑡 +∑𝛼5𝑑𝑙𝑛𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛_𝑒𝑥𝑝𝑒𝑐𝑡𝑖𝑜𝑛𝑡−𝑖 𝑞 𝑡+∑𝛼6𝑑𝑙𝑛𝑖𝑛𝑓𝑙𝑎𝑡𝑖𝑜𝑛_𝑒𝑥𝑝𝑒𝑐𝑡𝑖𝑜𝑛𝑡−1 𝑞 𝑡+∑𝛼7𝑑𝑙𝑛𝑙𝑖𝑞𝑢𝑖𝑑𝑖𝑡𝑦_𝑎𝑠𝑠𝑒𝑡𝑡−1 𝑞 𝑡 +∑𝛼8𝑑𝑙𝑛𝑛𝑝𝑙𝑡−1 𝑞 𝑡+𝛾𝐸𝐶𝑀𝑡−1 +𝜀𝑡 Here; 𝛼0 – constant, 𝜀𝑡 – standard error, 𝛼1,…,𝛼6 – short-term coefficient, 𝛽1,….,𝛽5 – long-term elasticity, ECМ – error correction coefficient, γ-rate of change (coefficient of variation). Table 4: Cointegration form and short-run dependence test 5 Variables Coefficient Standard Error F-Statistics Plausibility C 0.776507 1.612508 0.481552 0.6335 lnloans(-1)* -0.266068 0.166923 -1.593957 0.1211 lndeposit _interest_rate** 0.730321 0.353464 2.066180 0.0473 lnprofit(-1) -0.045504 0.088709 -0.512959 0.6116 dlncapital** -0.570674 4.563119 -0.125062 0.9013 dlndeposit (-1) 2.060875 3.761403 0.547901 0.5877 dlninflation_expectation(-1) -2.809001 1.698064 -1.654237 0.1082 dlninterest_rate(-1) 4.901237 3.273745 1.497134 0.1445 dlnliquid_asset(-1) -3.225751 1.398354 -2.306820 0.0279 dlnnpl** 0.838444 0.462327 1.813530 0.0094 д(lnloans(-1)) -0.541513 0.182851 -2.961504 0.0058 д(lnloans(-2)) -0.541615 0.161115 -3.361669 0.0021 д(lnprofit) -0.022161 0.076230 -0.290708 0.7732 д(lnprofit(-1)) 0.108585 0.066072 1.643416 0.1104 д(dlndeposit ) -2.778726 2.773083 -1.002035 0.3241 д(dlninflation_expectation) -0.897396 0.583638 -1.537592 0.1343 д(dlninflation_expectation(-1)) 1.717651 0.907544 1.892636 0.0678 д(dlninflation_expectation(-2)) 1.009430 0.523139 1.929564 0.0629 д(dlninterest_rate) -0.814477 1.245350 -0.654015 0.5179 д(dlninterest_rate(-1)) -5.874942 2.254097 -2.606340 0.0139 д(dlninterest_rate(-2)) -3.453453 1.794196 -1.924791 0.0635 д(dlninterest_rate(-3)) -2.708201 1.341460 -2.018846 0.0522 д(dlnliquid_asset) -0.000610 0.957940 -0.000636 0.9995 д(dlnliquid_asset(-1)) 1.362469 0.518258 2.628938 0.0132 еc = lnloans - (2.7449*lndeposit _interest_rate -0.1710*lnprofit -2.1448*dlncapital + 7.7457*dlndeposit - 10.5575*dlninflation_expectation + 18.4210*dlninterest_rate-12.1238*dlnliquid_asset + 3.1512*dlnnpl - 2.9185) *** indicates statistical significance at 1%, ** indicates statistical significance at 5%, * indicates statistical significance at 10%. 5 Prepared by the author using Eviews-12.0 software based on data from the Central Bank of the Republic of Uzbekistan. Econometric Analysis of Financial Support for Business Entities JEFMS, Volume 08 Issue 12 December 2025 www.ijefm.co.in Page 7667 Based on the data in Table 4, we conducted a cointegration form and short-term correlation test. After the errors are corrected, the growth of commercial bank loans to business entities is positively affected by the independent variables of the interest rate on commercial bank deposits and the increase in commercial bank problem loans. Based on the hypothesis test, the coefficients of the interest rate on commercial bank deposits and the increase in commercial bank problem loans are positively significant with a probability value of 0.0000. The coefficient of the interest rate on commercial bank deposits is 0.0473, indicating a 95% confidence level. In addition, the coefficient of the problem loans of commercial banks is 0.0094, indicating a 99% confidence level. The error correction coefficient CointEq(-1) has a probability value of 0.0000, and the confidence level of the coefficient is 99 percent. The error correction coefficient means that the deviation from the long-term to the short-term is corrected by 99 percent every quarter. The coefficient CointEq(-1) of -2.9185 indicates that the deviation from the long-term to the short-term is corrected by 2.91 percent. The short-term correlation test shows that the growth of commercial bank loans to small business entities is associated with the interest rate on commercial bank deposits and commercial bank non-performing loans. The fact that these variables have a long-term correlation is shown in the table below (Table 5). Table 5: Long-term dependency test 6 Variables Coefficients Standard Error F-statistics Plausibility LNDEPOSIT _INTEREST_RATE 2.744867 2.959234 0.927560 0.3608 LNPROFIT -0.171025 0.521736 -0.327799 0.7453 DLNCAPITAL -2.144845 14.54280 -0.147485 0.8837 DLNDEPOSIT 7.745672 25.01016 0.309701 0.7589 DLNINFLATION_EXPECTATION -10.55746 11.78231 -0.896043 0.3771 DLNINTEREST_RATE 18.42100 15.14592 1.216235 0.2331 DLNLIQUID_ASSET -12.12378 16.27969 -0.744718 0.4621 DLNNPL 3.151241 3.048171 1.033814 0.3092 C 2.918453 4.576047 0.637767 0.5283 *** indicates statistical significance at 