Journal of Credit Risk

Risk.net

Three ways to improve the systemic risk analysis of the Central and Eastern European region using SRISK and CoVaR

Marta Karaś and Witold Szczepaniak

  • The paper presents three modifications to measures of systemic risk – SRISK and ΔCoVaR – and increases the number of analyzed systemically important institutions in the CEE region. The study is more comprehensive than its predecessors with respect to the region.
  • Presented methodology integrates the information about banks' systemicness, interconnectedness, complexity, significance and size. Additionally, we include the data of broad domestic market indexes into the calculation of ΔCoVaR. The new methodology takes into account endogenous systemic risk factors, both at the bank level and the system level, as well as exogenous factors such as the exchange rate or the impact of the biggest European stock markets.
  • The empirical results confirm increased systemic risk for CEE countries in two periods, 2008–2009 and 2011–2013, at the same time identifying countries more prone to fragility and to contagion. The rankings confirm the high potential cost of systemic risk materialization in Central and Eastern Europe. We also show that in some cases, this burden was greater for small CEE countries than it was for Germany that has, by far, the biggest value of systemic risk.
  • The study confirms that systemic risk in the CEE region has the same theoretical properties as it does for advanced economies. The results also show that during periods of calm, SRISK and CoVaR indicate different financial institutions as the weakest links in the given financial system. These results have a macroprudential consequence: the regulatory bodies in the analyzed countries should consider both the information provided by CoVaR and by SRISK when monitoring systemic risk, especially when they want to put banks under scrutiny.

This paper proposes three modifications to the calculation of SRISK and CoVaR. These modifications make it possible to apply the two measures to an additional 31 systemically important financial institutions in the Central and Eastern European (CEE) region. They also add information about interconnectedness and complexity, and illuminate risk factors that are endemic to CEE and Western European stock markets. We empirically analyze Bulgaria, Estonia, Czechia, Hungary, Latvia, Lithuania, Poland, Romania and Slovakia in the period from 2006 to 2018. The results confirm increased systemic risk in the years 2008–9 and 2012–13. Systemic risk rankings demonstrate the significant scale of systemic risk relative to gross domestic product for many of the countries under analysis. The results also confirm that systemic risk in the CEE region has the same theoretical properties as it does in advanced economies. This finding underlines that it is necessary to analyze the CEE region using measures of systemic risk that are at least as sophisticated as those used in the most developed countries.

1 Introduction

Systemic risk is a very complex and challenging concept to measure. According to Benoit et al (2017, p. 131) it is “one of the most elusive concepts in finance”. The European Systemic Risk Board defines systemic risk as “a risk of disruption in the financial system with the potential to have negative consequences for the internal market and the real economy” (Constâncio 2010). As such, systemic risk affects all other types of risk that exist in financial systems.

Studies have proposed multiple methods to quantify systemic risk in advanced economies. (For a thorough analysis of the applicability of 54 such methods to the Central and Eastern European region we refer the reader to Karaś (2019).)

However, such methods are only partially transferable to countries with less advanced financial systems because these countries have much smaller stock markets than advanced economies.

To the best of our knowledge, no other work proposes the methodology outlined in this paper. Thus, the paper adds to recent advances in risk modeling. We propose a set of modifications for the computation of two well-known systemic risk measures: SRISK and CoVaR. The modifications make these risk measures applicable to a much wider range of credit institutions in emerging economies. Moreover, they allow for

  • including more information regarding the structural and ownership characteristics of the studied financial systems (using systemic importance scores (SISs) for system-wide risk aggregation); and

  • using information from domestic and global stock markets to enable a more effective analysis of both local and broader environments.

Further, we adopt a nonstandard approach to calculate CoVaR and ΔCoVaR, as we replace infrequent book-value data with daily stock returns. This replacement increases the sensitivity to behavioral phenomena and their effect on the quantification of systemic risk. It also improves transparency.

We empirically apply the proposed methodology to nine Central and Eastern European (CEE) countries (Bulgaria, Estonia, Czechia, Hungary, Latvia, Lithuania, Poland, Romania and Slovakia). The results indicate how systemic risk levels – specifically fragility (SRISK) and the potential of risk spillover (ΔCoVaR) – change throughout the period from 2006 to 2018. We also compile a set of rankings to facilitate comparison of the studied countries with respect to the price of systemic risk and its burden on the state. Our methodology is also used to calculate the same measures for Germany. This calculation enables a comparison of the scale of systemic risk experienced by the CEE region with the most advanced European economy.

Finally, we run regressions to test three hypotheses.

  1. (1)

    SRISK and ΔCoVaR provide different information about systemic risk in the studied CEE countries.

  2. (2)

    The difference between the results obtained with SRISK and ΔCoVaR relates to the degree of complexity of the studied financial systems.

  3. (3)

    SRISK- and ΔCoVaR-based risk signals converge in periods of higher risk (ie, crises).

The paper proceeds as follows. Section 2 presents a review of systemic risk in CEE countries. In Section 3, we discuss the relevance of SRISK and CoVaR. In Section 4, we outline our proposed modifications. Section 5 presents the technical aspects of the proposals. Section 6 is our empirical analysis, and Section 7 concludes.

2 Systemic risk measurement and the Central and Eastern European region

At least two facts about the financial systems of the CEE region have crucial implications for systemic risk analysis. The first is a typical feature of all emerging markets: the banking sector is the most central part of the financial system; thus, it is also “the main source of risk for financial stability” (Karkowska 2013, p. 4). The second fact is that foreign owners control the CEE banking sector: they possess over 90% of the total banking assets in the region (Radulescu et al 2018, pp. 7–8). As a consequence, CEE countries have relatively homogeneous financial systems, in which foreign-owned banks are major credit suppliers (Dumicic 2018, p. 2).

Interestingly, foreign-owned banks do not list on the local stock markets until these markets become sufficiently large and deep. The Polish example illustrates this phenomenon well. In Poland, all larger banks are listed on the Warsaw Stock Exchange,11 1 See the WIG Banking Index: https://bit.ly/3jmbn5y. while in the other analyzed countries almost none of the foreign-owned banks are locally listed. As a consequence, a much broader set of systemic risk measures are applicable to the Polish financial system than for the other CEE countries.

The fact that so many CEE banking systems depend on foreign-owned banks to supply credit poses a significant macroprudential challenge. Regulators in CEE countries pioneered the use of a wide variety of macroprudential tools, and were using them even before these tools were known as “macroprudential”, that is, before the global financial crisis (Dumicic 2018, p. 10). However, most of the subsidiaries of big Western European banks were able to circumvent the regulations thanks to their close ties with parent institutions (Dumicic 2018). This evasion was likely facilitated by the existing internal markets identified in an empirical study by De Haas and van Lelyveld (2010).

A notable example may be the sharp deleveraging of the Romanian and Bulgarian foreign-owned banks, which affected the supply of credit to the real economy (Radulescu et al 2018, pp. 3–4). Despite their efforts, local regulators were not able to limit this process. Correspondingly, Jočienė (2015, p. 51) concludes that foreign-owned bank subsidiaries in Lithuania, Latvia and Estonia were adopting the business models of their foreign parents even when these were not optimal models for their local environments. Barkauskaite et al (2018) find that a dependence on the foreign parent in strategic decisions is an important macroprudential factor. As they point out, improvement is needed in the methods currently used by regulators for the quantification of systemic risk for such banks.

Since 2015, the European Union has required the regulators of its member states to calculate SISs, which identify other systematically important institutions (O-SIIs). These O-SIIs have been described by the European Banking Authority (EBA) as “institutions that, due to their systemic importance, are more likely to create risks to financial stability whilst maximizing private benefits through rational decisions [and] may bring negative externalities into the system and contribute to market distortions” (European Banking Authority 2014a). Foreign-owned banks, most often the subsidiaries of much bigger Western European and Nordic banks, compose the majority of such O-SIIs in the CEE region.

All central banks in the European Union analyze the external financial environment in their financial stability reports. The long historical experience of systemic risk monitoring leaves no doubt as to the usefulness of such information. The works of Bisias et al (2012), Hattori et al (2014) and Benoit et al (2017) provide an overview of an extensive number of studies that propose various systemic risk measures.22 2 In the course of the larger study, we have identified 54 systemic risk measures used by regulators and researchers. However, only a few of them are applicable to CEE countries due to data limitations. (For details, see Karaś (2019). The work may be accessed at http://www.wir.ue.wroc.pl.)

