“Determinants of the credit risk in developing countries: a case of Kosovo banking sector”

The determinant of the credit risk of banks in a developing country have limited data to analyze and limited participation in literature. Determinants of credit risk are very important in order to define the non-performing loans (NPL) in Kosovo banking systems. Even though banking system in Kosovo is the newest in region, it is comparable with banking systems to all places in regions (Albania, Serbia, Montenegro, Macedonia, Bosnia and Herzegovina, etc.). The main purpose of this paper is to classify some factors that influence credit risk in commercial banks in Kosovo. The research includes seven commercial banks for the period 2006–2015. Data analysis and interpretation are processed with Statistical Program for Social Sciences SPSS v.19.0. The effect of variations in the determinants of credit risk exposure is based on using a multivariate panel regression model. Our empirical results suggest that a significant relationship exists between credit risk and the following variables: Profitability (ROE and ROA), Inefficiency (IE), Loans to deposit ratio (LDR), Credit growth (CG) and Deposit rate (DR), while variables of Solvency (SR) and Credit rate (CR) are not statistically significant in terms of credit risk.


INTRODUCTION
Financial markets play a key role in facilitating risk sharing and efficient allocation of assets among investors (Chabakauri & Han, 2016).Financial sector in Kosovo is under supervision of Kosovo Central Bank (KCB), which supervises financial system in compliance to current law competences.Actually there are ten banks that operate in Kosovo where eight of them operate with foreign capital including 93.1% of overall banking sector capital and two other banks operate with native capital including 6.9% of overall banking sector capital (KCB, 2015).Credit giving activity from commercial banks continues to represent one of the main sources of commercial banks (Shkodra et al., 2012).Credits play a very important role for countries in transition such as Kosovo in creating, increasing and developing business activities (Shkodra et al., 2011).
Financial sector is categorized as a very important pillar for economic stability and in Kosovo this sector compared to other places in the region is very stable (Shkodra et al., 2012).The last financial crisis and recession have made non-performing loans one of the major concerns for both bank managers and regulatory authorities.The recent financial crisis has called the attention to the consequences that banking crises can have on the economy (Agnello & Sousa, 2011).For this reason, some economists look again at the factors that may trigger a banking crisis (De Grauwe, 2010; Laeven & Valencia, 2010).Bank credit risk in emerging markets showing causal relationship between economic growth, real interest rate, the net interest margins, the real exchange rate appreciation, and non-performing loans is analyzed by Fofack (2005).
The risk is a process that needs to be identified and managed with a continuous and dynamic precisely accuracy in every financial institution while including all risks spectrum such as: a) credit risk, b) market risk, c) liquidity risk, d) operational risk, and e) transferred risk 1 .Credit risk management is one of the most serious challenges of financial sector which is influenced mostly by both internal factor and external factors which should not be neglected.KCB which supervises financial institutions in Kosovo uses risks matrix (see Table 1) in order to collect the levels of risks of a particular financial institution activity (KCB, Financial Supervisor Manual, 2015).
In order to assess and to identify identical and comprehensive risks in all banks and then to apply in compliance with current laws and rules the KCB has determined assessment systems CAMELS and CAELS as a general frame for monitoring risks.
Assessment system CAMELS involves: assessment of capital sufficiency, credit risk and financial means quality, management, gains, liquidity risk, sensitivity from trade risk, operational risk, internal checks, internal auditing, compliance with KCB recommendations and focused examination during month.Assessment system CAELS involves: general risk, credit risk, liquidity risk, sensitivity from trade risk and operational risk.
Based on the fact that debts include the major group of banking activities and are major source of gain for each bank as well, we consider that credit risk has an important influence and requires a special concentration.Credit risk management is a systematic issue which requires special attention from each institution that deals with credit giving.
Based on probability categorization of credit risk -a) low; b) medium; and c) high -in Kosovo credit risk is assessed to be high, whereas according to the way categorization of credit risk -a) decreasing; b) stable; and c) increasing -credit risk is assessed to be increasing.
Because banking industry is considered as a sensitive industry taking into account the fact that it faces high and different risks KCB should take a special consideration when creating rules and laws to keep financial system stable which may influence the economy of Kosovo.According to KCB (KCB, Clients debt in banks in Kosovo, 2015) overload of debts in banks apart from direct risk to financial sector stability involves also social and psychological influence of debtors and society in general.
KCB has undertaken a series of measures to keep financial stability of the country and to protect from unperformed credits of commercial banks such as: a) obligated reserves measure -where every bank is obligated to deposit means in KCB and to use them when credit converge became impossible; b) credit provision measure -where every bank is obligated to pay provision to KCB for each credit issued; c) liquidity measure -where every bank is obligated to keep the level of liquidity determined by KCB in order to cover obligation toward client in every time.
Weak credit risk management will cause the increase in unperformed credits which can lead to financial instability or, according to Cooper et al. (2003), variations in credit risks would lead to variations in the health of banks' loan portfolio which in turn affects bank performance.Gonzalez-Hermosillo et al.
(1997) define the higher share of NPLs to total loans leads to the greater probability of banking failure.The level of NPL and provisions in commercial banks in Kosovo are represented in Figure 1.
Figure 1 shows that NPL to total credit ratio has been increased each year from 2008 to 2013; in 2008 NPL included 3.3% of overall credits in banking system in Kosovo, whereas in 2015 NPL reached 6.2%, so the NPL in 2014 started decreasing trend.From Table 2 we see that Kosovo has a very good performance in comparison to the countries in the region.

