OCCASIONAL PAPER SERIES. No 21 / 2018 OVERVIEW OF LITHUANIA S BANKING SECTOR SUSTAINABILITY IN THE POST-CRISIS ENVIRONMENT

Transcription

1 BANK OF LITHUANIA. WORKING PAPER SERIES No 1 / 08 SHORT-TERM FORECASTING OF GDP USING LARGE MONTHLY DATASETS: A PSEUDO REAL-TIME FORECAST EVALUATION EXERCISE 1 OVERVIEW OF LITHUANIA S BANKING SECTOR SUSTAINABILITY IN THE POST-CRISIS ENVIRONMENT OCCASIONAL PAPER SERIES No 21 / 18

2 ISSN (ONLINE) OCCASIONAL PAPER SERIES No 21 / 18 OVERVIEW OF LITHUANIA S BANKING SECTOR SUSTAINABILITY IN THE POST-CRISIS ENVIRONMENT By Jokūbas Markevičius Lietuvos bankas, 18 Reproduction for educational and non-commercial purposes is permitted provided that the source is acknowledged. Totorių g. 4 LT Vilnius, Lithuania The views expressed are those of the author(s) and do not necessarily represent those of the Bank of Lithuania.

3 Table of Contents PART I: BANKING SECTOR ENVIRONMENT... 3 POST-CRISIS CHANGES IN MARKET STRUCTURE, CONCENTRATION AND INTERCONNECTEDNESS... 3 POST-CRISIS CHANGES IN RISK APPETITE AND SOLVENCY OF BANKS... PART II: BUSINESS MODELS PROFITABILITY INTEREST PROFIT MARGINS INCOME DIVERSIFICATION... ASSET ALLOCATION LENDING INTEREST RATES IMPAIRMENT COSTS FUNDING COSTS ADMINISTRATIVE COSTS CONCLUDING REMARKS

4 Abstract In the aftermath of the financial crisis, Lithuania s banking sector faced structural changes related to higher concentration, decreased interconnectedness and lower risk appetite. At the same time banks were able to maintain strong profitability levels measured by the EU standards largely due to increased efficiency, very low funding costs, reduced impairments and stable commission income. This paper describes the banking sector of Lithuania in the post-crisis environment and argues that the post-crisis structural changes in general had positive effects on the banking sector s resilience and in the first instance on profitability. In particular, high concentration of the sector was likely to help banks achieve higher efficiency while reduced risk tolerance had direct positive effects on lower impairments as well as indirect effects on lower funding costs. On the other hand, good profitability of the sector has been largely dependent on two largest market participants while smaller banks and branches had less prosperous profitability prospects. In the environment where large banks appear to have particularly good cost management practises, the possibilities for entry of new market players of significant size are plausible only if the newcomers are able to reach the prevailing level of high efficiency. However, without a sizeable market share this might be difficult to achieve. 4

5 GR ES FR NL DE SE IT DK CY PT AT GB SI HU IE BE LV PL MT FI BG SK LU CZ HR RO LT EE PART I: BANKING SECTOR ENVIRONMENT Post-crisis changes in market structure, concentration and interconnectedness The market structure of Lithuania s banking sector has changed significantly since the 1990s as the sector became dominated by several Nordic capital banks. The establishment of the first commercial banks started right after the declaration of Lithuania s independence. The number of local banks sprung up as the lack of adequate regulation meant that banks with no sustainable business model could operate (figure 1). The number of operating banks in Lithuania peaked at 27 in 1993, but very soon was followed by a sharp contraction and the first banking crisis of the country in Following the crisis, banking regulation was tightened making the sector more prone to fraud and more attractive to foreign capital. As a result, a number of foreign banks started entering Lithuania s banking market mostly though the acquisition of local banks (Garbaravičius and Kuodis, 02). Despite a significant increase in market participation by foreign banks, during , the total number of banks remained rather stable varying from 11 to 14. However, as the profitability outlook became more optimistic in the light of rising economic activity and demand for credit, a new wave of banks started entering the market. Between 06 and, the number of banks operating in Lithuania rose from 11 to 18. However, this trend soon reversed after the global financial crisis (that started in 08) hit the profitability of banks. In fact, the crisis resulted in yet another market consolidation and as of Q4 17 there were 12 actively performing banks in Lithuania. Only two (relatively smaller) of those banks were owned by shareholders based in Lithuania. Overall, at the end of 16, foreign banks accounted for almost 92 % of total assets of the banking sector. Although foreign bank dominance is a common feature in the post-soviet countries of Europe, the level of domestic banks market share in Lithuania was particularly low at the end of 16, i.e. the second lowest in the EU (figure 2). 1. Over the years foreign capital became dominant in Lithuania s banking sector Number of banks and foreign bank branches in Lithuania at the beginning of the year Number of banks Bln. EUR 2. Domestics banks occupy a small market share in Lithuania compared with other EU countries Share of domestic bank assets vs. total banking sector assets in the EU countries, Foreign branches Banks Total banks' assets (RHS) Source: ECB CBD2. 5