Despite this profusion, only a few studies present sophisticated systemic risk measures for the CEE countries. Among them are the financial instability index (FII) proposed by Jakubík and Slačík (2013) and the composite indicator of systemic stress (CISS) (Hollo et al 2012) calculated by Kubinschi and Barnea (2016) for Czechia, Hungary, Poland and Romania. The European Central Bank (2019) computes the CISS on an ongoing basis for the advanced economies of the eurozone as well as for Czechia, Hungary and Poland. Unfortunately, the FII and CISS use only broader market indexes (from the money market, the equity and bond markets and the exchange rate market). Therefore, these two measures do not enable the analysis of microprudential bank data such as leverage or various exposures.

In the sample period, among the central banks of the CEE countries selected for this study, only two used more advanced methods of analysis. One is Poland, which employs a relatively elaborate early-warning model using the estimated output gap (Narodowy Bank Polski 2019). The other is Lithuania, which uses the gross-domestic-product-at-risk measure (Lietuvos Bankas 2019). Both approaches allow for combining data regarding the wider financial environment with microprudential indicators to obtain a single score for systemic risk (an indicator or index).

Quantile-based systemic risk measures that combine bank-level and market-level data are even more rarely used in CEE countries. Further, when such measures are calculated, data limitations restrict the sample of banks by excluding the foreign-owned banks that are not listed on the domestic stock exchanges. Table 1 presents details regarding all (to the best of our knowledge) the studies that use this type of systemic risk measure for the CEE region published in the last decade.

Table 1: The systemic risk measures that combine bank and stock market data: application to selected CEE countries, by study. [Total number of financial institutions and the number of O-SIIs included in the studies. Source: own elaboration based on the lists of O-SIIs 2015–19 (European Banking Authority 2019) and the sources cited in the table.]
  Karkowska Engle et al Jajuga et al Andrieş et al Current  
  (2013) (2019) (2017) (2020) paper  
             
  No. of No. of No. of No. of No. of No. of No. of No. of No. of Total
Country institutions O-SIIs institutions O-SIIs institutions O-SIIs institutions O-SIIs O-SIIs O-SIIs
Bulgaria 4 3 3 2 4 3 9 10
Czechia 1 1 2 1 1 1 5 7
Estonia 3 4
Hungary 2 1 2 1 2 1 1 1 6 8
Latvia 1 0 3 6
Lithuania 3 2 1 1 4 4
Poland 8 8 12 8 11 8 10 9 8 12
Romania 2 1 2 2 2 2 3 2 7 9
Slovakia 1 1 2 2 4 2 4 5
In total 21 16 19 13 20 15 24 19 49 65
% of O-SIIs   25   20   23   31 75  
Horizon 2006–12 2000–to date 2005–16 2005–12 2006–18  
Frequency Quarterly Monthly Daily Weekly Daily  
Measure CCA CCA, SRISK SRISK, SRISK,  
      SRISK     CoVaR CoVaR  

As seen in Table 1, the studies that use quantile-based measures for systemic risk quantification in CEE countries were markedly ineffective in capturing the effect of O-SIIs. In all cases except Poland, no more than 50% of the existing O-SIIs were included in the sample. The last column indicates how our methods increase the number of systemically important banks that may be analyzed. In some cases, the increase is by 50–60% (Bulgaria, Czechia, Lithuania, Latvia and Slovakia) and in others it is by about 70–75% (Hungary and Romania). Our methods also enable the first analysis of Estonia. In total, we add 31 new O-SIIs to SRISK and ΔCoVaR computation. Thus, out of 65 O-SIIs identified by the EBA, we analyze a total of 49.

3 SRISK and CoVaR: relevance

Karkowska (2013) reports that systemic risk in the European banking sector is driven initially by bank-specific risk premiums and subsequently by market determinants, such as volatility and stock-exchange capitalization. Large financial leverage, significant maturity mismatches, turbulence in the financial markets and the interactions between market liquidity and financing significantly increase the degree of systemic risk.

Since adverse shocks spill over between financial institutions, systemic risk should not be analyzed using risk measurements focused on individual institutions. Especially in relatively small financial systems (such as those in the CEE region), the risk assumed by one larger institution may cause significant spillover effects. Further, herding behavior, which becomes more likely as the market size decreases, may cause exposure to adverse systemic events, even if individually financial institutions are characterized by a low level of risk. As Adrian and Brunnermeier (2011, p. 2) point out, “regulation based on the risk of institutions in isolation can lead to excessive risk-taking along systemic risk dimensions”. Therefore, there is a strong need to analyze risk in the financial system as a whole (see Duffie 2019; Engle and Ruan 2019).

At the same time, insight into the financial condition of the institutions that comprise the financial system is essential for quantifying systemic risk because institutional fragility entails an increased sensitivity to systemic events. This sensitivity refers especially to the build up of undercapitalization, which makes institutions vulnerable to liquidity fluctuations and spillover effects. Brownlees and Engle (2017, p. 57) argue that “because of the extensive use of leverage in the financial sector, this industry is particularly vulnerable to downward market movements”.

Many fragility-focused measures have received significant recognition in the literature. Comparative studies point to significant advantages of SRISK relative to other measures (see Benoit et al 2014; Acharya et al 2014; Laeven et al 2014, 2016; Brownlees and Engle 2017; Jajuga et al 2017). One such advantage is that SRISK uses information about the build up of undercapitalization. Another is SRISK’s use of the technique for calculating quasi-leverage proposed by Brownlees and Engle (2017), which retains the daily frequency. In effect, SRISK is adaptable to rapidly changing market conditions. The fact that this measure also contains information about the size of the bank further improves its informational properties.

Based on a series of robustness checks, Brownlees and Engle (2017) conclude that the precrisis levels of SRISK are a good predictor of external capital injections, bailouts and worsening macroeconomic conditions (specifically, industrial production and the unemployment rate). Even more interesting is the fact that SRISK is a predictor of other institution-focused risk measures that are based on proprietary regulatory data, such as stress tests. Moreover, since the measure is based on publicly available information, it is very transparent. Finally, because SRISK assumes that the triggering systemic event takes the form of a prolonged market decline, its conditionality is very intuitive (cf, Engle and Zazzara 2018).

The other systemic risk measure selected for this study is ΔCoVaR. Its authors (Adrian and Brunnermeier 2016) describe it as complementary to high-frequency measures based on the marginal expected shortfall (MES), such as SRISK. Similarly, Kuziak and Piontek (2018) confirm that ΔCoVaR does not correspond to the condition of financial institutions or their financial soundness indicators. This empirical observation confirms that ΔCoVaR is a measure of risk spillover and not a measure of fragility as such.

Interestingly, Benoit et al (2014) show that CoVaR reflects the bank’s tail risk (its VaR) conditional on market returns, while SRISK is associated mostly with leverage. Measures based on MES, such as SRISK, are able to determine the institutions most exposed to a potential financial crisis (fragility), whereas CoVaR indicates the institutions that contribute most to a potential crisis (risk spillover potential).

Benoit et al (2014) also report that, in their comparison of various measures, the lowest concordance was that between SRISK and ΔCoVaR (only 9.9%, with similarities between the other measures of about 50%), demonstrating that the differences in measurement output come from the characteristics of the two measures and not from model risk or instability. Additionally, they show that SRISK and ΔCoVaR (ΔCoVaR expressed in euros) produce different rankings of systemically important financial institutions (SIFIs), which also confirms that these two measures have a different focus and capture different information about systemic risk. In the empirical part of this paper, we produce similar rankings that confirm this conclusion for our sample of CEE countries.

Measuring the contribution of a given financial institution to contagion often relies on confidential data about positions and risk exposures that banks provide to the regulator. However, such risk may also transmit indirectly, for instance, through price effects and liquidity spirals. According to Adrian and Brunnermeier (2011, p. 8), the indirect spillover effects are “quantitatively more important”, as a fire sale of assets may lead to losses for all the market participants who hold similar exposures. In the same sense, in a low liquidity environment, each market participant creates a state in which their subsequent individually optimal responses to adverse events are likely to cause externalities for others. Arguably, such exposures are more significant in smaller, more shallow markets typical of emerging economies. ΔCoVaR captures this risk by looking at the increase in tail co-movement between the institution and the system.

The characteristics that align CoVaR-based measures with the theoretical foundation described above include the following (Adrian and Brunnermeier 2011, pp. 9–10).

Conditionality:

the conditioning of the risk measurement on an institution’s value-at-risk (VaR) with a threshold that facilitates a choice of a (very) low probability scenario.

Focus on tail distribution:

the adverse-scenario-based conditioning that typically shifts the mean downward, increasing the variance and potentially also increasing the higher moments, such as negative skewness and kurtosis.

Noncausality:

not distinguishing between causal co-movement and co-movement driven by a common factor, and thereby capturing direct and indirect spillovers.

Directionality:

the ability to measure how the system affects the risk of the institution and how the risk of the institution affects the system, while noting that these relations may be asymmetric.