ANALYSIS OF CREDIT RISK DETERMINANTS
The data are taken from publication reports, but in some cases we need to calculate average data because the data hasn't been checked.Now we will see calculations done for assessment of selected variables which will not significantly impact credit risk and those which have impact on credit risk.

RESULT INTERPRETATION
Table 3 presents correlation analysis for all variables including dependent and independent variables.To determine parametric or nonparametric test the Kolmogorov -Smirnov goodness of fit test is used.
The coefficients of Pearson correlation in Table 3 show a positive correlation what is important, but we find not significant correlation between LDR (Loans deposit rate) and CR (credit risk), between SR (solvency) and CR (credit risk), as well as between the DR (deposit rate) and the CR (credit risk).The opposite of that is a negative correlation, which we show at the Pearson correlation coefficients relation between cr (credit rate) and CR (credit risk).
Table 4 showed also the problem of multicollinearity by using the VIF (variance inflation factor).There is no serious problem with the use of our model of the multicollinearity, because all VIF values are identified less than 2 (the highest VIF has value of 1.869).
This holds that the presence of multicollinearity is minimal.In order to examine the heteroscedasticity problem in the model, Durbin -Watson (DW) test was used.While examining the heteroscedasticity in Table 5, one may note that the observed positive autocorrelation is 1.851 for credit risk.Endogeneity in the model was tested by Hausman test.The Hausman test based on Chi-squared statistic (9.77, and prob.0.5994) indicates that corresponding effects are statistically insignificant, so the null hypothesis is accepted and random effect model is preferred.
The Hausman test simultaneously examines the justification for the use of instrumental variables as possible solutions to the endogeneity problem.
In terms of the significance of all the independent variables taken together the F-test was used with significance of 5%.It implies that our model is a good fit because in Table 4 it can be seen the F statistic (6.811) of the probability for the overall regression relationship is 0.005 < . It can be concluded that we reject the null hypothesis that all coefficients are simultaneously zero and accept that the overall regression is significant.Furthermore, individual t-tests show that four variables in our Model (IE, ROE, CG and DR) are found to be statistically significant at the empirical significance level of less than 5%.Linear combination of explanatory variables formed the regression function in our model that provides R-squared coefficient of determination of 31.27%.The regression results are given in Table 4.The R-Squared statistic indicates that all these 8 predictor variables combined explain 31.27% of the variance in credit risk.The remaining nearly 68.73% of the variations in Credit risk can be explained by factors that are not included in our model.The standard error of the estimate shows the standard deviation of the residuals to be .03454.The adjusted R-squared statistic, which is more suitable for comparing models with different numbers of independent variables, is 21.29%.When assessing the impact of independent variables on credit risk variable ROE has the most influence whose regression standardized coefficient by a variable ROE (beta is 0.41), followed by a variable CG (beta is -0.111) and variable DR (0.147).As it can be observed from the summary of regression output, all other regression coefficients were not statistically significant because the p-value is larger than 0.05.Besides, the regression results also revealed that only four variables have significant impact on the model.The characteristic variables in our model are ROA (-0.11410),IE (-0.00421),CG (-0.01203) and cr (-0.00031) which have reciprocal relationship with the CR.It means that when each of these variables increases, it leads to lower CR.On the other side of the coin, positive relationships with CR have the following variables: ROE (0.11422), LDR (0.00212), DR (0.03132) and SR (0.03919) which means that when each of these variables decreases, it leads to raise CR.

CONCLUSIONS
Following on from the results we found out that a significant relationship exists between credit risk and the following variables: Profitability (ROE and ROA); Inefficiency (IE); Loans to deposit ratio (LDR); Credit growth (CG) and Deposit rate (DR), while variables Solvency (SR) and Credit rate (cr) are not statistically significant in terms of credit risk (CR).
The findings of this study showed that banking credit risk is significantly negatively affected by IE and CG.It means that with the growth of credit risk the banks' cost efficiency and lending of the banks decline.Moreover, the negative statistically significant value of IE and CG suggests that the both variables have a substantial impact on credit risk.Further, a negative coefficient of IE (0.00421) implies that increase in efficiency leads to decrease in credit risk as well as with CG (0.01203).
Coefficient of 0.03132, which has DR shows that any increase in this variable leads to increase in CR.The value of 0.11422 is a positive coefficient of ROE which means that increase in ROE leads to CR.Moreover, the results of the study demonstrate that ROA, LDR and cr coefficients estimate is positive, however, statistically not significant.The low coefficient of LDR 0.00212 suggests that LDR has weak impact on the CR.
The relationship between the and CR is positive and also significant.Lastly, ROA and SR also don't have any significant relationship with CR.In conformity of econometric methods in this research we can confirm the hypothesis in the case of Kosovo that the highest importance in explaining the variability of credit risk in banks is inherent to the following variables: Profitability (ROE), Inefficiency (IE), Credit growth (CG) and Deposit rate (DR) while variables Profitability (ROA), Loans to deposit ratio (LDR), Solvency (SR) and credit rate (cr) are not statistically significant in terms of credit risk.

Figure 1 .
Figure 1.NPL and provisions for the period of 2008-2015

Table 1 .
Banking risk matrix

Table 3 .
Correlation matrix of bank specific variables

Table 3 (
cont.).Correlation matrix of bank specific variables Note: * Correlation is significant at the 0.05 level.** Correlation is significant at the 0.01 level.

Table 4 .
Common effect model

Table 5 .
Model summary