Comparative Review of the Serbia and Czech Republic

Macroeconomic stress test-

Comparative review of the Serbia and Czech Republic

ViktorijaMisic[1]

Abstract:

After the collapse of investment bank Lehman Brothers back in 2008 and the US subrime crisis, the main question became the quantification of financial sector vulnerabilities.Since the collapse of commercial banks can lead to financial instability, one of adequate measures would be to examine the effect of shocks to various risk factors on the financial condition of banks. Credit risk, being a crucial risk in banks, is the risk that a borrower will default on his financial obligations. It can lead to big losses in banking books.

Within the framework of macroeconomic stress test we investigate two countries Czech Republic and Republic of Serbia. The Czech Republic became the member of European Union in 2004 and published first Financial Stability report in 2005.The Czech financial system proved resilient to the effects of the global financial crisis. During the last three years, banks further strengthened capitalization levels, with total capitalization increasing to 15.9 percent by June 2011. Hence, the Czech banking sector was one of the few in Central and Eastern Europe (CEE) which, so far, did not require public support (International Monetary Fund, 2012). AlthoughSerbia became a candidate for European Union membership in 2012, starting from 1. July 2010, the National Bank of Serbia made affords in strengthening stability of the financial system.As a part of the Financial Sector Assessment program (FSAP) stress test were conducted to assess the resilience of the Serbian banking sector to a set of extreme but plausible shocks.

This paper describes the methodology of credit stress test, implementation and practically apply macroeconomic stress test as well as the results of Serbianand Czech banking sector. Our focus is on calculation of default rate for corporate and household sector under the scenario 1, scenario 2 and scenario 3 in case of Czech Republic and Republic of Serbia.

Key words: bank, financial stability, stress tests, credit risk, macro stress test, non-performing loans.

JEL classification: C22, E24, E31, E27, G21, G28.

1. Introduction

The stress tests are carried out by EBA (European Banking Authority) and the national supervisory authorities in EU member states, in close cooperation with the European Systemic Risk Board, the European Central Bank and the European commission. It has become a standard element of Financial Sector Assessment Programs (FSAPs), implemented jointly by the IMF and the World Bank (IMF (2003)). The ESRB was established at 2011 with aim to identify systemic risks and issue recommendations to mitigate those risks.

Stress testing is also important part of the New Basel Accord (BIS, 2004). Basel II (Pillar 1-minimum capital requirements and Pillar 2-supervisory review process) requires banks to perform stress tests. The bank which applies internal rating based approach has to target those requirements. Basel III proposals are designed to lead to greater strength of commercial banks. The capital requirements are getting raised and new regulatory requirements regarding bank liquidity and leverage are introduced, as well as additional capital buffers. (BCBS 2010) [2].

Many empirical studies have employed these macro credit models. Boss (2002) uses the macroeconomic credit model to analyze the stress situation for bank default probability in Austria and finds that industrial production, inflation rate, stock index, nominal short-term interest rates, and oil prices are the determinant factors of default probability.

BaboucekJancar (2005) employ the vector autoregression model (VAR) using the NPLs and the macroeconomic factors for the Czech Republic. Through macro prudential analysis they use an unrestricted VAR model to empirically investigate transmission involving a set of macroeconomic variables who characterized Czech economy.

Drehmann, Sorensen & Stringa (2010) estimate the integrated impact of the credit and the interest rate risks on the banks' portfolios, assessing the banks' economic value, the future earnings and the capital adequacy.

PeuraJokivuolle (2003) measure the capital adequacy by simulating the difference between the bank's actual capital and the minimum capital requirements and they determine whether the estimated bank's capital buffer is sufficient over the business cycles.

Pain (2003) found an empirical relationship between banks’ loan loss provisions and macroeconomic indicators such as GDP growth, real interest rates, credit growth and the concentration of the loan portfolio.

In their work Kalirai and Scheicher (2002) model the impact of key macroeconomic variables, such asindicators of general economic activity, price stability, households’ and corporate sectors’ situation, financial market and external events, on aggregated loan loss provisions (LLP) using a linear regression model and a sensitivity analysis for macro stress testing. Short-term interest rates, GDP growth rates, the stock index and industrial production are found to influence LLP.

Pesola (2001, 2007) confirmed that macroeconomic shocks jointly with financial fragility generate banks loan losses. He employs an econometric model based on panel data to assess the relationship between the ratio of banks’ loan losses and enterprise bankruptcies per capita and macroeconomic variables. His findings suggest that high corporate and household indebtedness, combined with negative macroeconomic shocks contributed to the banking crisis in Sweden, Norway and Finland.

All of these studies confirmed that macroeconomic variables could affect the bank’s portfolio and increase credit risk measured by LLP[3] and NPL.

1. Credit risk and macroeconomic stress test- Serbia and Czech Republic

The main credit risk parameters are PD- probability of default, LGD- loss given default, EAD- exposure at default. PD is used to predict the volume of the gross inflow of non-performing loans NPL’s. Changes in the risk factors can lead to upgrades as well as downgrades of risk parameters (The PD is by far the most popular risk parameter which is followed in stress tests). For example, an increase in price of resources such as oil or energy can have a negative impact on PDs in the automobile or any other industry consuming lots of energy, but it could have a positive impact on the PDs in the country trading these resources.[4]

Depending on the availability of data, credit risk factors and their correlations with macrovariables can be estimated using data on loan performance (historical NPLs, default rates, recovery rates, loan-loss provisions (LLPs) or cost of credit) or using microdata on corporate sector from credit registries and eventually household sector data (Čihak, 2007).

Non-performing loan (NPL) is one of important indicators to evaluate status of portfolio in commercial banks in Serbia and Czech Republic. As NPL rate gets higher, bank need more provision to cover losses on these non-performing loan. Therefore, NPL ratio can reasonably represent the default risk of commercial bank. The share of NPLs in Serbia’s total banking sector loans has been increasing since 2008 and achieve the peak in Q1 2012 where total loans of the banking sector past due for more than 90 days making up 20,4 percent of gross loans(National Bank of Serbia, 2012).

According to Czech National bank, in contrast to 2009–2010, there was no further strong growth in NPLs, and their ratio to total loans declined slightly to 6% at the end of 2011 (compared to 6.3% at the end of 2010). An international comparison between selected EU countries shows that the NPL ratio in the Czech Republic is similar to that in Slovakia (5.6%), higher than in Austria (2.7%) and Belgium (2.8%), and lower than in Poland (8.2%), Slovenia (11.8%). A moderate decline in credit risk is also indicated by the evolution of risk costs, defined as net provisioning relative to total loans and by the evolution of loan restructuring in both the household segment and the non-financial corporations segment. As we can see from the figure below, corporate NPL in 31.10 reached 7.51 of total loans. In 2002 at the end of January the NPL reached peak of 17.7 and after that never was at the same level. NPL house was in 2008 in Q2 in the lowest level at 2.9 of total loans, but after that constantly growing. (Czech National Bank, 2011/2012).

Figure 1.Total NPL ratio, NPL corporate and NPL household ratios- Czech Republic

Source: Author’s calculation.

The increase in the unemployment rate in 2012 Q3 could make problem for household in payment loan obligations. The increasing unemployment causes the default rate to grow as more people lose jobs and their creditworthiness decreases.According to Ministry of labourand social affairs of Czech Republic the lowest unemployment rate was in 2008 Q3 and reached 5.0% and in Q32012 reached 8.4 percent. The peak was in Q4 2010 was 9.6%.

The Czech financial sector is dominated by a few large banks. Banks account for84 percent of the financial sector assets.The banking system’s assets grew rapidly from 2000 to 2008, especially household loans, but balance sheet growth had almost stopped in 2009 as a result of the crisis. The 5 largest banks control more than 70 percent of total bank assets, and the 3 largest ones about 60 percent. (International Monetary Fund, 2012)

The Serbian banking sector at end-Q3 2012 comprised of 33 banks- 21 in foreign and 12 in domestic ownership. Among domestically owned banks, 9 banks were state-owned (either by holding a majority share or being the largest individual shareholder) and 3 were in the ownership of private individuals. Foreign-owned banks dominated the market – they accounted for 74% of total assets, 74% of total capital and 71% of employment of the banking sector, and posted profit of RSD 17.5 bln.(National Bank of Serbia, 2012)

Serbia’s banking sector is well capitalized.According to National Bank of Serbia Capital adequacy ratio in Q3 2012 is 16.4 percent (in Q3 2011 was 19.7%). The share of non-performing loans in total loans is rising, mostly as a result of foreign exchange-induced credit risk. In Q1 2012 the capital adequacy ratio decreased as a result of the combined effect of a fall in regulatory capital and a rise in risk-weighted assets triggered, among other things, by dinar depreciation. Banking sector assets reached 83,5 of GDP in 2011. Bank’s loan portfolio is still dominant, accounting for close to 60% of total banking sector assets in 2011. (National Bank of Serbia, 2011).

Table 1.Selected parameters of the Serbian banking sector

Number of banks / Profit in bln. / Assets in bln / Assets in % / Capital in bln / Capital in %
Total domestic banks / 12 / 21.7 / 685 / 26 / 135 / 25
Total foreign banks / 21 / 22.9 / 1.965 / 74 / 411 / 75

Source:

1. Macroeconomic credit risk model

SorgeVirolainen (2006) highlight two approaches that explicitly link the default probabilities and macroeconomic variables- Wilson (1997a, 1997b) and Merton (1974) models.

These two authors adopt the Wilson framework to perform a macro stress test on credit default probability in Finland and find that default probability distribution by Monte Carlo simulation is significantly different from its normal distribution in stress situations. In comparison, the Merton model integrates asset price changes into default probability evaluation.

Merton's model (1974) was originally developed for the firm`s level but extended for the purposes of the macro stress testing. Merton model integrates asset price changes into default probability evaluation. Merton’s type model for the Czech economy was used in Jakubik (2007). JakubikSchmieder (2008) apply the model on the household and the corporate sectors for the Czech Republic and Germany. They test theeffects of macroeconomic variables on NPL as a measure of the default rate. They concludethat key macroeconomic determinants, such as interest rates, exchange rates, inflation, GDPgrowth and the level of indebtedness, can meaningfully simulate corporate default rates for bothcountries. The results show the greater macroeconomic shocks in Czech Republic than in Germany.

Hamerle, Liebig & Scheule (2004) use factors models to forecast the default probabilities of the individual borrowers in Germany. Merton's model was used also in Drehmann (2005) for the stress testing the corporate exposures of the banks in the UK. He concluded that is quite reassuring as even in the worst conditions expected losses of banks corporate exposures are not high enough to cause a bank failure.

One of the few credit risk models that explicitly links macroeconomic factors and corporate sector default rates was developed by Wilson (1997a, 1997b). Wilson's logistic model was used in studies of Boss (2002) and Virolainen (2004). Boss (2002) and Boss et al. (2006) estimate the relationship between the macroeconomic variables and the credit risk for the corporate default rate in the Austrian banking sector. Virolainen (2004) and Virolainen, JokivuolleVähämaa (2008) develop the macroeconomic credit risk model that estimates the probability of default in Finnish industries. The idea is to model the relationship between default rates and macroeconomic factors and to simulate the evolution of default rates over time by generating macroeconomic shocks to the system. These simulated future default rates and estimates expected and unexpected credit portfolio losses including also the current macroeconomic situation.

For purpose calculation default rates we use Wilson model (1997a,b) in line with Virolainen (2004). First, the average default rate for industry j is modeled by the logistic functional form as

ds,t= 1/1+exp(ys,t)

whereds,tis the default rate in industry s at time t, and ys,tis the industry-specific macroeconomic index, whose parameters must be estimated.

According with Boss (2002)we adopt formulation thata higher value for ys,timplies a better state of the economy with a lower default rate ds,tand vice versa. The logistic functional form is given by:

L(ds,t ) ln (1-ds,t/dj,t)=ys,t

The logit transformed default rate (the industry-specific macroeconomic index) is assumed to be determined by a number of exogenous macroeconomic factors.

Ys,t=ß s,0 + ß s,1 x1,t + ß s,2 x2,t+….+ ß s,nxn,t+ εs,t ,

where βs=(βs,1βs,2,…., βs,n ) is a set of regression coefficients to be estimated for the sth industry, xn,t(x1,t,x2,t…xn,t) is the set of explanatory macroeconomic factors (e.g. GDP, exchange rate, unemployment rate, etc.), and εs,t is a random error assumed to be independent and identically normally distributed.

3.1.Macroeconomic credit risk model for corporate sector-Serbia and Czech Republic

Typically, the credit risk models include a measure of credit risk as dependent variable and macroeconomic variables (i.e., output measures, interest rates, inflation, and the exchange rate) as explanatory variables. The macroeconomic data for the Czech Republic have been taken from the time series archives (ARAD) of the Czech National Bank. For corporate sector as independent variables we used GDP growth rate, unemployment rate, exchange rate CZK/EUR andexchange rate CZK/USD.

Table 2.Summary Statistics, Czech Republic, Corporate sector, observations from 2002:1 to 2012:3

Variable / Mean / Median / Minimum / Maximum / Std. Dev.
gdp / 0.704 / 0.8 / -3.3 / 2.4 / 1.05
un / 8.323 / 8.6 / 5.0 / 10.3 / 1.34
czk_usd / 22.43 / 21.23 / 15.89 / 36.23 / 4.88
czk_eur / 27.90 / 27.78 / 24.29 / 32.98 / 2.75

Source: Author’s calculation.Summary Statistics, Gdp is gross domestic product, un- unemployment rate, exchange rate czk_usd, exchange rate czk_eur.

The macroeconomic credit risk model for corporate sector for Czech Republic is:

ln (1- pdcorp,t/pdcorp,t) = α + ß1gdp,t+ ß2un,t+ ß3czk_eur,t+ß4czk_usd,t

Table 3. Czech Republic, corporate sector, using observations 2002:1-2012:3 (T = 43)

Dependent variable: ln_nplc

Variable / Denoted / Coefficient / Std. Error / t-ratio / p-value
Constant / const / -3.01529 / 0.460675 / -6.5454 / <0.00001 / ***
GDP growth rate / gdp / -0.0710079 / 0.032693 / -2.1720 / 0.03617 / **
Unemployment rate / un / 0.249528 / 0.0279239 / 8.9360 / <0.00001 / ***
Exchange rate CZK/EUR / czk_eur / -0.130102 / 0.0248833 / -5.2285 / <0.00001 / ***
Exchange rate CZK/USD / czk_usd / 0.089606 / 0.0137262 / 6.5281 / <0.00001 / ***
R-squared / 0.841888 / Adjusted R-squared / 0.825244
F(4, 38) / 50.58383 / Hannan-Quinn / -7.776307
P-value(F) / 1.02e-14 / Akaike criterion / -11.02369
rho / 0.603986 / Durbin-Watson / 0.797412

Source: Author’s calculation.Significant at 1% level. Dependent variable: ln_corp.

As shown in the table, the outcome of our analysis demonstrates a important influence of the exchange rate czk_usd. According to t- test we might say that unemployment rate and exchange rate CZK/USD are the substantional explanatory variable.An appreciated exchange rate raises the prices of domestic goods in foreign currency. Appreciated exchange rates results in a higher default rates in the corporate sector for both countries.Positive impact of the CZK/EUR exchange rate on the default rate might be the results of the preference for loans denominated in the euro. GDP have negative signs means that increasing GDP affects positively demand for company’s goods. Such increases may lead in better creditworthiness of the firms.

The negative impact of depreciation of the domestic currency on the default rate is given by the fact that the currency depreciation favours domestic exporters and increases their profits. The increasing GDP stimulates the demand for goods that corporations produce and that increases their profits and ability to repay the debt. The probability of default decreases.

Figure 2. Independent and Dependent variablesfor corporate sector- Czech Republic

Source: Author’s calculaton.Un- unemployment rate, nplc- is nonperforming loan for corporate sector, exch is exchange rate CZK/EUR and exchczk_usd is exchange rate CZK/USD.

The figures shows that nonperforming loan achieve a peak in 2010Q1. Increased unemployment, negative industrial production and local currency depreciation caused problem in repaying debt.

Figure 3. Actual and estimates value for Czech corporate sector ln_npl

Source: Author’s calculation.

Figure 3.shows performance of the estimated model. In 2008 default rate was at the lowest level at about 3% default rate. Furthermore, leading the collapse in worldwide market and distress situation during the crisis the default rate constantly increased. This situation reflects on corporate sector ability to pay debt. After 2008 the default rate constantly decrease and peak 9.1% in 2010.The explanatory variables of corporate sector in case of Serbia we use GDP, producer price index, GDP Euro area and industrial production.