ΔCoVaR is a statistical measure, so it does not require explicit reference to structural economic models, which makes it readily applicable to almost any financial system, even one with scarce data. The corresponding assumptions are that public market data, such as stock returns, reflect information about publicly traded financial institutions, and also depict the behavioral phenomena that accompany trading.

The measures selected for this study are also relatively well suited to distinguish between the risk triggers which are of systemic importance and those that are not. They enable us to limit the input data to O-SIIs, which means that any kind of distress related to a nonsystemic bank is automatically disregarded. Also, as the measurements are based on stock market reactions, they depict the sentiment of market participants toward the expected effect of various events. An example might be sale of a large branch of one systemically important bank to another. If the markets view such an event with optimism, eg, because this solution means a future healthier business position for the seller, then we observe a positive reaction in the markets. This is the case even if the event initially stresses the books of the seller by putting a strain on the asset side of its balance sheet.

A good illustration of such an event was the withdrawal of Deutsche Bank (one of the Polish O-SIIs) from Poland. Despite a drastic drop in assets, the downscaling did not act as a systemic trigger for the banking system, because the bank’s loans were secured by another global SIFI that was expanding into the Polish market at the time (Santander Bank). The German and Polish stock markets did not react badly to news of the withdrawal and did not record any increase in systemic risk. The reaction of the risk measures selected for this study reflected the actual situation: the takeover did not trigger systemic risk in the Polish or German financial systems.

Another interesting example relates to one of the SIFIs present in the CEE region: Crédit Agricole (CA). CA initially took over the Greek Emporiki Bank, only to later sell it to another systemically important Greek bank, Alpha Bank. This was significant for some CEE countries, such as Romania, in which Emporiki Bank had a major presence. The move allowed CA to grow in Albania, Bulgaria and Romania.

As the sovereign debt crisis unfolded, CA became exposed to Greek government bills and bonds. Because the Greek branches had by that point been unprofitable for some time, CA decided to sell them at a loss (for €1 per stock) to Alpha Bank. This decision was actually welcomed by the stockholders of CA and in the following weeks investors saw a price recovery for CA’s stock.

The systemic risk measures applied in this paper did not react to this sale as they would to a systemic risk trigger, despite the fact that CA had realized a book-value loss from the transaction. There was no actual systemic risk that either Greece or the parent institution could transfer to the CEE region. Thus, the systemic risk measures reacted appropriately.

4 Three proposals of modification

In order to effectively apply SRISK and ΔCoVaR to the CEE region, we propose three modifications. The first allows for the expansion of the sample of institutions. The second consists in adding a broad domestic market index to the input data, which increases the sensitivity of risk measures to domestic macroeconomic conditions. The third proposal introduces a new way to establish the model of the financial system by modifying how the weights of banks are calculated. Proposals 1 and 3 refer to both risk measures, while Proposal 2 refers specifically to ΔCoVaR. Further, we follow an approach similar to Benoit et al (2014) in which we only use stock market data for the calculation of CoVaR and ΔCoVaR.33 3 The precise methodology we use is an extension of the approach used in our earlier studies (see Karaś and Szczepaniak 2017).

This approach maximizes the sensitivity of the measures to the behavioral aspects of systemic risk and decreases dependence on book-value data.

4.1 Proposal 1: proxy approach

The inclusion of as many O-SIIs as possible in systemic risk analysis is of the utmost importance. For the CEE region, this is challenging because many O-SIIs are banks owned by foreign SIFIs and are not listed on the CEE stock exchanges. While these O-SIIs are exposed to the risk generated in equity markets, this risk is generated mostly abroad, where the parent SIFIs are listed. This exposure justifies the inclusion of the systemic risk factors that are generated in the relevant foreign stock exchanges in the quantification of the CEE region’s systemic risk.

To quantify this exposure, we use a proxy for each foreign-owned subsidiary O-SII, while correcting for the O-SII’s size. This simple solution is theoretically justified because

  • foreign-owned subsidiaries have limited freedom to choose the risks they expose themselves to,

  • they follow the business model of the parent,

  • they depend to a significant extent on parent–subsidiary financing, and

  • they are affected by the reputational risk and financial condition of the parent.

Further, CEE regulators have limited influence over the functioning of such banks. At the same time, the exposure of these O-SIIs to local stock markets and the risk factors that they generate is limited.

In practical terms, our solution tackles the issue of data availability by allowing for the addition of 31 new O-SIIs to the sample, increasing the calculability of SRISK and ΔCoVaR by more than a factor of two in comparison with previous studies. The number of stock markets included in the analysis increase by seven: the Frankfurt Stock Exchange, Euronext, the Athens Stock Exchange, the Vienna Stock Exchange, the Berlin Stock Exchange, the Italian Stock Exchange in Milan and the Oslo Stock Exchange.

The CEE region’s O-SIIs that are still excluded from the analysis include the banks that recently either entered or exited the analyzed financial systems. Thus, the data needed to cover the whole sample period are unavailable. Also excluded are banks that do not use any form of stock-market-based funding (either directly or by proxy): cooperatives, credit union banks and banks owned privately or by state agencies. Thus, these banks have business models that are different from the majority of the region’s O-SIIs, and they are not exposed to the risk factors important for SRISK or ΔCoVaR.

Our proxy approach also accounts for two channels of contagion significant to the CEE region: the large European stock markets and the effect of risk (including reputational risk) generated by the parent SIFIs. The independence from local listings makes our approach applicable to a greater number of emerging economies, which significantly expands the possible reach of the quantile-based systemic risk measurement in less developed countries.

In order to retain the information about the scale of each local O-SII’s potential impact on systemic risk, we use the data from the publicly available financial statements of each foreign-owned subsidiary. This solution allows us to capture book-value changes such as the growth of the analyzed banks, sudden downsizings and sales and mergers, as well as changes in leverage.

We use the size difference between the parent and its subsidiary to proportionally rescale the price of stocks and the market capitalization of the parent bank (see (5.3)–(5.5)). In effect, we obtain a set of data where assets, liabilities, debt, etc, are those of the subsidiary and the stock prices correspond to its size. At the same time, by rescaling we generate a price time series and a market-value time series that capture the price effects and other risks generated in the foreign stock markets.

The sale of Emporiki Bank by CA mentioned above illustrates the two aspects of our proxy approach. The transaction affected the assets side of CA’s book; however, it did not mechanically affect its stock price. Since our approach only considers the assets of the local CEE branch, it de facto disregards the changes in assets of the parent institution. Simply put, the book-value loss of CA is not reflected by the SRISK or ΔCoVaR of the subsidiaries. However, the price movements of the parent’s stock are captured by these measures. In the CA case, the recorded systemic risk effect is minimal because the sale was not perceived as negative by the stockholders. Our approach therefore has the benefit of limiting the potential to transfer risk through the assets side so long as this risk does not destabilize the parent institution. Additionally, we are able to keep a balance between the effect of the subsidiary’s distress and the contagion effect.

4.2 Proposal 2: stock market indexes

The general premise of the second proposal is that the prices of banks’ stocks do not capture all of the systemic risk significant to the countries in the CEE region. We also need to be able to measure the fraction of this risk that relates to the economic and financial distress of the specific country under analysis. In order to take this endogenous risk into account, we use broad domestic stock market indexes. In each case, we select the broad stock index composed of the largest listed companies. Each index represents not only stock market conditions but also the state of the economy, as it includes the most prominent national companies from various nonfinancial industries.

In practice, we include a domestic market index in ΔCoVaR through the following procedure. We consider each index as if it were an average-size O-SII (see (5.6) and (5.7)) and we follow the calculation procedures applied to financial institutions. This solution allows us to increase the number of stock exchanges included in the analysis by another three from the CEE region: the Czech, Estonian and Latvian domestic stock exchanges. Thus, in total we consider sixteen stock exchanges in our study.

ΔCoVaR is a measure of risk spillover and the index makes it possible to add a vital contagion channel to the quantification of systemic risk. This modification is feasible only for ΔCoVaR, as SRISK requires, inter alia, data on leverage that do not exist for a stock market index. From a theoretical perspective, SRISK is a measure of fragility, which is a bank-specific characteristic. Thus, introducing a contagion-type risk factor in its calculation would be inappropriate.

4.3 Proposal 3: model of financial system based on scores

The third proposal relates to the model of the banking system. Our quantile-based measures, by assumption, use an embedded model of the system that consists of a collection of financial institutions. The existing approaches establish the weights of these institutions based on their size (either total assets or capitalization). We instead use SISs, recommended by the EBA for identifying O-SIIs. This approach takes advantage of the characteristics that make banks systemically important, acknowledging that, for the CEE region, there are other factors just as important as size.