1%, ** indicates statistical significance at 5%, * indicates statistical significance at 10%. The data in Table 5 show that the long-term correlation test showed some variation in the impact of the factors. In particular, the increase in the volume of commercial bank loans to small businesses is not affected by the increase in the interest rate on commercial bank deposits and the increase in problem loans of commercial banks, which is not statistically significant. The coefficients of these indicators are statistically insignificant, and we can see that the effect of these coefficients is not present even at the 99% confidence level. As can be seen from the econometric model, the selected variables have only a short-term correlation, and therefore provide the opportunity to forecast in the short term. In order to check the robustness of our econometric model, we perform the CUSUM test of the residuals (Figure 2). -20 -15 -10 -5 0 5 10 15 20 II III IV III III IV III III IV 2022 2023 2024 CUSUM 5% Significance 6 Prepared by the author using Eviews-12.0 software based on data from the Central Bank of the Republic of Uzbekistan. Econometric Analysis of Financial Support for Business Entities JEFMS, Volume 08 Issue 12 December 2025 www.ijefm.co.in Page 7668 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 II III IV III III IV III III IV 2022 2023 2024 CUSUM of Squares 5% Significance Figure 2. CUSUM test of the total sum of residuals 7 As can be seen from the CUSUM test [16], the total sum of the residuals of the variables is stable. In addition, the sum of the squares of the residuals of the variables is also stable. Because the interval of the residuals does not go beyond the interval of statistical significance of five percent. The econometric model formed based on the results of the CUSUM test can only make short-term forecasts. In conclusion, it can be said that the result of the in-depth econometric analysis conducted in the autoregressive distributed lag (ARDL) model was the study of the effect of the interest rate on commercial bank deposits and problem loans of commercial banks, which are arbitrary variables, on the growth of commercial bank loans to small business entities. As a result, with other factors remaining unchanged (seteris paribus), an increase in the interest rate on commercial bank deposits by 1 percent leads to an increase in the volume of commercial bank loans to small business entities by 0.73 percent. Also, an increase in problem loans of commercial banks by 1% will lead to an increase in the volume of commercial bank loans to business entities by 0.83%. As can be seen from our research, it is necessary to pay attention to the resolution of short-term problem loans of commercial banks, taking into account the short-term importance of short-term deposit interest rates and liquidity assets, and we believe that monetary policy and bank liquidity management strategies should be closely linked with credit policy in order to effectively influence the volume of commercial bank loans to business entities in the economy. Based on the developed econometric model, we make a forecast of the volume of commercial bank loans to business entities until 2030. Table 6: Forecast of the growth trend of the volume of commercial bank loans to business entities, billions of soums [17] Years The Volume Of Loans Allocated By Commercial Banks To Small Businesses Volume Of NonPerforming Loans (NPL) Average Interest Rates On Bank Deposits 2024 92 424,8 21 185,00 19,60 2025 93 866,63 22 943,36 21,03 2026 95 330,95 24 847,65 22,57 2027 96 818,11 26 910,01 24,21 2028 98 328,47 29 143,54 25,98 2029 99 862,40 31 562,45 27,88 2030 101 420,25 34 182,14 29,91 As can be seen from the data in Table 6, the volume of loans granted by commercial banks to small businesses is on an upward trend and is expected to reach 101,420.25 billion soums in 2030. We can see that this figure is expected to increase by 9.7% compared to 2024. As a result of the increase in the volume of loans granted by commercial banks to small businesses, the 7 Prepared by the author using Eviews-12.0 software based on data from the Central Bank of the Republic of Uzbekistan. Econometric Analysis of Financial Support for Business Entities JEFMS, Volume 08 Issue 12 December 2025 www.ijefm.co.in Page 7669 volume of non-performing loans (NPL) in commercial banks is expected to increase by 1.61 times and the average interest rates on