Determinants of loans in Slovakia

Transcription

1 Determinants of loans in Slovakia Ing. Kristína Kočišová, PhD. Technical University of Košice, Faculty of Economics, Department of Banking and Investment; Nemcovej 32, Košice exclusive e-journal Abstract The aim of this paper is to analyse the determinants of loans in the Slovak banking sector. Analysis of selected individual characteristics of banking sector is realized in period of on the basis of quarterly data published by National Bank of Slovakia. Data on profitability; operating efficiency; quality of bank portfolio; share of the interest income; basic interest rate of the National bank of Slovakia (NBS) and concentration of banking sector will be used as independent variables for construction of econometric model. Loans, accounted dependent variable, will be measured as total loans to clients in log form. The results of regression analysis will show variables with positive and negative significant impact to loans. Key words: Banking sector, Total loans to customers, Profitability, Operating efficiency, Nonperforming loans, Concentration 1. INTRODUCTION A majority of studies relating to the issuing of credits are concerned with monetary financial institutions (banks) and their relations towards households and corporates, because banks represent a crucial source of loans for many families and firms. Structure of banking activities and the role of banks has changed in recent years. One of the factors that influenced the development on the banking sector was also financial crisis and its consequences. However, bank loans' demand has increased during the last decade in many countries, also in case of the Slovak Republic. In the Slovak Republic there is the lack of studies examining influence of banking sector characteristics on total volume of loans. Trends on the Slovak credit market are described for example in the study of Urbanovičová (2011). In her study she analysed the impact of financial crisis on the Slovak banking sector. Through the regression model she analysed how total deposits, rate of inflation, profit and capital affected the total loans in the period of Vokorokosová a Peller (2013a, 20013b) examined the impact of unemployment rate, average interest rate for credits to households, disposable income of households, earnings from property of households to household credits in Slovakia. They used quarterly data from 1995 to 2004 for modelling the relation between evaluated variables and they found out, that credits to households were positively influenced by disposable income and negatively by unemployment rate, real interest rate and earnings from property. Kiseľáková (2010) in her paper focused on analysis of credit risk in Slovakia, dependence of selected risk factors and quantification of impacts through models of regression analysis during the period As the depended variables she used total loans or profit and as the independent variables were

2 used for example market interest rate, total deposits, share of non-performing loans to total loans and capital adequacy. In her study she found out, that total deposits positively influenced total loans. This paper aims to analyse the impact of profitability; operating efficiency; quality of bank portfolio; share of the interest income; basic interest rate of central bank and concentration of banking sector on bank lending. Time-series analysis is realized in period on the basis of quarterly data of Slovak banking sector, published by the National Bank of Slovakia. The next part presents the data and describes the variables used. Relationships between selected individual characteristics of banking sector and loans are described in section 3. The paper ends with summary and conclusion. 2. DATA DESCRIPTION We will use quarterly data from Q4/2005 to Q4/2013 for the Slovak banking sector, obtained from the National bank of Slovakia. Data on total loans to clients; profitability; operating efficiency; quality of bank portfolio; share of the interest income; basic interest rate of NBS and concentration of banking sector will be used as independent variables in construction of econometric model. The dependent variable in the model will be total loans to clients and will be measured as total loans to clients in log form. The development of total loans to clients is presented in Figure 1. As can be seen in the figure, volume of total loans to clients increased throughout the analysed period, with faster growth rate at beginning of period. In 2009 loan growth stagnated, which may be consequence of the global economic crisis, which in this period fully reflected in the Slovak banking sector. Loan growth was accompanied by a rise of share of loans to total assets of banking sector. While at the beginning of the analysed period (Q4/2005) loans accounted for about 38% of total assets at the end of the analysed period (Q4/2013) loans accounted for about 63% of total assets of banking sector. Figure 1: Loans to clients Source: Author s calculations on the basis of [2] Structure of loans in term of target segment is shown in Figure 2. It can be seen that over the analysed period the majority of total loans (about 80%) were located in the segment of retail and corporates. Other segments (non-bank financial companies, general government, non-residents) had a minor proportion of total loans. At the beginning of analysed period loans to corporates accounted for nearly 50 % of total