Each SIS is derived based on the criterions for the assessment of O-SIIs that are pursuant to Article 131 (3) of Directive 2013/36/EU (European Union 2013). These criterions are a compromise between a unified approach to the identification of O-SIIs in all European countries and the specific factors that make each European region unique (European Banking Authority 2014a). They are outlined in each yearly notification to the EBA and include size (weighted between 25% and 40%), interconnectedness (15–25%), relevance to the economy (25–30%) and complexity (15–25%).44 4 The original proposal by the EBA postulated equal (25%) weights for all four characteristics (see European Banking Authority 2014b). However, the EBA has later allowed a certain level of variability (for details please compare notifications at the European Systemic Risk Board website: https://bit.ly/3wWxkwj). The SISs are calculated and disclosed by national regulators. Once we use the SISs to derive the weights of the institutions in each system, all the above information regarding the characteristics of the O-SSIs becomes embedded in our systemic risk analysis.

There is a degree of freedom in how the regulators obtain the scores. The process includes, for instance, expert assessments of complexity. Also, local regulators individually set the thresholds for which institutions qualify as O-SIIs. For example, the score needs to be at least 350 for Hungary but only 275 for Germany. Further, each country may set its own approach to obtain the criterions for systemic scoring. Importantly, every departure from the baseline must be accepted by the EBA. This solution ensures a compromise between the comparability of scores across countries and flexibility. The EBA recognizes that local regulators are best equipped to assess which institutions are indeed systemically important for their countries.

In effect, the SIS not only identifies systemic institutions, but also quantifies the differences in their significance for systemic risk. We use the SISs to derive the weights of the institutions in each financial system modeled by our study. We apply the average SIS score for each institution, obtained as the mean of all scores disclosed by the regulator within the sample period for the particular institution (see (5.1) and (5.2)). The method is applied systematically in all ten financial system models.

We chose the average because, in the studied period, the variability of each of the score-based weights for approximately 90% of the cases did not exceed 4%. For 75% of cases, it was within the range of 0.1% to 1.5%. However, we have developed a solution for using time-varying weights that is readily applicable for any case where the scores would vary significantly over time. We report the values of the standard deviation of the scores for each bank in Table 2 together with the average score.

4.4 Maximizing the impact of systemicness and behavioral phenomena by data selection

Several papers propose alternative ways to calculate CoVaR (see, for example, Wong and Fong 2011; Adams et al 2014; Adrian and Brunnermeier 2016). We base our computations only on stock market data, which is similar to Benoit et al (2014), who use the dynamic conditional correlation-generalized autoregressive conditional heteroscedasticity (GARCH-DCC) model instead of quantile regression. Our approach assumes the creation of a synthetic index of the banking sector of each CEE country. The results in Benoit (2014) show that the use of a domestic index is crucial, and we argue, moreover, that it is necessary to use a domestic index focused on systemic risk.

Our approach corresponds to the concept of a financial-system model described in Proposal 1: it allows us to measure systemic risk based only on those institutions that are deemed systemically important by the regulators of each country.55 5 The assumption is that each national regulator properly assesses the potential effect of each O-SII on the local financial system and environment. At the same time, Proposal 2 mitigates the risk of underestimating the local macroeconomic conditions when measuring systemic risk.

This approach also enables us to use the SISs to properly weight the effect of each O-SII on the risk within the index (Proposal 3).

Finally, this approach is necessary if we want to have a coherent universal method to apply to a larger set of developing countries, because financial (or banking) market indexes vary significantly between different emerging economies. In some cases, such indexes do not even exist. Therefore, a synthetic index is both beneficial and necessary.

Basing the calculations of ΔCoVaR on stock market data is also desirable from the perspective of behavioral finance. In particular, by replacing book-value data with stock-market data we account for several phenomena that increase systemic risk (see, for example, Gale and Allen 2003; Shiller 2003; Brunnermeier 2008; Acharya and Merrouche 2013; Gale and Yorulmazer 2013). We consider the market sentiment reflected in the pricing of stocks, including the effect of systemically important events, such as runs on financial instruments and market freezes.

Moreover, access to book-value data is difficult and time-delayed, especially when the data are published only in monthly or quarterly reports prepared by banks for regulators. Using stock-market data also tackles the problem of creative accounting and window dressing by financial institutions in distress. Ultimately, the issue of transparency aside, the daily time series has better signaling properties than a weekly or monthly time series.

From a technical standpoint, it is necessary to apply the same estimation methods for both measures used in this paper in order to minimize the effect of model risk on comparative results. Only the use of stock market data allows us to obtain a time series that is rich enough in observations to confidently estimate the necessary GARCH models. As a consequence, the modified ΔCoVaR is a better candidate for contemporaneous systemic risk analysis in this study than the original proposal.

5 Technical description of the modified measures

For a general case, a national financial system s is defined as a composite of i O-SIIs, where i=1,,N. The rate of return of financial system s at time t is defined as

  rs,t=i=1Nwiri,t,   (5.1)

where ri,t is the rate of return on the equity of institution i, and wi is the weight of financial institution i in system s, which is equal to

  wi=SISiSISs,   (5.2)

where SISi is the systemic importance score of institution i, and SISs is the total SIS of all O-SIIs in the national financial system s.

In the case of O-SIIs that are not listed in the domestic stock market, the rate of return ri,tdomestic and the market capitalization MVi,tdomestic are established as

  ri,tdomestic=cri,tri,tparent,   (5.3)
  MVi,tdomestic=cri,tMVi,tparent,   (5.4)

where the conversion ratio is equal to

  cri,t=Ai,tdomesticAi,tparent,   (5.5)

Ai,tdomestic is the book value of the O-SII’s assets and Ai,tparent is the book value of the parent institution’s assets.

For CoVaR, the set of O-SIIs is supplemented by the domestic stock index. In this case, financial system s consists of i financial institutions and the stock index. Here, the rate of return rs,t of system s at moment t is defined by (5.6), where ri,t is the logarithmic rate of return on institution i and wi is the institution’s share in the system, that is, the ratio of the institution’s score value scri to the total score of the entire system: scrs=i=1N-1scri. In the case of the stock index, rd,t is the logarithmic rate of return on the domestic index, while the index share wd is the value of the institutions’ average share (mean wi) in each analyzed financial system

  rs,t=i=1N-1ri,twi+rd,twd,   (5.6)

where ri,t is the logarithmic rate of return on institution i at time t, rd,t is the logarithmic rate of return on the domestic index at time t, wi=scri/scrs is the share of institution i in the system s and wd=i=1N-1wi/(N-1) is the share of the domestic index in the system s.

When determining the currency-denoted value of system-level CoVaR, we establish the index capitalization MVd as the average capitalization of financial institutions for each considered system:

  MVd=i=1N-1MViN-1,   (5.7)

where MVi is the capitalization of institution i.

The VaR of financial institution i in system s at the level of confidence (1-q) equals

  VaRi,tq(ri,t)=inf{r:F(r)q},   (5.8)

where Fi is the cumulative distribution function of ri,t.

Since P(ri,tVaRi,tq)=q, VaRi,tq can be determined as a q-quantile of the distribution Fi and the VaR of an individual financial institution is given by

  VaRi,tq=σi,tFi-1(q),   (5.9)

where σi,t is the volatility of the rates of return at time t.

The expected shortfall (ES) of system s is defined as the average of all the losses that are greater than or equal to the VaR:

  ESs,t(c)=Et-1(rs,trs,t<VaRs,tq)=i=1Nwi,tEt-1(ri,trs,t<VaRs,tq),   (5.10)

where c is a quantile of the distribution of rs equal to VaRs,tq for q=1%.

The marginal expected shortfall (MES) is given as a partial derivative computed as

  MESi,t(c)=ESs,t(c)wi,t=Et-1(ri,trs,t<c).   (5.11)

The long-run marginal expected shortfall (LRMES) is defined as the expectation that the financial institution’s multiperiod return on equity is conditional on the systemic event,

  LRMESi,t(c)=1-exp(-γMESi,t(c)),   (5.12)

where γ is the correcting factor relative to the length of the assumed horizon.

The long-run market decline corresponds to a 40% loss over a horizon of six months (cf, Engle and Zazzara 2018). By assumption, we ignore the negative values of SRISK (ie, any potential capital surpluses). Thus, SRISK is defined as

  SRISKi,t=max[0;k(Di,t+(1-LRMESi,t)Vi,t)requiredcapital-(1-LRMESi,t)Vi,tcurrentcapital],   (5.13)

where Di,t is the value of debt at time t, Vi,t is the market value of equity at time t and k is the regulatory minimum capital ratio.

We set the k parameter at 8%, following Brownlees and Engle (2017) and Engle and Zazzara (2018). As presented before, the system-wide SRISKs,t of financial system s is computed as the sum of all SRISKi,t for all O-SIIs in that system, where i=1,,N.

CoVaR captures the total price of risk in the financial system that is conditional on an institution being in distress. Formally, CoVaRs,t of financial system s corresponds to the VaRtq of the market return that is obtained conditionally from event c(ri,t) that is observed for financial institution i. In computation, we use the concept of the financial system and its VaR that are outlined in (5.1)–(5.7) as well as the following probability equation:

  P(rs,tCoVaRsc(ri,t)q)=q.   (5.14)

ΔCoVaRi,t captures the marginal contribution of institution i to overall systemic risk in a noncausal sense. The distress state c of financial institution i is defined as VaRi,tq for q=1%, and distress is understood as every instance when losses are at least equal to VaRi,tq:

  ΔCoVaRi,tq=CoVaRs|ri,tVaRi,t(q)t-CoVaRs|ri,t=Median(ri,t).   (5.15)

To establish the value of ΔCoVaR, we multiply the ΔCoVaR of each institution by its size (market capitalization) expressed in euros. We use daily euro exchange rates for each domestic currency.

5.1 Estimation

For the sake of comparability and coherence, we apply the same estimation procedures to both systemic risk measures. We follow the literature (Brownlees and Engle 2017; Benoit et al 2014, 2017) and use GJR-GARCH for conditional volatility and GARCH-DCC for dynamic correlation:66 6 GJR denotes Glosten–Jagannathan–Runkle. Thorough documentation of the properties of each of these models as well as details regarding their construction may be found at the V-Lab homepage: https://vlab.stern.nyu.edu/docs.

  rt=Gtυt,   (5.16)

where rt is the vector of (rs,t,ri,t) at time t, and Gt is the conditional variance–covariance matrix at time t.

The matrix takes the form

  Gt=(σs,t2σi,tσs,tρi,tσi,tσs,tρi,tσi,t2),   (5.17)

where σs,t is the conditional standard deviation of system s at time t, σi,t is the conditional standard deviation of financial institution i at time t and ρi,t is the time-varying conditional correlation.

Following the assumed GARCH method, we use the vector vt of independent and identically distributed random variables (εs,t,εi,t) such that E(υt)=0 and E(υtυt)=I2 is a 2×2 unit matrix. The expected shortfall is determined based on the estimator

  MES^i,t(VaRs,tq)=σ^i,tρ^i,tE^t-1(εs,tεs,t<κ)+σ^i,t1-ρ^i,t2E^t-1(εi,tεs,t<κ)   (5.18)

for

  κ=VaRs,tqσs,t,K(x)=-x/hk(u)du,  

where k(u) is a normal distribution density function and h=T-1/5,

  E^t-1(εs,tεs,t<κ) =t=1TK(κ-εs,t/h)εs,tt=1TK(κ-εs,t/h)   (5.19)
and
  E^t-1(εi,tεs,t<κ) =t=1TK(κ-εs,t/h)εi,tt=1TK(κ-εs,t/h).   (5.20)

6 The empirical application of the proposed measures

The sample period for this study is from 2006 to 2018, which includes the global financial crisis, the sovereign debt crisis and the period of the subsequent economic downturn. The total sample of banks equals 53 (49 for the CEE region and 4 for Germany; see Table 2).

Table 2: SIFIs considered in the study ordered by country. [Average SIS and its standard deviation (SD) in the sample period. *Named DNB and Nordea before 2018. **The SIS established by the regulator via the expert method.]
Country Bank Avg. SIS SD Type Ultimate EU parent (owns or controls)
Bulgaria UniCredit Bulbank A.D. 1901 1.51 O-SII UniCredit S.p.A.
  United Bulgarian Bank A.D. 0867 3.58 O-SII KBC Group NV
  First Investment Bank A.D. 1171 0.75 O-SII
  DSK Bank EAD 1240 5.00 O-SII OTP Bank Nyrt.
  Societe Generale Expressbank A.D. 0513 4.43 O-SII Société Générale S.A.
  Raiffeisenbank (Bulgaria) EAD 0674 0.17 O-SII Raiffeisen Bank International A.G.
  Eurobank Bulgaria A.D. 0725 1.34 O-SII Eurobank Ergasias S.A.
  Central Cooperative Bank A.D. 0493 0.79 O-SII
  Piraeus Bank Bulgaria A.D. 0255 2.21 O-SII Piraeus Bank S.A.
Czechia Československá obchodní banka. a.s. 2140 1.72 O-SII KBC Group NV
  Komerční banka. a.s 1455 1.44 O-SII Société Générale S.A.
  Česká spořitelna. a.s 1508 0.61 O-SII Erste Group Bank A.G.
  UniCredit Bank CZ and SK. a.s 0912 2.55 O-SII UniCredit S.p.A.
  Raiffesenbank a.s. 0440 0.24 O-SII Raiffeisen-Landesbanken-Holding GmbH
Estonia Swedbank AS 3194 10.75 O-SII Swedbank A.B.
  AS SEB Pank 2029 4.94 O-SII Skandinaviska Enskilda Banken A.B.
  AS Luminor Bank 1902 3.82 O-SII Luminor Group A.B.*
Hungary OTP Bank Nyrt. 2891 1.72 O-SII
  UniCredit Bank Hungary Zrt. 0937 0.81 O-SII UniCredit S.p.A
  Kereskedelmiés Hitelbank Zrt. 0809 0.75 O-SII KBC Group NV
  Erste Bank Hungary Zrt. 0600 0.32 O-SII Erste Group Bank A.G.
  Raiffeisen Bank Zrt. 0596 0.46 O-SII Raiffeisen Bank International A.G.
  CIB Bank Zrt. 0434 0.47 O-SII Intesa San Paolo S.p.A
Latvia Swedbank A.S. 3080 0** O-SII Swedbank A.B.
  SEB banka A.S. 2434 0** O-SII Skandinaviska Enskilda Banken A.B.
  DNB banka A.S. 0430 0** O-SII Luminor Group A.B.*
Lithuania SEB bankas A.B. 3649 5.71 O-SII Skandinaviska Enskilda Banken A.B.
  Luminor Bank A.B. 2276 3.72 O-SII Luminor Group A.B.*
  Swedbank A.B. 2046 7.69 O-SII Swedbank A.B. (itself; Sweden)
  Šiauliu bankas A.B. 0840 4.08 O-SII
Poland PKO BP S.A. 1480 1.37 O-SII
  Bank Polska Kasa Opieki S.A. 1085 1.43 O-SII
  Bank Zachodni WBK S.A. 1041 1.37 O-SII Banco Santander
  ING Bank Ślaski S.A. 0870 0.81 O-SII ING Bank N.V.
  mBank S.A. 1017 1.15 O-SII Commerzbank A.G.
  Bank Handlowy w Warszawie S.A. 0518 1.17 O-SII
  Millennium Bank S.A. 0424 1.53 O-SII Banco Comercial Portugues
  Deutsche Bank Polska S.A. 0259 2.82 O-SII Deutsche Bank A.G.
Romania Banca Transilvania S.A. 1276 3.97 O-SII
  UniCredit Bank S.A. 1432 3.68 O-SII UniCredit S.p.A.
  Banca Comercialală Română S.A. 1503 4.32 O-SII Erste Group Bank A.G.
  BRD – Groupe Societe Generale S.A. 1165 2.05 O-SII Société Générale S.A.
  Raiffeisen Bank S.A. 0988 1.43 O-SII Raiffeisen Bank International A.G.
  Alpha Bank România S.A. 0422 0.92 O-SII Alpha Bank
  OTP Bank Romania S.A. 0305 0.38 O-SII OTP Bank Nyrt.
Slovakia Všeobecná Úverová Banka A.S. 2189 1.36 O-SII Intesa San Paolo S.p.A
  Slovenská Sporitel’ňa A.S. 1759 0.80 O-SII ERSTE Group Bank A.G.
  Tatra Banka A.S 1362 0.91 O-SII Raiffeisen-Landesbanken-Holding GmbH
  Československá Obchodná Banka A.S. 1223 0.39 O-SII KBC Group N.V.
Germany Deutsche Bank A.G. 2727 0.91 G-SIFI
  Commerzbank A.G. 0837 1.53 O-SII
  Unicredit Bank A.G. 0448 1.87 O-SII UniCredit Group
  ING DiBa A.G. 0129 0.53 O-SII ING Bank N.V.
SRISK for the CEE region in the period 2006--19.
Figure 1: SRISK for the CEE region in the period 2006–19.
DeltaCoVaR for the CEE region in the period 2006--19.
Figure 2: ΔCoVaR for the CEE region in the period 2006–19.

The data comprise approximately 170 000 entries. The data sources are Thomson Reuters DataStream and the publicly available financial statements of the sample banks, as well as official documentation related to the identification of O-SIIs published by national regulators. The results include system-level ΔCoVaR and SRISK presented in Figures 112, the rankings of countries based on these measures presented in Tables 3 and 4 and the regressions of ΔCoVaR and SRISK presented in Table 5. Additionally, we provide granular bank-level results in the online appendix.

6.1 Systemic risk between 2008 and 2018 in the CEE region: system-level view

The periods of intensified systemic risk are 2008–9 and 2011–13. These periods are indicated by both SRISK and ΔCoVaR. However, the two measures display a different scale of risk and a different view of its relative changes over time (see Figure 1 for SRISK and Figure 2 for ΔCoVaR).

Moreover, comparison of the two measures indicates differences between countries (see Figures 312). For several countries, such as Latvia, fragility (SRISK) is larger than the risk spillover potential (ΔCoVaR); for others, the situation is the opposite. This is especially evident for Poland, where almost all banks run surpluses during the less turbulent times, while SRISK spikes only during the periods of the highest distress. On the other hand, Figures 1, 2 and 10 show that Poland is characterized by a very high level of contagion risk, which reflects its strong ties with the Western European financial systems. For other countries, such as Germany, both indicators are comparably high. The rankings for the most turbulent periods (2008–9 and 2011–13) as well as the ranking for the end of the sample period (2017–18) confirm these observations (see Tables 3 and 4).

SRISK and DeltaCoVaR for Bulgaria between 2008 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Figure 3: SRISK and ΔCoVaR for Bulgaria between 2008 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
SRISK and DeltaCoVaR for Czechia between 2007 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Figure 4: SRISK and ΔCoVaR for Czechia between 2007 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
SRISK and DeltaCoVaR for Estonia between 2007 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Figure 5: SRISK and ΔCoVaR for Estonia between 2007 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
SRISK and DeltaCoVaR for Germany between 2007 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Figure 6: SRISK and ΔCoVaR for Germany between 2007 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
SRISK and DeltaCoVaR for Hungary between 2007 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Figure 7: SRISK and ΔCoVaR for Hungary between 2007 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
SRISK and DeltaCoVaR for Latvia between 2007 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Figure 8: SRISK and ΔCoVaR for Latvia between 2007 and 2018. (a) System CoVaR (in %). (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
SRISK and DeltaCoVaR for Lithuaniua between 2007 and 2018. (a) System CoVaR (in %), (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Figure 9: SRISK and ΔCoVaR for Lithuaniua between 2007 and 2018. (a) System CoVaR (in %), (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
SRISK and DeltaCoVaR for Poland between 2007 and 2017 . (a) System CoVaR (in %), (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Figure 10: SRISK and ΔCoVaR for Poland between 2007 and 2017 . (a) System CoVaR (in %), (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
SRISK and DeltaCoVaR for Romania between 2007 and 2018. (a) System CoVaR (in %), (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Figure 11: SRISK and ΔCoVaR for Romania between 2007 and 2018. (a) System CoVaR (in %), (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
SRISK and DeltaCoVaR for Slovakia between 2007 and 2018. (a) System CoVaR (in %), (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Figure 12: SRISK and ΔCoVaR for Slovakia between 2007 and 2018. (a) System CoVaR (in %), (b) System SRISK (in %). (c) System CoVaR (in thousands of euros). (d) System SRISK (in thousands of euros).
Table 3: Country rankings based on the total price of systemic risk expressed in euros in three periods: 2008–9, 2011–13 and 2017–18. [Rank 1 means the highest risk; rank 10 is the lowest.]
  2008–9 2011–13 2017–18
       
Rank CoVaR SRISK CoVaR SRISK CoVaR SRISK
01 Germany Germany Germany Germany Germany Germany
02 Poland Czechia Poland Czechia Poland Czechia
03 Czechia Hungary Czechia Hungary Czechia Hungary
04 Romania Latvia Hungary Latvia Hungary Latvia
05 Hungary Romania Romania Romania Romania Romania
06 Bulgaria Poland Estonia Slovakia Lithuania Bulgaria
07 Slovakia Slovakia Lithuania Bulgaria Estonia Slovakia
08 Estonia Bulgaria Slovakia Lithuania Slovakia Poland
09 Lithuania Lithuania Bulgaria Estonia Bulgaria Lithuania
10 Latvia Estonia Latvia Poland Latvia Estonia

6.2 Systemic risk between 2008 and 2018 in the CEE region: institution-level view

The analysis of the granular bank-level results (see figures in the online appendix) leads to some further observations. For instance, there is a clear distinction in terms of scope between SRISK and ΔCoVaR. SRISK results are bank-specific, that is, they show significant differences between the fragility of individual banks, both in scale and over time. Different financial institutions have high risk spikes at different points in time and show different trends that are clearly dependent on changes in leverage.

ΔCoVaR presents differently. High correlations may be observed not only at the bank level or within one country (see, for example, Bulgaria or Hungary in the online appendix) but also at the cross-border level, especially in economically linked regions such as the Baltic states (see Figures 258 and 9). Importantly, the correlations are high regardless of whether a given bank is listed domestically or on a large European stock market (ie, when the proxy is used). This demonstrates that the systemic risk measured with ΔCoVaR is system-specific and not bank-specific and that there are strong channels of contagion between Western, Northern, Central and Eastern Europe.

The bank-level results for SRISK (see the online appendix) are expressed in euros and depict two groups of banks with different funding structures. One group are the banks that run relatively high, but less cyclical, leverage. These banks are predominantly the subsidiaries of big Western and Northern European financial institutions (see Raiffeisen, Erste Group, Swedbank and Skandinaviska Enskilda in the online appendix). The other group of banks run a capital surplus in normal conditions (which is denoted by SRISK equal to zero), but in turbulent periods they experience high spikes in systemic risk that show a much more significant cyclicality. These are Western European banks (see Société Générale and KBC in the online appendix) as well as some CEE domestic institutions (see Banca Transilvania, Šiauliu Bankas and PKO BP in the online appendix).

Such trends refer to all local subsidiaries regardless of the country. They show that sizable international parent banks execute their business models on a European level, with little variability at the local scale. Interestingly, this observation holds throughout the entire study period and confirms the necessity for the effective introduction of the Basel III regulations. It also indicates a limited effect of the Basel III countercyclical buffers in the period 2015–19 for the analyzed countries, which may be because the buffers were still being phased in and were working at a very limited capacity over this period.

Currency risk is also worth mentioning in the context of the CEE countries. The euro exchange rate is a significant macroeconomic risk factor for developing countries in the European Union that are outside the eurozone. Applying daily exchange rates when quantifying systemic risk77 7 See the discussion that refers to the effect of currency on systemic risk measurement results in Benoit et al (2019). Here, we use daily exchange rates purposefully, as developing countries are affected by currency risk quite differently from the developed countries analyzed in Benoit et al (2019). captures the effect of this exchange rate on the levels of risk denoted in euros. Both the system-level and bank-level results depict the points in time where a currency’s stability or instability, respectively, enlarged or mitigated the endogenic systemic risk peaks. This is visible in the data for all the analyzed CEE countries (compare the upper and lower charts in Figures 312 as well as the figures in the online appendix).

6.3 Systemic risk between 2008 and 2018 in the CEE region: country rankings

The rankings presented below are based on SRISK (by default expressed in currency) and ΔCoVaR (cf, Benoit et al 2019). They facilitate the comparison of the analyzed countries in two ways: in currency, by making it possible to analyze the size of the risk expressed as its price; and relative to gross domestic product, enabling the comparison of the countries’ capacities to cope with the burden of the financial crisis.

The first ranking (Table 3) is based on the price (value) of the systemic risk measures, that is, the price of systemic risk denoted in euros. This ranking demonstrates that Czechia and Hungary are characterized by the highest systemic risk among the CEE countries, regardless of the period, in terms of both fragility and risk spillover. At the same time, Germany remains the greatest source of systemic risk in the sample. For Poland, ΔCoVaR and SRISK place it at opposite ends of the ranking, which points to a high potential for risk spillover but relatively low fragility. The Baltic states have the lowest ΔCoVaR risk ranking. In the case of SRISK, a significant difference exists between Latvia and other Baltic countries, in that Latvia is characterized by much higher fragility.