commercial bank deposits are expected to increase by 52%. Therefore, the volume of loans granted by commercial banks to small businesses will lead to an increase in the volume of non-performing loans (NPL) in commercial banks and the average interest rates on commercial bank deposits. CONCLUSION The scientific innovation is based on the proposal to forecast the growth of loans granted by commercial banks to small businesses until 2030 based on the autoregressive distributed lag (ARDL) econometric model. As a result of the practical implementation of this scientific result, the long-term and short-term impact of the average interest rate on commercial bank deposits and problem loans on the growth of loans granted by commercial banks to small businesses was assessed. As a result, with other factors remaining unchanged (seteris paribus), an increase in the interest rate on commercial bank deposits by 1 percent leads to an increase in the volume of commercial bank loans to small businesses by 0.73 percent. It is also justified that the increase in problem loans of commercial banks by 1 percent and the increase in the volume of commercial bank loans to business entities by 0.83 percent. REFERENCES 1) Cho, J. S., Greenwood‐Nimmo, M., & Shin, Y. (2021). Recent developments of the autoregressive distributed lag modelling framework. Journal of Economic Surveys, 37(1), 7. https://doi.org/10.1111/joes.12450 2) Cho, J. S., Greenwood‐Nimmo, M., & Shin, Y. (2021). Recent developments of the autoregressive distributed lag modelling framework. Journal of Economic Surveys, 37(1), 7. https://doi.org/10.1111/joes.12450 3) Sofro, A., Aidha, S. D. N., & Khikmah, K. N. (2023). Forecasting poverty in East Java using vector autoregressive method and vector error correction model. E3S Web of Conferences, 450, 3004. https://doi.org/10.1051/e3sconf/202345003004 4) Sajid, M. J., Khan, S. A. R., Sun, Y., & Yu, Z. (2022). The long-term dynamic relationship between communicable disease spread, economic prosperity, greenhouse gas emissions, and government health expenditures: preparing for COVID-19like pandemics. Environmental Science and Pollution Research, 30(10), 26164. https://doi.org/10.1007/s11356-02223984-9 5) Chen, P. (2024). Vector Error Correction Models with Stationary and Nonstationary Variables. Economic Analysis Letters, 3(2), 37. https://doi.org/10.58567/eal03020004 6) Natsiopoulos, K., & Tzeremes, N. (2022). ARDL: An R package for the analysis of level relationships. The Journal of Open Source Software, 7(79), 3496. https://doi.org/10.21105/joss.03496 7) Khalid, W. (2023b). The Impact of Renewable and Non-Renewable Energy Consumption on Aggregate Output in Pakistan: Robust Evidence from the RALS Cointegration Test. Research Square (Research Square). https://doi.org/10.21203/rs.3.rs3715285/v2 8) Caetano, R. V., & Marques, A. C. (2023). Nonlinear relationships between Foreign Direct Investment decisions and environmental degradation in highand middle-income countries. Journal of Industrial and Business Economics, 51(1), 135. https://doi.org/10.1007/s40812-023-00282-0 9) Awosusi, A. A., Eweade, B. S., & Ojekemi, O. S. (2024). Analyzing the environmental role of resource efficiency, economic globalization, and biomass usage in Malaysia: a time-varying causal approach. Environment Development and Sustainability. https://doi.org/10.1007/s10668-024-05196-y 10) Erdoğan, S., Pata, U. K., Okumuş, İ., & Solarin, S. A. (2024). Uncovering the linkage among tourism diversification, economic development, energy consumption, population density, technology, and carbon emissions: empirical evidence from New Zealand. Air Quality Atmosphere & Health. https://doi.org/10.1007/s11869-024-01594-3 11) Thampanya, N., Wu, J., & Cowton, C. J. (2021). Carbon neutrality targets, optimal environmental management strategies & the role of financial development: New evidence incorporating non-linear effects and different income levels. Journal of Environmental Management, 297, 113352. https://doi.org/10.1016/j.jenvman.2021.113352 12) https://www.researchgate.net/publication/285050843_Vector_Autoregressive_and_Vector_Error_Correction_Models 13) Dickey, D. A.; Fuller, W. A. (1979). "Distribution of the Estimators for Autoregressive Time Series with a Unit Root". Journal of the American Statistical Association. 74 (366): 427–431. 14) Phillips,P.C.B.; Perron,P.(1988)."Testing for a Unit Root in Time Series Regression". Biometrika. 75 (2): 335–346.