3 assets. Over the analysed period this share declined. The year 2009 can be considered as a turning point for the banking sector, when the economic crisis hit Slovakia with full force. Banks restricted their investment activities in sectors with a worsening financial position, while they invested more in conservative assets. Financial crisis affected bank lending to the corporate sector. The proof was starting decline in lending in corporate sector, which with slight fluctuations persist until the end of analysed period. The development in the retail segment was not so much affected by financial crisis. Lending to households was affected by the crisis with a certain lag. Therefore we can talk about relatively strong growth in 2009.The growth in retail segment and decline in corporate sector were so important that they changed shares of market segments. Since 2010 the volume of retail loans was higher than the volume of loans to corporates. This trend continued until the end of analysed period, when retail loans accounted for 52% and loans to corporates approximately 37% of total loans. Figure 2: Loans to customer according segments Source: Author s calculations on the basis of [2] Indicators Return on Assets (ROA) will measure profitability of banking sector, as the first independent variable, which is determined as total profit after tax to total assets. This indicators is used to measure how profitable are the companies in use of their assets. Higher value of indicator indicates higher rate of banks profitability. Internationally accepted good value in banking sector is generally considered a value ROA equal to one. In case of the Slovak banking sector the value of ROA was in the range between 0,21% and 1,27%. The maximum value was reached in Q4/2006 so profitability in this period can be marked as very good. At the end of 2006 the profit of banking sector increase of almost 27,7% year-onyear. The profitability of banking sector continued to show the positive trends seen in previous years. The minimum value was reached in Q1/2009 and in this period we can talk about weak profitability. At the beginning of 2009 the profitability of banking sector as a whole fell by more than 50% and this negative trend affected almost all banks in the sector. The negative trend in bank profitability in 2009 was mainly influenced by two factors: the global economic crisis and the euro adoption. The second independent variable will be indicator of operating efficiency, which will be measured by socalled Cost to Income Ratio (C/I). Cost to income ratio represents share of operating costs to operating income. This indicator tells us what percentage of the operating income the banks use for its operation. Decreasing value of this indicator suggests that banking sector uses its resources rationally and effectively. The maximum value of cost to income ratio was reached at the beginning of observed period (Q4/2005). At the end of 2005 the operating efficiency of banking sector deteriorated year-on-year which was influenced by the increasing operating costs of banking sector as a whole by 6%. The

4 minimum value was reached in Q2/2011. This minimum value indicated that in this period banking sector used its resources most rationally and effectively. The third independent variable, quality of bank portfolio, will be measured as the ratio of nonperforming loans to total loans. Non-performing loans represent a total volume of claims, which are at risk of payment and represent potential threat of a bank. Therefore these loans require the creation of provisions, which a bank must use to cover the depreciation of its loan portfolio; the consequence is also decrease of bank s profit. The minimum value of non-performing loans to total loans (NPL/TL) was reached in Q4/2007 that indicates that the quality of loan portfolio was the highest and the credit risk was the lowest. The maximum value in Q3/2010 indicates the highest credit risk and the lowest quality of loan portfolio, which was affected especially by poor quality of loan portfolio in corporate segment. The fourth independent variable will be share of the interest income to total gross income from banking activities (II/TI). This variable points to the importance of interest income. As can be seen the interest income accounted for more than 50% of gross income. The lowest values were reached at the beginning, while the maximum value was reached at the end of analysed period. This fact suggests that in Slovak banking sector there is interest rate policy applicable and its importance in the observed period increased. The fifth independent variable will be average basic interest rate (IR) of the National bank of Slovakia in evaluated period. Interest rate was till 2009 determined as the basic interest rate of NBS for two-week REPO tender and after our entry into Eurosystem was determined as interest rate for main refinancing operations of Eurosystem. The last independent variable will be concentration of banking sector on loan market measured by concentration ratio for 3 largest banks in the loan market (CR3). The CR3 index is measured as the sum of market shares for 3 largest banks and is used to indirectly measurement of the level of competition. Increasing value of CR3 index indicates that the level of competition in banking sector decreases and the market power is concentrated in hand of low number of banks. The elementary descriptive statistics and expected relationship to loans to clients are presented in Table 1; changes between first and last observed period are presented in Figure 3. Table 1 Descriptive statistics (log)tl ROA C/I NPL/TL II/TI IR CR3 MEAN 7, , , , , , , MEDIAN 7, , , , , , , MAX 7, , , , , , , MIN 7, , , , , , , STDEV 0, , , , , , , Observations Expected relationship Source: Author s calculations on the basis of [2]; [3]; [4] Between profitability measured by ROA and lending we expect a positive relationship (Table 1). According Řepková (2010), low profitability of banks leads to their lower stability and reduces their competitiveness and limits the supply of products on the financial market. This fact can be seen in Figure 3. Between beginning and end of analysed period the level of profitability decreased, this can led to the lower level of competition which is reflected in the increasing concentration of the loan market