Table 4: Country rankings based on the total price of systemic risk expressed in euros relative to GDP in three periods: 2008–9, 2011–13 and 2017–18. [Rank 1 means the highest risk; rank 10 is the lowest.]
  2008–9 2011–13 2017–18
       
Rank CoVaR SRISK CoVaR SRISK CoVaR SRISK
01 Poland Latvia Poland Latvia Poland Latvia
02 Czechia Germany Czechia Germany Estonia Hungary
03 Hungary Estonia Estonia Hungary Czechia Germany
04 Estonia Hungary Germany Czechia Latvia Bulgaria
05 Bulgaria Czechia Latvia Bulgaria Lithuania Czechia
06 Germany Lithuania Hungary Estonia Hungary Slovakia
07 Romania Bulgaria Lithuania Slovakia Germany Estonia
08 Latvia Slovakia Bulgaria Lithuania Bulgaria Lithuania
09 Lithuania Romania Romania Romania Slovakia Romania
10 Slovakia Poland Slovakia Poland Romania Poland

The second ranking (Table 4) represents the burden of systemic risk relative to GDP. Despite the smaller value scale of systemic risk in CEE countries compared with Germany, this risk is large on the domestic scale and puts a heavy burden not only on the previously identified high-risk CEE countries but also on the Baltic states. In the case of SRISK, for Poland and Romania we see the effect of relatively strong economies during the global financial crisis. This places the countries lower in the ranking, showing a greater capacity to carry a potential bail-out burden. For Hungary, we see the opposite effect. In turn, the strong ties of Poland and Czechia with the Western European countries are visible in the higher ΔCoVaR ranks. Germany shows a lower burden of systemic risk relative to GDP than in absolute terms, especially in the last period. This shifts it downward in the ranking.

The analysis across the three selected periods, the joint analysis of Figures 112 and the rankings all show that there are two different groups of countries whose systemic risk has increased. One group is characterized by a significant growth in the total value of assets in the financial system that is not accompanied by similar growth in GDP, which is likely responsible for the increase in the burden of systemic risk (eg, Slovakia and Bulgaria). In the other group, the role of the contagion channels has increased, perhaps inter alia, due to joining the eurozone within the sample period (eg, the Baltic states). Beyond these specific observations, the rankings also show that SRISK and ΔCoVaR provide different information by identifying the countries with high fragility and those with a greater potential of spillovers in systemic risk, respectively.

6.4 SRISK and ΔCoVaR: regression results

In the final part of the study, we regress the SRISK daily results on ΔCoVaR daily results in order to verify hypotheses (1)–(3) outlined in Section 1. We present the summary results of the regressions in Table 5 along with the confirmation of their statistical relevance (reported methods include maximum likelihood and the Lagrange multiplier). The statistical significance of the results in all cases but one was confirmed by the p-value being below 0.05.

Table 5: Quantile regressions of system-level SRISK on ΔCoVaR for three scenarios (quantiles): 5% best returns case, average returns case and 5% worst returns case. [ML: maximum likelihood. LM: Lagrange multiplier. The R2 statistic depicts the level of similarity between SRISK and ΔCoVaR per country per scenario. The p-value is presented for two tests, ML and LM, and in each case the null hypothesis is H0: β=0 (quantiles: 0.05, 0.5, 0.95).]
Quantile Results Bulgaria Czechia Estonia Germany Hungary Latvia Lithuania Poland Romania Slovakia
q=0.05 R2 statistic (%) 27 37 40 53 24 41 41 24 21 29
  ML p-value < 0.040 < 0.001 < 0.0001 < 0.0001 < 0.033 < 0.0001 < 0.0001 < 0.009 < 0.043 < 0.000
  LM p-value < 0.0001 <0.0001 < 0.030 < 0.0001 < 0.0001 < 0.002 < 0.007 < 0.178 < 0.000 < 0.0001
q=0.5 R2 statistic (%) 33 59 70 70 45 70 70 33 39 49
  ML p-value < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001
  LM p-value < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001
q=0.95 R2 statistic (%) 59 87 95 93 73 95 95 73 69 88
  ML p-value < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001
  LM p-value < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001

The results provide evidence only partially confirming hypothesis (1) – that SRISK, which identifies fragility, and CoVaR, which identifies the potential for risk spillover, provide different information about systemic risk for the studied CEE countries.88 8 Similar characteristics are observable for Germany, but the general level of similarity between SRISK and CoVaR in this case is greater, which is in accordance with the greater dependence of the developed banking sectors on stock market funding.

This hypothesis holds true for the 5% quantile, for which the reported R2 values point to a similarity of only 21–41%. For the 50% quantile, the results still point to significant differences between the risk signals detected by SRISK and CoVaR, but these differences are smaller: the similarity is in the range of 30–77%. In the highest studied quantile (95%), the two measures present almost the same information for smaller CEE systems (R2 of around 90% or more) and slightly less similar information for more complex systems, such as Romania, Bulgaria, Hungary and Poland, with a similarity of about 60–70%. These results are in line with the theoretical expectation that, for smaller systems, more convergence between fragility and contagion is observed during a crisis. This observation confirms hypothesis (2).

Hypothesis (3) is also confirmed by the value of R2 increasing with the the analyzed quantile. Such an increase means that, when the levels of risk are low, SRISK and ΔCoVaR signal different levels of risk for the same entities; but when the risk increases, they peak together. This may be because, during a crisis, the feedback effect between fragility and contagion leads to convergence. This observation, reported in the literature for advanced economies, is now confirmed for the CEE region. Importantly, the low similarity of the results for the two analyzed measures in the low and medium quantile (corresponding to calm periods) shows that macroprudential bodies should use both measures to identify and monitor financial institutions if they want to effectively recognize the weak links in the financial system before a crisis unfolds.

7 Conclusions

This paper presented three proposals to modify two well-established measures of systemic risk, SRISK and ΔCoVaR. Thanks to these modifications we significantly increased the number of analyzed systemically important institutions in the CEE region. We also included information about banks’ systemicness, interconnectedness, complexity, significance and size. Additionally, we integrated the data of broad domestic market indexes into the calculation of ΔCoVaR.

Our methodology takes into account endogenous systemic risk factors, at both the bank level and the system level, as well as exogenous factors such as the exchange rate and the impact of the largest European stock markets. For this reason, the study is more comprehensive than its predecessors with respect to the CEE region.

The empirical results confirmed increased systemic risk for CEE countries in two periods, 2008–9 and 2011–13, at the same time identifying countries more prone to fragility and to contagion. The rankings confirmed the high potential cost of systemic risk materialization in the CEE region. We also showed that, in some cases, this burden was greater for small CEE countries than it was for Germany, which has, by far, the greatest value of systemic risk.

We verified three postulated hypotheses and confirmed that systemic risk in the CEE region has the same theoretical properties as it does for the advanced economies. These properties underline the necessity of analyzing this risk by using measures at least as sophisticated as the ones used by the most developed countries.

The results also show that during periods of calm, SRISK and ΔCoVaR each indicate different financial institutions as the weakest links in the given financial system. These results have a macroprudential consequence: the regulatory bodies in the analyzed countries should consider the information provided by both SRISK and CoVaR when monitoring systemic risk, especially when they want to put banks under scrutiny.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Acknowledgements

We are extremely grateful to special issue Co-Editor Mark Carey and anonymous reviewers for their substantial contributions, because the manuscript was significantly improved with their assistance. This paper forms part of a research project funded by the National Science Centre, Poland, under Agreement UMO-2018/29/N/HS4/02783.

References

  • Acharya, V., and Merrouche, O. (2013). Precautionary hoarding of liquidity and interbank markets: evidence from the subprime crisis. Review of Finance 17(1), 107–160 (https://doi.org/10.1093/rof/rfs022).
  • Acharya, V., Engle, R., and Pierret, D. (2014). Testing macroprudential stress tests: the risk of regulatory risk weights. Journal of Monetary Economics 65, 36–53 (https://doi.org/10.1016/j.jmoneco.2014.04.014).
  • Adams, Z., Füss, R., and Gropp, R. (2014). Spillover effects among financial institutions: a state-dependent sensitivity value-at-risk approach. Journal of Financial and Quantitative Analysis 49(3), 575–598 (https://doi.org/10.1017/S0022109014000325).
  • Adrian, T., and Brunnermeier, M. (2011). CoVaR. Staff Report 348, Federal Reserve Bank of New York (https://doi.org/10.3386/w17454).
  • Adrian, T., and Brunnermeier, M. (2016). CoVaR. American Economic Review 106(7), 1705–1741 (https://doi.org/10.1257/aer.20120555).
  • Andrieş, A., Nistor, S., and Sprincean, N. (2020). The impact of central bank transparency on systemic risk – evidence from Central and Eastern Europe. Research in International Business and Finance 51 (https://doi.org/10.1016/j.ribaf.2018.06.002).
  • Barkauskaite, A., Lakstutiene, A., and Witkowska, J. (2018). Measurement of systemic risk in a common European Union risk-based deposit insurance system: formal necessity or value-adding process? Risks 6(4), article 137 (https://doi.org/10.3390/risks6040137).
  • Benoit, S. (2014). Where is the system? International Economics 138(August), 1–27 (https://doi.org/10.1016/j.inteco.2013.11.004).
  • Benoit, S., Colletaz, G., Hurlin, Ch., and Pérignon, Ch. (2014). A theoretical and empirical comparison of systemic risk measures. Working Paper FIN-2014-1030:42, HEC Paris Research Papers.
  • Benoit, S., Colliard, J.-E., Hurlin, Ch., and Pérignon, Ch. (2017). Where the risks lie: a survey on systemic risk. Review of Finance 21(1), 109–152 (https://doi.org/10.1093/rof/rfw026).
  • Benoit, S., Hurlin, Ch., and Pérignon, Ch. (2019). Pitfalls in systemic-risk scoring. Journal of Financial Intermediation 38(April), 19–44 (https://doi.org/10.1016/j.jfi.2018.05.004).
  • Bisias, D., Flood, M., Lo, A., and Valavanis, S. (2012). A survey of systemic risk analytics. Annual Review of Financial Economics 4(1), 255–296 (https://doi.org/10.1146/annurev-financial-110311-101754).
  • Brownlees, Ch., and Engle, R. (2017). SRISK: a conditional capital shortfall measure of systemic risk. Review of Financial Studies 30(1), 48–79 (https://doi.org/10.1093/rfs/hhw060).
  • Brunnermeier, M. (2008). Bubbles. In New Palgrave Dictionary of Economics, 2nd edn. Palgrave (https://doi.org/10.1057/978-1-349-95121-5_44-2).
  • Constâncio, V. (2010). The establishment of the European systemic risk board: challenges and opportunities. Introductory Remarks, Eurofi Financial Forum. URL: https://bit.ly/3wW759h.
  • De Haas, R., and van Lelyveld, I. (2010). Internal capital markets and lending by multinational bank subsidiaries. Journal of Financial Intermediation 19(1), 1–25 (https://doi.org/10.1016/j.jfi.2009.02.001).
  • Duffie, D. (2019). Prone to fail: the pre-crisis financial system. Journal of Economic Perspectives 33(1), 81–106 (https://doi.org/10.1257/jep.33.1.81).
  • Dumicic, M. (2018). Effectiveness of macroprudential policies in Central and Eastern European countries. Public Sector Economics 42(1), 1–19 (https://doi.org/10.3326/pse.42.1.1).
  • Engle, R., and Ruan, T. (2019). Measuring the probability of a financial crisis. Proceedings of the National Academy of Sciences 116(37), 18 341–18 346 (https://doi.org/10.1073/pnas.1903879116).
  • Engle, R., and Zazzara, C. (2018). Systemic risk in the financial system: capital shortfalls under Brexit, the US elections and the Italian referendum. The Journal of Credit Risk 14(4), 97–120 (https://doi.org/10.21314/JCR.2018.247).
  • Engle, R., Capellini, R., and Reis, B. (2019). Systemic risk analysis summary. Web Page, V-Lab. URL: https://vlab.stern.nyu.edu/welcome/srisk.
  • European Banking Authority (2014a). Guidelines on criteria to assess other systemically important institutions (O-SIIs). Report, European Banking Authority. URL: https://bit.ly/3i3eRaY.
  • European Banking Authority (2014b). Guidelines on O-SIIs assessment EBA/GL/2014/10. Report, European Banking Authority. URL: https://bit.ly/2TXUkwa.
  • European Banking Authority (2019). List of O-SIIs notified to the EBA. Report, European Banking Authority. URL: https://bit.ly/3hDhxNm.
  • European Central Bank (2019). CISS – composite indicator of systemic stress. Web Page, ECB Statistical Data Warehouse. URL: https://bit.ly/3qpHzqk.
  • Gale, D., and Allen, F. (2003). Asset price bubbles and stock market interlinkages. In The Implications for Monetary, Regulatory, and International Policies, Hunter, W., Kaufman, G., and Pomerleano, M. (eds). MIT Press, Cambridge, MA.
  • Gale, D., and Yorulmazer, T. (2013). Liquidity hoarding. Theoretical Economics 8(2), 291–324 (https://doi.org/10.3982/TE1064).
  • Hattori, A., Kikuchi, K., Niwa, F., and Uchida, Y. (2014). A survey of systemic risk measures: methodology and application to the Japanese market. IMES Discussion Paper 14-E-03, Institute for Monetary and Economic Studies, Bank of Japan.
  • Hollo, D., Kremer, M., and Lo Duca, M. (2012). CISS – a composite indicator of systemic stress in the financial system. Working Paper, European Central Bank Working Paper Series (https://doi.org/10.2139/ssrn.1611717).
  • Jajuga, K., Karaś, M., Kuziak, K., and Szczepaniak, W. (2017). Ryzyko systemu finansowego. Report, Materialy i Studia, Volume 329, Narodowy Bank Polski.
  • Jakubík, P., and Slačík, T. (2013). Measuring financial (in)stability in emerging Europe: a new index-based approach. Oesterreichische Nationalbank Financial Stability Report 25, 102–117.
  • Jočiėne, A. (2015). Scandinavian bank subsidiaries in the Baltics: have they all behaved in a similar way? Intellectual Economics 9(1), 43–54 (https://doi.org/10{}.1016/j.intele{}).2015{}.09{}.002.
  • Karaś, M. (2019). Measuring financial system stability: analysis and application for Poland. PhD Thesis, Wroclaw University of Economics.
  • Karaś, M., and Szczepaniak, W. (2017). Measuring systemic risk with CoVaR using a stock market data based approach. In Contemporary Trends and Challenges in Finance, Jajuga, K., Orlowski, L. T., and Staehr, K. (eds), pp. 135–143. Springer Proceedings in Business and Economics. Springer (https://doi.org/10.1007/978-3-319-54885-2_13).
  • Karkowska, R. (2013). Systemic risk in Central and Eastern European banking and its determinants. Merton option model approach. Working Paper, Social Science Research Network (https://doi.org/10.2139/ssrn.2349017).
  • Kubinschi, M., and Barnea, D. (2016). Systemic risk impact on economic growth – the case of the CEE countries. Romanian Journal of Economic Forecasting 16, 79–94.
  • Kuziak, K., and Piontek, K. (2018). Assessment of systemic risk in the Polish banking industry. In Contemporary Trends and Challenges in Finance, Jajuga, K., Locarek-Junge, H., and Orlowski, L. T. (eds), pp. 145–158. Springer Proceedings in Business and Economics. Springer (https://doi.org/10.1007/978-3-319-76228-9_14).
  • Laeven, L., Ratnovski, L., and Tong, H. (2014). Bank size and systemic risk. Staff Discussion Notes of the International Monetary Fund 14(4) (https://doi.org/10.5089/9781484363720.006).
  • Laeven, L., Ratnovski, L., and Tong, H. (2016). Bank size, capital, and systemic risk: some international evidence. Journal of Banking and Finance 69(August), 25–34 (https://doi.org/10.1016/j.jbankfin.2015.06.022).
  • Lietuvos Bankas (2019). Lietuvos Bankas financial stability review. Annual Report, Lietuvos Bankas. URL: https://bit.ly/3j9YwU8.
  • Narodowy Bank Polski (2019). Narodowy Bank Polski financial stability report. Report, June, Narodowy Bank Polski. URL: https://nbp.pl/en/systemfinansowy/fsr201906.pdf.
  • Official Journal of the European Union (2013). Directive 2013/36/EU of the European Parliament and of the Council of 26 June 2013 on access to the activity of credit institutions and the prudential supervision of credit institutions and investment firms, amending Directive 2002/87/EC and repealing Directives 2006/48/EC and 2006/49/EC text with EEA relevance. Official Journal of the European Union 56(L176). URL: http://data.europa.eu/eli/dir/2013/36/oj/eng.
  • Radulescu, M., Logica, B., and Sinisi, C. (2018). Developments of the CEE banking sectors after the financial crisis. Proceedings of the International Conference on Business Excellence 12(1), 851–863 (https://doi.org/10.2478/picbe-2018-0076).
  • Shiller, R. (2003). From efficient markets theory to behavioral finance. Journal of Economic Perspectives 17(1), 83–104 (https://doi.org/10.1257/089533003321164967).
  • Wong, A. Y.-T., and Fong, T. P. W. (2011). Analysing interconnectivity among economies. Emerging Markets Review 12(4), 432–442 (https://doi.org/10.1016/j.ememar.2011{}).06{}.004.

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact info@risk.net or view our subscription options here: http://subscriptions.risk.net/subscribe

You are currently unable to copy this content. Please contact info@risk.net to find out more.

You need to sign in to use this feature. If you don’t have a Risk.net account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here