6 Since 11, concentration of Lithuania s banking sector has increased significantly due to bankruptcies of banks, significant merger and portfolio acquisitions. Since the inception of Lithuania s commercial banking activities, the banking sector concentration has always varied at relatively high levels with only few banks holding the major part of the sectors assets. The Herfildahl-Hirchman index, calculated in terms of total assets of banks, reached its highest levels in 00, when one of the biggest banks Hermis was acquired by the Swedish capital bank SEB (figure 3). Since then, the level of concentration started falling as domestic banks Snoras and Ūkio bankas were able to take up a significant market share from their Scandinavian-based rivals. However, market concentration decline reversed after these two local banks were declared insolvent in 11 and 13 due to fraudulent activities of their management. In 15 and 16, concentration increased even further when Danske bank s retail portfolio was acquired by Swedbank and bank Finasta was taken over by Šiaulių bankas. Finally, in Q4 17, concentration increased again after the merger of DNB bank and Nordea branch in Lithuania which previously were the third and the fourth largest banks in terms of assets. These acquisitions and the merger were motivated by low profitability outlook due to inability of banks to reach sufficient economies of scale. Although profitability of the country s banking sector appears to be strong according to the EU standards, it is rather uneven across banks and strongly depends on the relatively better performance of two largest banks: Swedbank and SEB (see more in Profitability in Part II). Indeed, as of Q4 18, three largest banks in Lithuania accounted for 81 % of total assets of the banking sector forming a strongly oligopolistic market structure. The literature about the impact of concentration on financial stability in different countries provides ambiguous results. Researches that were based on the episode of the 08 financial crisis have shown that for Canada and Australia higher concentration has contributed to greater resilience to the crisis, while in such countries as Switzerland and the Netherlands the relationship was rather opposite (ECD, ). In the case of Lithuania, where before 08 concentration had been particularly high, the crisis outcome was particularly severe. As Scandinavian-based large banks induced credit crunch, non-financial private sector was left with no options of external financing. Therefore, it can be argued that concentration is likely to amplify systemic risk in the periods of turmoil. Judging from the set of key performance indicators, high concentration stands among the major weakness of Lithuania s banking sector (figure 4). 6

7 Since 11, Lithuania s banking sector has become more concentrated Market share of banks in Lithuania in terms of total assets and HHI index HHI index Herfindahl hirschman index (right-hand scale) Note: areas of different colours represent individual banks' market share in terms of total assets on the left hand scale High concentration is the major weakness of Lithuania s banking sector Main indicators of Lithuania s banking sector in terms of ranking across countries in Europe. Q2 17 Concentra tion (HHI index) Liquidity Coverage coverage 25 ratio of ratio NPLs Costincome ratio Tier 1 capital ratio Return on assets Leverage ratio Return on equity Net interest margin NPL Loan-todeposit ratio Source: EBA and Bank of Lithuania. Note: green area indicates leading positions while red area indicates bad relative performance High market concentration can be partially explained by a small size of the market; however, there is a number of EU markets of a similar size that enjoy lower levels of concentration. Positive relationship between the country s economic output and its banking sector concentration is illustrated in figure 5 suggesting that banks are attracted by higher potential business volumes. However, there are many examples where countries with similar GDP have lower banking sector concentration than Lithuania, such as Slovenia, Malta, Croatia or Latvia, indicating that similar-size markets can be served by more banks. Moreover, the indebtedness of Lithuania s private sector at the end of 17 was among the lowest in the EU suggesting that Lithuania s market potential has not been exploited as much as in other Central-Eastern European countries. Therefore, from the macro perspective Lithuania s market still has relatively more space for stronger penetration. Evidence suggests that more concentrated banking systems perform better in terms of efficiency. Although most research finds that lower efficiency in more competitive-concentrated markets is in line with the Queit Life Hypothesis (Repkova and Stavarek, 13), this relationship might not hold if competition and concentration were independent. Indeed, the data of 16 suggest a positive relationship between the cost-toasset ratio and the Herfildahl-Hirchman index among the EU banking sectors (figure 6). One possible explanation of such relationship could be related to economies of scale being reached in more concentrated sectors. For instance, a bank may be in a better position to invest in efficiency (e.g. IT systems) when a market is more concentrated as investment in digitalisation often requires relatively large financial resources which might bring benefits only if the bank s network is large enough or the bank has stable market power. Hence, countries with many small-scale banks, such as Germany or France, face efficiency problems while countries where few banks dominate the market seem to be able to reach higher economies of scale and, 7

8 hence, higher efficiency. Given that this rationale holds, higher concentration does not necessarily infers less competition. 5. High concentration partially reflects small market potential Annual GDP vs. HHI index for the EU countries GDP, bln. Eur Lithuania R² = 0, ,05 0,1 0,15 0,2 0,25 0,3 HHI index, thousands Sources: ECB, Eurostat and Bank of Lithuania calculations. 6. More concentrated markets seem to enjoy better efficiency Cost-to-asset vs. HHI index across the EU countries Cost-to-assets (efficiency), percentage R² = 0, ,05 0,1 0,15 0,2 0,25 0,3 HHI index, thousands Source: ECB and Bank of Lithuania's calculations. Lithuania In Lithuania s credit market lower and less persistent concentration is evident only for small business and consumer loans. More precisely, persistently high concentration is evident in the market of housing loans as well as in the market of relatively large business loans, i.e. larger than EUR 1 million (figure 7). In fact, only in the market of smaller loans (either business or consumer) concentration seems to be substantially lower, allowing for a wider range of market participants. The recovery in credit growth since 15 has increased lending concentration even further for medium size (EUR million) corporate loans and housing loans, as most of the new lending was provided by largest banks. For instance, as of 3Q 17, only three largest banks were actively engaged in the housing loans provision business. In Lithuania s corporate credit market, the level of concentration varies depending on the economic sector. The highest degree of corporate credit concentration is observed in the public sector, IT, utilities and mining economic sectors, as illustrated in figure 8 (branches are excluded from the sample due to limited data). These sectors are mostly dominated by a few large companies and, hence, are typically served only by few large banks that are able to provide big loans with attractive interest rates (also see Interest rates in Part II). On the other hand, corporate segments with the lowest concentration include sectors of accommodation and catering as well as agriculture, where the average loan is relatively smaller and the volume of loans - higher, meaning that these economic sectors can be more easily served by smaller banks. 8

9 7. Concentration level is particularly high in segments of larger loans, while segments of smaller loans are noted for stronger competition Rolling yearly average of the HHI (market concentration) index in different loan market segments in Lithuania HHI index Source: Bank of LIthuania. HHI of < 0,25 mln. loans to NFC HHI of 0,25-0,5 mln. loans to NFC HHI of > 1 mln. loans to NFC HHI of housing loans 8. Concentration level of varies in the corporate credit market depending on economic sectors HHI index for bank corporate loan portfolios across different economic sectors as of Accomodation and catering Agriculture Health Administrative activities Proffesional activity Manufacturing Education RE development Construction Other actitities Water Trade Artistic activity Transportation Mining Utilities IT Public governance Note: branches are excluded HHI index After the financial crisis of 08, financial interconnectedness between banking systems of Lithuania and Nordic countries has weakened significantly. Credit expansion of required significant foreign funding sources as local deposits were exiting the country due to large current account deficits. As a result, banks increased their share of foreign funding to the level accounting for half of their total assets in 08 and most of this funding naturally came from parent banks based in Scandinavia (figure 9). The situation started changing since the outbreak of the financial crisis in 09, because deteriorated financial situation urged banks to deleverage their portfolios (also see Risk appetite and Solvency in Part I) and parent banks started withdrawing their funds from Lithuania. Between 09 and 16, the share of foreign funding in banks has decreased from 46.6 % to 11.2 % and the loan-to-deposit ratio fell from over 0 % to around 0 %. Interconnectedness within Lithuania s banking system has also weakened due to lower activity in the local interbank market. The local interbank market was rather active during the pre-crisis period as well as several years after the crisis (figure ). However, since 12, it practically disappeared as banks shifted to other sources of short-term funding in the background of growing distrust among banks amidst bankruptcies of bank Snoras and Ūkio bankas. On the other hand, foreign branches and subsidiaries continued pursuing short-term interbank funding activities with their parent banks. In fact, they significantly increased their net borrowing from abroad during and reduced it after the introduction of the euro, although this occurrence might have been related to plain liquidity management at the banking group level. As of 17, the net interbank lending to foreign banks was at historically low level. 9

10 9. Dependence on foreign funding has reduced since 08 Funding obtained from other banks as % of total banking sector assets Source: bank of Lithuania. Note: different color areas represent individual banks.. Interbank market activity in Lithuania practically disappeared after the euro introduction Monthly local and international interbank lending volumes Mln. euro Local interbank market lending Net lending to foreign banks Post-crisis changes in risk appetite and solvency of banks Following the financial crisis, Lithuania s banks significantly reduced their asset portfolios. Significant losses experienced during the crisis have urged the Lithuanian banks to cut on new lending leading to a significant reduction of their loan portfolios. Most of the other EU banking sectors avoided such behavior in the first years after the financial crisis, but followed the suit after the outbreak of the Euro zone sovereign debt crisis during (Altavilla, Paries and Nicoletti, 15). Compared to 08, the lowest point of the EU median banking sector s asset value was recorded at 89 % in 16, while for Lithuania s banks at 73 % in 12 (figure 11). Even more evident was the fall in risk-weighted assets which for the EU median banking sector and Lithuanian banks accounted for 73 % and 50 % of the pre-crisis levels, respectively. One of the main reasons for the portfolio reduction of Lithuanian banks was the structural change in their target levels of risk aversion. In 08, the average risk weight of Lithuanian banks stood at 59 % and was significantly above the EU median (46 %); however, in the aftermath of the financial crisis it contracted substantially (i.e. to 34 % in 16) falling behind the EU median of 39 % (figure 12). Such behavioral shift can be viewed as a structural change in the way banks manage their credit risk, as risk weights of the Lithuanian banks continue to be low despite the credit provision recovery which started in 15.

11 11. Following the financial crisis Lithuania s banks significantly reduced their assets Dynamics of bank assets across the EU countries since 08 Index, 08 = Source: ECB CBD2 5th-95th percentile EU median Lithuania 12. Lithuania s banks average risk weights fell sharply after the crisis Ratio of risk-weighted assets to total assets across the EU countries Source: ECB CBD2 5th-95th percentile range Lithuania EU median The fall in the risk appetite of banks was mostly evident in the corporate loan segment due to its relatively higher credit risk. Corporate loans tend to carry higher risk (as well as higher risk weights) due to a typically lower level of collateralization and smaller financial support from the government (e.g. social benefits), as compared to housing loans (also see asset allocation in Part II). In fact, it was corporate loans that caused most of the losses for Lithuania s banking sector during the crisis. As a result, banks cut corporate lending volumes the most (i.e. from EUR 6.7 to 2.2 billion) between 08 and 14 (figure 13). The average risk premia (lending margin) of Lithuanian banks for smaller corporate loans temporarily increased after the crisis, but fell back in 16, supporting the argument that banks have limited their exposure to riskier corporates. Overall, since 08 interest rates on relatively riskier corporate loans as well as small corporate loans moved in line with typically safer housing loans. Although the relative risk premia increased somehow for small corporate loans during 09-14, it fell back to the level of 08 in 16 (figure 14). The stability of the risk premia might indicate lack of banks willingness to engage in lending to higher risk corporate clients for which they would require higher interest rates (Neri, 13). This is in line with the fact that average risk-weights of banks have not grown despite the recovery in credit issuance. 11

12 13. Corporates faced the largest decrease in lending volumes after the financial crisis New lending volumes by types of borrowers (rolling 12-month sums) Mln. Eur Interest risk premia on small corporates temporarily rose after the financial crisis Weighted-average interest rates on new loans by types of borrowers (rolling 12-month averages) Percentage Corporate loans < 1 mln. Eur Corporate loans > 1 mln. Eur Housing loans Source: ECB MIR Corporate loans < 1 mln. Eur Corporate loans > 1 mln. Eur Housing loans Source: ECB MIR Lithuania s banks have strengthened their capital adequacy ratios significantly since 08. During the period of 08-16, the overall capital adequacy ratio of the banking sector increased twofold, i.e. from 12 % to 24 %, before reducing to % in 17. This was in contrast to the pre-crisis period (00-08) when capital adequacy ratios of banks kept declining and stood relatively close to the required minimum just before the onset of the financial crisis. As of Q2 17, largest banks in Lithuania were relatively far from the current minimum of.5 %, although some domestic banks were quite close to the minimum requirement level after Pillar II buffers have been taken into account (figure 15). Much of the banks improvement in capital adequacy resulted from portfolio downsizing and reduction in risk weights. As already indicated in the paragraph above, the nominal value of risk-weighted assets of Lithuania s banks fell almost twice during and had a significant positive impact on solvency levels of banks. In other words, with the same amount of capital banks could now cover twice as large likely losses which boosted their capital adequacy ratios to become among the highest in the EU (figure 16). In addition, the level of bank capital base was sustained by strong and stable levels of profitability, although a significant part of retained profit was paid out to shareholders in the beginning of

13 15. Lithuania s banks have significantly improved their solvency position since 09 Distribution of capital adequacy ratios of Lithuania s banks of different size and the minimum capital adequacy requirement largest domestic banks 3 largest Nordic banks Min. capital adequacy requirement Min. capital adequacy requirement & capital cons. Buffer Source: Bank of Lithuania Notes: only subsidiaries and including banks Snoras and Ukis Much of the improvement in capital adequacy resulted from reduction of risk weights Contributions to annual change in capital adequacy ratio Percentage points Contribution of annual change in risk weighted assets Contribution of annual change in capital Annual change in capital adequacy ratio Capital adequacy ratio (right-hand scale) Šaltinis: Lietuvos banko skaičiavimai Despite post-crisis portfolio downsizing banks in Lithuania appear to be more leveraged than their peers in Eastern Europe, although significantly less compared to the EU average. The leverage ratio (or equity-to-assets ratio) of Lithuania s banks stood at 8.6 % in Q1 17, i.e. was significantly above the EU average of 5.3 % (figure 17). It is true though that the EU average is largely affected by the Western European banking sectors which hold relatively lower levels of capital as historically they have experienced significantly lower impairment losses due to lower economic volatility. On the other hand, Lithuania s banks appear to be more leveraged compared to banking sectors of other Central-Eastern European countries. It can be viewed as a direct consequence of reduced average portfolio risk weight suggesting that Lithuania s banks carry lower credit risk for the same nominal value of assets. Distribution of individual banks solvency rates suggests that Nordic-Baltic banks are among the healthiest in Europe, although data gaps add some comparison limitations. As of 16, the highest share of banks with solvency ratios close to or even below the minimum requirements was among the Southern European countries that had been dealing with the legacy issues of non-performing loans (ESRB, 17). For instance, nearly two thirds of banks in Italy had solvency ratios below 16 % (figure 18). On the contrary, banking sectors in Lithuania and in other Nordic-Baltic countries (i.e. Sweden, Finland, Latvia) had virtually no banks with alerting solvency levels, showing little contagion risk and stability of financial systems of the Nordic-Baltic Region overall. 13

14 EE HR GR BG PL RO CY IE SI LV LT HU SK MT PT FI NO AT LU CZ ES GB IT BE EU FR DE SE NL DK 17. Leverage of Lithuania s banks is below the EU average Leverage (equity/asset) ratio across EU countries, Q Source: EBA. 18. Solvency position of Lithuania s banks is good compared to other EU banks Solvency ratio distribution in terms of the number of banks falling within respective intervals, Source: CBD2. IT SI GR PL AT DE PT ROHR BE MT LU SK CY DK CZ LV SE FI NL LT Solvency ratio < 8 % Solvency ratio 8-12 % Solvency ratio % Solvency ratio 16- % Solvency ratio > % NA Overall, the structural change in risk tolerance levels of banks has positively influenced their solvency and profitability, but might have negatively affected the country s economic growth potential. From a profitability perspective, lower risk appetite decreased impairment costs of banks and might have lowered their funding costs (also see Part II). The rationale behind the latter relationship is rather non-trivial: lower credit provision decreased Lithuania s imports and helped balance its current account, meaning that a higher share of fiat money stayed in the economy (Wray, 12), allowing banks to reduce its deposit interest rates relatively more than in other EU countries (also see funding in Part II). On the other hand, reduced credit provision might have had a negative macroeconomic effect, as credit rationing might have slowed the economic recovery after the crisis by limiting potential investment and particularly the growth of smallmedium enterprises (Armeanu at al., 15). Nevertheless, from a financial stability point of view reduction of banks risk appetite made the financial system less prone to systemic risks. 14

15 PART II: BUSINESS MODELS Analysis of profitability of Lithuania s banks and its determining factors is key for the assessment of sustainability of banks business models. Although banks business models can be assessed through a variety of angles, the long term profitability can be considered a core measure of sustainability of banks as it sums up the overall performance of banks. Given that Lithuania s banking sector consists of long-standing players with rather stable market shares, the current profitability levels are close to the long-term levels (i.e. profitability is not distorted by short-term actions, such as aggressive take up of market share). The following assessment of banks business models requires decomposition of profitability into its main components. From the income side, the most notable components are interest income, commission income and trading income. From the expense side, the most significant components include funding costs, impairment costs and administrative costs. Further analysis takes a closer look into the interest income and all three expense components in an attempt to explain the good profitability performance of Lithuania s banking sector relative to the median EU country (figure 19). 19. Bank business model structure Profitability Income Expenses Interest income Commission income Trading income Funding costs Impairment Administrati ve costs i) Asset composition ii) Interest rates iii) Portfolio size i) desposit interest rates; ii) share of deposit funding i) NPL level ii) Impairment flows iii)collaeralized loans share i) IT costs ii) personel costs Profitability Profitability of Lithuania s banking sector has stood above the EU median in the aftermath of the financial crisis of Overall, the sector s profitability has remained strong despite a significant reduction in the loan portfolio of banks and the period of low interest rates. As of 16, return on assets of Lithuania s banking sector stood at 0.98 % while the median return at the EU level was 0.42 % (figure ). On one hand, such difference might be related to the upswing of the economic cycle which was also ongoing in other highly profitable Eastern and Central European banking sectors. If this is the case, currently high profitability might be vulnerable to deceleration of economic activity once the economic upswing slows down. On the other hand, high profitability might also be a result of structural factors, such as increased 15

16 EE LV BG HU HR CZ RO LT SI SK MT IE PL SE BE DK AT LU FI NL FR ES GB DE PT CY IT GR efficiency and better risk management which could in turn prove the stability of Lithuania s banks business models. Good profitability of Lithuania s banking sector has been largely influenced by two largest banks, while profitability levels of smaller banks and branches have been rather subdued. On one hand, the share of highly profitable banks in Lithuania is relatively high compared to the EU standards: around 90 % of Lithuania s banks (excluding branches) had return on equity ratio (RoE) higher than 5 %, i.e. above the EU average, in 16. It is true though that many EU countries were experiencing profitability challenges related to legacy issues of non-performing loans, environment of low interest rates and low economic growth (ECB, 17). On the other hand, if we include branches and consider other profitability metrics, such as the ratio of return on asset (RoA) or return on risk-weighted assets (RoRWA), profitability distribution of Lithuania s banks becomes less proportional. For instance, in 14-17, six out of nine banks with a market share above 1 % had an average RoA below the sector s mean (figure 21). In fact, the average was mainly outperformed by two largest banks (Swedbank and SEB) and Šiaulių bankas driving the whole sector s profitability upwards, while other smaller banks had notably weaker profitability prospects.. Lithuania s banking sector profitability was above the EU median in 16 Percentage 2,0 1,5 1,0 0,5 0,0-0,5 Return on asset ratio for banking sectors across the EU 21. Lithuania s banking sector profitability is largely influenced by two largest banks Return on assets and return on risk-weighted-assets of individual banks in Lithuania as the average. Percentage ,0-1,5 Source: ECB CBD2. RoA RoRWA Note: RoRWA defined as return on risk-weighted-assets. Banks' names are shortened for visual purposes. Profit of Lithuania s banks is strong relative to risks they assume, measured by the EU standards. The ratio of net profit to risk-weighted assets (RoRWA) of Lithuania s banking sector has been above the EU median, since 11 (figure 22). It is worth saying that such profitability indicator allows for a more accurate comparison of profitability of different banks as opposed to the regular RoE or RoA indicators. It is superior to RoA as it includes the asset-related risk and it is superior to RoE as it does not take into account a portion of capital held above the requirements. This extra capital usually represents accumulated profit set aside for the distribution to the shareholders or is held to strengthen banks solvency reputation. Therefore, given its 16

17 sensitivity to the bank risks, RoRWA suggests that for the same level of profit Lithuania s banks are exposed to significantly lower risks than majority of the EU countries and, hence, need less capital to cover those risks. It could also be said that banks in Lithuania would have one of the highest RoE ratios in the EU if all banks reduced their capital adequacy ratios down to the required minimum. Compared to Nordic parent institutions, Lithuania s banks achieve better profitability relative to the assumed risks. In terms of the ratio of net profit to risk-weighted assets the largest banks in Lithuania have largely outperformed their Scandinavian parent institutions since 11 (figure 23), although it was not the case with the DNB bank which consequently merged with Nordea in an attempt to boost profitability. Overall, the analysis suggests that Lithuanian subsidiaries of SEB and Swedbank were generating higher riskweighted returns than their parent institutions. This is in contrast to the paper by Jociene (15) which argues that Baltic subsidiaries largely underperform their parent banks in terms of profitability as measured by RoE. As indicated in the paragraph above, RoE has its flaws, especially for subsidiaries of large banks. First, profit in subsidiaries is typically retained for a certain time period before it is paid out, hence, artificially lowering RoE. Second, even after dividend pay-outs, capital held in these subsidiaries is considerably above the requirements compared to parent banks, possibly due to reputational reasons or portfolio growth expectations. Overall, RoRWA measure supports the fact that the Lithuanian market is indeed successful for subsidiaries of SEB and Swedbank while not so much in the case of DNB. 22. Return on risk-weighted assets has exceeded the EU median since Return on risk-weighted assets in the EU countries 23. Return on risk-weighted assets is higher in Lithuania s banks than in their parent institutions Return on risk-weighted assets of SEB, Swedbank and DNB in Lithuania vs. parent banks th-95th percentile range EU median Lithuania Source: CBD Source: SNL finance and the Bank of Lithuania. Range across Lithuania Range across group Volatility of net profit of Lithuania s banks decreased following the financial crisis suggesting higher matureness of the sector. High volatility of the profit of banks observed in has stabilized after the financial crisis mostly due to stabilization of interest income and impairment costs (figure 24). From 17

18 13 onwards, the profit of banks has been historically stable and strong with 16 being the third most profitable year since the inception of commercial banking in Lithuania. The helicopter view suggests that profitability of Lithuania s banks was higher than the EU median in 16, mostly due to relatively lower interest expenses, impairment costs and administrative costs. Expressing the main banks income and expense items as a percentage of total assets provides a tool for bank performance comparison across individual EU countries. Plotting Lithuania s banking sector indicators against the EU median suggests that Lithuania s banks tend to outperform in terms of lower funding, impairment and administrative costs, but significantly underperform in terms of interest income (figure 25). In other words, low interest income appears to be outweighed by good cost management Profitability of the banking sector has been historically stable and strong Mln. Eur 00 0 Lithuanian banks net profit and its main components Net result of other operations Impairment Depreciation Administrative costs Trading income Net commission income Net interest income Profit 25. Low interest expenses, administrative costs and impairment costs are the main factors contributing to good profitability Bank income and expense indicators as a ratio to total assets. Lithuania s banking sector vs. the EU median % of total assets 3,0 2,5 2,0 1,5 1,0 0,5 0,0-0,5-1,0 Beter than EU median Interest exp. -1,5-1,50-1,00-0,50 0,00 0,50 1,00 1,50 2,00 2,50 3,00 Source: ECB CBD2 Impairment Admin. exp. Return Trading & FX results Commission income Interest income Worse than EU median Interest profit margins Lithuanian banks interest profit margin (interest income over costs not covered with other types of income) has been above the EU average. To maintain profitability banks have to cover the costs associated with their activities, such as administrative expenses, deposit interest and impairment costs. A part of administrative costs can be covered by non-lending-related income such as commission or trading income, since such costs could be considered as inputs for generation of non-lending income. On the other hand, the uncovered part of administrative costs, together with deposit and impairment costs must be compensated by charging interest rates on loans. The difference between banks interest income and the aforementioned uncovered costs divided by total assets can then be viewed as interest profit margin of banks, as illustrated in figure 26. Estimated in such a way, in 16, profit margin on lending of Lithuania s banks was above the EU average, largely because of successful management of administration, interest and impairment costs. 18

19 Interest profit margins explain most of the EU banks profitability variation. There is a strong positive relationship between the estimated interest profit margins of the EU banks and their return-on-assets ratios. In fact, single linear regression explains 71 % of such relationship, indicating that 5 items (i.e. interest income, non-interest income, administrative costs, interest expenses and impairment) can indeed be focused on in assessing sustainability of banks business models (figure 27). 26. Lending rates and cost management have significant influence on interest profit margins of banks Profit margins, estimated using 4 components, across individual EU banking sectors 6,0 27. Estimated interest profit margins explain profitability dispersion of the EU banks with 70 % accuracy Profit margins, estimated using 4 components, vs. RoA across individual EU banking sectors RoA, percentages 2,0 5,0 4,0 3,0 2,0 1,0 0,0-1,0 1,5 1,0 0,5 0,0-0,5 Lithuania -2,0 HUBG LV HRROSK SI CZ EE PL LT IE ES AT MTBE SE NL FR DK LUGBDEGRCY FI IT PT Profit margin Inpairment Interest expenses Administrative exp. not covered by non-interest income Source: Bank of Lithuania -1,0 R² = 0,5299-1, Profit margin, percentages Source: Bank of Lithuania Interest profit margins on loans of individual banks of Lithuania are rather different suggesting that for some banks higher costs are not equally compensated by higher lending interest rates. The estimated interest profit margins are quite wide for some banks and rather narrow for others (figure 28). For some banks estimated interest profit margins are reduced by inability to cover administrative expenses with non-interest income (e.g. due to the lack of efficiency or economies of scale) as well as higher deposit costs. In fact, it appears that some banks might be affected by a vicious circle, i.e. weak cost management by some banks forces them to lift lending rates which consequently may lead to higher risk tolerance (as ability of banks to compete for lower-risk clients reduces) and, hence, higher impairment-related losses. Household loans appear to carry higher net profit than corporate loans for largest banks when administrative expenses are distributed equally. Corporate loans appear to have lower interest rates and carry higher impairment costs compared to household loans which are dominated by well-collateralized housing loans. As a result, they appear to generate higher net profit for banks with largest loan portfolios (figure 29). On the other hand, household loans might carry higher administrative costs given that the number of household loans is typically significantly larger than that of corporate loans. However, data 19

20 limitations do not allow for accurate splitting of administrative costs for the respective corporate and household loans Bank interest profit margin on loans depends on their offered interest rates and cost management Bank lending rate components based on average data of Profit margin New loan impairments Deposit interest costs Administrative exp. not covered by other income Lending rate 29. Bank profit is a result of interest profit margins and lending volumes Bank interest profit margins on loans vs. net product profit: corporate loans (green circles) & household loans (yellow circles). The size of the circle represents the portfolio size. Profit margin, in percentages 7,0 6,0 5,0 4,0 3,0 2,0 1, Bank 1 Bank 2 Bank 3 Bank 4 Bank 5 Bank 6 Source: Bank of Lithuania 0,0-1,0 0,0,0,0 60,0 80,0 0,0 Product profit, Source: Bank of Lithuania millions Eur Income diversification Net interest income has historically been the major source of bank income. This indeed reflects the fact that business models of Lithuania s banks rely on traditional banking operations, i.e. lending. Although the ratio of net interest income to commission income has varied since 08, net interest income has never fallen below the amount of other income sources (figure ). As of 17, the ratio of net interest income to commission income was 198 %. This in turn means that a great deal of importance in maintaining strong profitability comes from the ability of banks to control interest income and interest expenses. Historically, commission income has proved to be the most stable source of bank income serving as a cushion in the low-interest-rate environment. Contrary to the literature that often suggests commission income to be the least reliable source of income during the period of economic downturns (DeYoung, 03), commission income has shown little volatility during the crisis years for Lithuania s banks. As interest income shrank due to deleveraging and low interest rate environment, commission income remained virtually unchanged during (figure 31). This indeed provides banks with some cushion in the period of stress and could be viewed as a risk mitigating factor.

21 . Net interest income is the major source of bank income 31. Commission income has proved to be the most stable income source Mln. Eur Main sources of banks income Gross interest and commission income in the banking sector Millions of EUR Net interest income Net commission income Trading gains Interest income Commission income Bank income is well diversified in terms of the EU standards. Interest income accounts for the largest share of total operating income of Lithuania s banks, although a significant share could be attributed to commission income (figure 32). As of Q1 17, net interest income accounted for 61.0 % of the total operating income, while net commission income made up 27.3 %. In general, compared to other EU countries, the income structure of Lithuania s banking sector can be viewed as well diversified, helping to reduce the risk of the prolonged period of low interest rate environment (ECB, 16). Larger banks tend to diversify away from interest income more compared to smaller banks. The simplified analysis shows that there is a weak relationship between income diversification (interest vs. commission income) and bank size (figure 33). This suggests that larger banks might be in a better position to use their market power to engage in fee-based business than smaller banks. In fact, by exploiting their market power, larger banks might be reducing fee-based business opportunities for smaller banks, hence, leading to even higher market concentration. On the other hand, higher income diversification can be viewed as a natural target for larger banks in order to mitigate the risk of lower interest income during economic downturns. From this perspective, higher risk diversification by large banks might also bring benefits of financial stability. 21

22 MT GR CY SK NL BG ES CZ BE EE AT HR HU DK PL SI PT LT DE RO IE SE GB LV IT FR FI LU 32. Bank income is well diversified in terms of the EU standards. Procentai 0% 90% 80% 70% 60% 50% % % % % 0% -% Composition of bank operating income Interest Commission Trading Other Source: ECB CBD2 data. 33. Larger banks tend to diversify income more than smaller banks Proportion of individual banks net interest income to net commission income versus respective banks total assets Net interest income vs. net commission income, ratio R² = 0, Total assets, Bln. EUR Asset allocation About two thirds of total assets in Lithuania s banking sector comprise traditional bank assets, i.e. loans to households and non-financial corporations. In fact, most of other EU banks tend to direct a higher proportion of their assets towards market instruments (e.g. debt securities) or other types of assets that include e.g. loans to government, other credit institutions and other financial institutions (figure 34). As in the most EU countries, the corporate loan portfolio of Lithuania s banks is mostly directed to real estate activities, manufacturing and trade sectors. These three economic sectors are dominant in the EU in terms of provision of corporate loans of banks, followed by construction and transport sectors (figure 35). In fact, Lithuania s banks stand approximately in a median position among the EU countries in terms of bank corporate portfolio concentration measured by the HHI index (i.e. degree of diversification of economic sectors). This in turn indicates that the corporate loan portfolio of Lithuania s banks is rather well diversified in terms of the EU standards. Generally, banks in Lithuania significantly rely on the real estate market. Household housing loans and corporate loans to real estate development and construction companies accounted for 54 % of total nonfinancial private sector lending portfolio of Lithuania s banks in Q3 17. In other words, the real estate market provides the main business opportunity for banks in Lithuania and, hence, is the main risk factor. It could be noted that since the real estate-related loans are usually covered with a high degree of collateralization, banks in Lithuania are relatively better able to reduce their idiosyncratic losses and the riskweighted-assets in general. On the other hand, they are more prone to systemic risks related to imbalances in the real estate market. 22

23 PT ES DK LT SK IE NL PL HR FI SE GR BE EE CY RO CZ SI FR IT BG LV HU AT MT DE LU DK SE FI EE DE LV AT IE HU LT NL FR CY SK IT PL LU BE PT ES MT BG HR RO GR SI 34. Lithuania s banks follow traditional banking models Composition of bank assets in the EU Member States Corporate loan portfolio of Lithuania s banks is mostly directed to RE activities Composition corporate loan portfolio of banks in the EU Member States Household loans Corporate loans Equity Debt securities Derivatives Cash Other Source: ECB CBD2. RE activities Trade Transportation Other services Agriculture Administrative activities Source: ECB CBD2. Manufacturing Construction Electricity Professional activities Accomodation and food IT Lithuania s banks are exposed to the level of risk that is close to the EU median suggesting a relatively low risk appetite compared to other Central-Eastern European countries. The average risk weight for banks operating in Lithuania stood at 36 % in Q1 17. Although this risk weight was significantly higher than in the Nordic countries (such as Sweden or Finland), it was also considerably lower than in most of the Central-Eastern European countries (figure 36). On one hand, this reflects the fact that more than a half of the loan portfolio of banks is related to the real estate sector which typically carries high-quality collateral. However, this also suggests that banks in Lithuania are engaged in relatively low risk lending. There is a strong positive relationship between risk weights and net lending margins across the EU countries suggesting that higher risk profile banking sectors are fairly compensated with higher risk premiums. Although due to data gaps the analysis does not cover all EU countries, a sample of 16 EU banking sectors seems to be sufficient to capture the aforementioned relationship (figure 37). A simple linear regression line allows comparing banking sectors that are relatively more/less compensated for risks they bear. In fact, the results show that banks in Lithuania are very close to such line indicating that they are relatively fairly compensated for the risks they bear. 23

24 36. Portfolio risks of Lithuania s banks is close to the EU median 37. Higher risk-taking is compensated with larger lending margins 50 Risk weighted assets to total assets, Average risk weights against lending margins in the EU countries Average risk weight, percentages R² = 0, LT HR PL PT CY SK LT EE CZ FR DK NL SE Source: ECB CBD Source: ECB CBD2 and MIR. Lending margin, percentages There is evidence that Lithuania s banks are rather risk-averse in lending to small and medium-sized enterprises, although these data are limited only to the IRB banks. As of Q2 17, the share of loans to small and medium-sized enterprises accounted for 41.8 % of total corporate loans. However, due to the lack of comparable data it is difficult to adequately compare exposures of banks to small and medium-sized enterprises across the EU countries. Still, one possible way is to compare the exposures of banks that use Internal Risk Based models. Measured this way, the share of bank loans to small and medium-sized enterprises accounted of 28 % of total corporate loans of the IRB banks during the same period, i.e. was close to the median of the EU countries (figure 38). These results are rather non-trivial as the share of small and medium-sized enterprises in the economy is usually relatively higher in the Central-Eastern European countries. Indeed, this might suggest that Lithuania s banks are rather conservative in lending to smaller companies, in other words more risk averse. Small and medium-sized enterprises carry relatively higher risk weights compared to other loan types, as reported by the IRB banks. Higher risk weights not only reflect higher default risk of SMEs but are also more costly for banks in terms of capital allocation. As of Q3 17, the average risk weight in respect of SMEs for Lithuania s banks using IRB approach was 59 %, while average risk weights for large corporates and retail loans stood at 45 % and 31 %, respectively (figure 39). 24