5 measured by CR3 index. Strong position of dominant banks in the market may be some signal of stability for clients; therefore we expect a positive relationship between CR3 index and loans to clients. Figure 3: Independent variables (Q4/2005; Q4/2013) Source: Author s calculations on the basis of [2]; [3]; [4] Decreasing value of Cost to income ratio (C/I) suggests that banking sector uses its resources rationally and effectively (Figure 3). This fact should be reflected in growth of loans to clients; therefore we expect a negative relationship. The ratio of non-performing loans to total loans (NPL/TL) represents a share of total volume of claims, which are at risk of payment and represent potential threat of a bank what can result in decrease of bank s profit. Growth value of non-performing loans increases the value of NPL/TL indicator and hence the probability of loss. Therefore we can say that the increasing value of NPL/TL has a negative impact on growth of total loans, therefore we expect negative relationship. The share of interest income to total gross income is likely to have a positive influence on the total loans, whereas the main source of banks interest income is mainly banks credit facilities. The last variable is interest rate of the central bank. We expect negative relationship between interest rate and total loans. The reason is that the rate of the central bank influences the level of interest rate of commercial banks. So if the interest rate of the central bank decreased, we could also observe a decrease in interest rate on the loan market. This decrease of interest rates represents a decrease of interest expenses for clients; therefore we can expect growth in demand for loans. 3. MODEL SPECIFICATION AND EMPIRICAL RESULTS The application part of the study examines, whether individual characteristics of the Slovak banking sector had an impact on total loans to clients. To determinate the impact of each variable to total loans we used the regression analysis. The basic regression model in following form was designed: log( TL) t 0 1 ROAt 2 ( C / I) t 3 ( NPL/ TL) t 4 ( II / TI) t 5 IRt 6 CR3t Where log(tl) t are total loans to clients in log form, ROA t is the profitability of banking sector measured by return on assets, (C/I) t is the operating efficiency of banking sector measured by cost to income ratio, (NPL/TL) t represents the quality of loan portfolio by the share of non-performing loans to total loans, (II/IT) t is share of the interest income to total gross income from banking activities, IR t is average basic (1) t

6 interest rate of NBS in evaluated period and CR3 t is concentration ratio for 3 largest banks on the loan market. For examination of the proposed model was used software R. R is a free software environment for statistical computing and graphics. It compiles and runs on a wide variety of UNIX platforms, Windows and MacOS and can be downloaded on the web page: R is very much a vehicle for newly developing methods of interactive data analysis. It is developing fast, and has been extended by a large collection of packages. However, most programs written in R are essentially ephemeral, written for a single piece of data analysis. ( The software R is used for testing the variables and the model as a whole; and verifying if proposed model was specified correctly. Individual tests are in detail described e.g. in [10]. When tested proposed model (1) through Durbin-Watson test the autocorrelation was found. To resolve the problem of autocorrelation there was used method of first differences. Testing model with variables in first differences indicates that the model is statistically significant (F-statistics), residuals are normally distributed (Jarque-Bera Normality Test), in model was not detected heteroskedasticity (Breusch-Pagan test) or autocorrelation (Durbin-Watson test) and the model is specified correctly (Ramsay Reset Test). The results of models testing are in Table 2 (Basic model; Model with first differences). Table 2 Results of models testing Basic model Model with first differences Model 1 Model 2 Model 3 Model 4 Intercept 6,90172*** 0,007769*** 0,007801*** 0,007514*** (35,199) (5,292) (5,441) (5,433) ROA 0, , (0,817) (0,218) C/I -1,02335*** -0,244475* -0,238194* -0,278429** (-8,259) (-2,166) (-2,224) (-2,934) NPL/TL -3,85274*** -1,678620*** -1,669845*** -1,859111*** (-6,027) (-3,747) (-3,813) (-5,008) II/TI 0,65422*** 0, , ,152058* (7,743) (1,525) (1,556) (1,945) IR -1,66653** 0, , (-2,843) (0,692) (0,827) CR3 1,66268*** 0,979845*** 0,992887*** 0,949197*** (5,884) (3,784) (4,015) (3,952) R 2 0,9899 0,6218 0,621 0,6111 F-statistics (p-value) 425,7 (2,2E-16) 6,849 (0, ) 8,522 (7,021E-5) 10,6 (2,689E-5) JB Test (p-value) 0,91 (0,524) 1,411 (0,315) 1,532 (0,278) 1,699 (0,233) BP test (p-value) 10,9633 (0,08952) 2,2631 (0,894) 1,0705 (0,9567) 0,6205 (0,9608) DW test (p-value) 1,195 (0, ) 1,5424 (0,06736) 1,5452 (0,0649) 1,5478 (0,0663) Reset test (p-value) 1,0327 (0,3713) 2,3145 (0,1214) 2,554 (0,09872) 1,3926 (0,267) *** 0,01 ** 0,05 * 0,1; t value in parentheses in the independent variables Source: Author s calculations in R The results in Table 2 shows that in model with first differences (Model 2) the profitability; share of interest income to total gross income; basic interest rate of NBS and concentration of the loan market had positive impact to loans. The operating efficiency and quality of loans portfolio had negative impact to loans. According the p-values in the independent variables was confirmed significance only in case of operating efficiency, quality of loans portfolio and concentration. Other variables were not marked as significant therefore we decided to exclude them gradually from the model. The variable with the highest p-value was profitability, therefore we suggested Model 3 without variable ROA. Model 3 confirmed

7 negative significant impact of operating efficiency, quality of loans portfolio and positive significant impact of concentration. In case of share of interest income to total gross income and basic interest rate of NBS the impact was not significant. Therefore there was tested Model 4 which discards variable with second highest p-value in Model 2 (IR). Model 4 marked all evaluated variables as significant. Table 2 presents results of the regression analysis and Model 4 testing, where we can see that the concentration on the loan market measured by CR3 index and share of interest income had positive impact to total loans. The operating efficiency and quality of loans portfolio had negative impact to total loans. If the cost to income ratio and share of non-performing loans to total loans were decreasing, total loans were rising. If the concentration ratio and share of interest income were rising, total loans were rising too. These finding are in line with expectations expressed in Table 1. Conclusion The aim of this paper was to analyse the determinants of loans in the Slovak banking sector in period of We examined impact of six individual characteristics of banking sector on total loans to clients on the basis of quarterly data published by the National Bank of Slovakia. Selected individual characteristics of banking sector included profitability measured by ROA; operating efficiency measured by cost to income ratio; quality of bank portfolio measured as a share of non-performing loans on total loans; share of the interest income on total gross income; basic interest rate of NBS and concentration of banking sector measured by CR3 index. These variables were used for construction of basic econometric model where dependent variable was total loans to clients in log form. The basic model was tested and as the result of testing was suggested Model 4 with variables in first differences, which excluded from basic model two variables without significant impact. The results of Model 4 regression analysis showed that the concentration on the credit market and share of the interest income had positive significant impact to loans and the operating efficiency and quality of portfolio had negative significant impact to total loans to clients. References 1. KISEĽÁKOVÁ, D.: Manažment kreditného rizika bank v SR a riziká v kontexte globálnej krízy. In: Ekonomická revue. 2010, 13, p ISSN: NBS: Analytické údaje finančného sektora. [online]. [cit ]. Dostupné na internete: < 3. NBS: Úrokové sadzby ECB. [online]. [cit ]. Dostupné na internete: < 4. NBS: Základné úrokové sadzby NBS do [online]. [cit ]. Dostupné na internete: < 5. ŘEPKOVÁ, I.: Structural Determinants of the Total Loans Volume in the Czech Republic. In: European Financial and Accounting Journal. 2010, 3-4, p ISSN: The R Project for Statistical Computing [online]. [cit ]. < 7. URBANOVIČOVÁ, M.: Vplyv hospodárskej krízy na slovenský bankový sektor. In: Mladí vedci 2011: zborník zo 4. medzinárodnej doktorandskej konferencie: Herľany, október, Košice: TU, 2011, p ISBN VOKOROKOSOVÁ, R., PELLER, F.: Credit to Households. What Impacts the Growth in Slovakia? In: Ekonomický časopis. 2013, 3, p ISSN:

8 9. VOKOROKOSOVÁ, R., PELLER, F.: Determinanty úverov domácností v SR. In: Ekonomika a informatika. 2013, 1, p ISSN: ŽELINSKÝ, T., GAZDA, V., VÝROST, T.: Ekonometria Košice: TU. ISBN: