1
How do banks react when things go rough? Bank lending behaviour
in Lithuania during the 2008-2009 crisis
Nerijus Černiauskas
Supervised by Dr. Margherita Saraceno
Helpful guidance and comments provided by Dr. Gabriele Galati and Dr. Sander Onderstal
Abstract
The thesis investigates the lending behaviour of foreign and domestic banks in Lithuania during the crisis of 2008-2009 against the background of changes in Lithuania’s financial system. An analysis based on macroeconomic data and evidence from a survey conducted by Bank of Lithuania reveals that a mixture of falling demand and tighter lending conditions led to reduced bank lending in Lithuania during the crisis. Using data on the balance sheets of banks in Lithuania provided by the Association of Lithuanian Banks, the lending behaviour of different types of foreign and domestic banks is compared. It is found that although foreign banks were in many respects different from domestic banks in Lithuania, they reduced total lending by a similar proportion during the crisis. It is also found that the distribution of loans changed between banks. Foreign banks reduced the fraction of housing loans to a lower extent than domestic banks. Swedish sister banks (the largest foreign banks in Lithuania) were found to curtail lending to households proportionally less than other foreign banks.
Master Thesis in Economics with a specialization in public economic policy in University of Amsterdam
2
Contents
1. Introduction ... 3
2. Review of the literature ... 5
3. Economic background and trends in Lithuanian financial sector ... 8
Data choice ... 8
Economic context ... 9
The financial system ... 14
Banks and financial institutions ... 23
4. Analysis of the banking behaviour in Lithuania ... 25
Data and grouping ... 25
Cross-section analysis ... 27
Time series analysis ... 34
A statistical assessment of differences in lending behaviour ... 40
5. Discussion ... 47
6. Conclusions ... 49
3
1. Introduction
With rising share of bank assets held by foreign banks in Eastern and Central Europe and developing countries there has been an interest in studying foreign bank behaviour and the effects they have on the banking sector. The 2008-2009 crisis reignited interest in foreign bank behaviour during periods of financial turmoil. Many studies used cross country data to examine bank lending during the 2008-2009 crisis and previous crises (Molyneux and Seth, 1998; Moshirian, 2001; Jeon and Miller, 2006; Choi, Gtierrez and Peria, 2013; Peek & Rosengren, 2000; Crystal, Dages & Goldberg, 2006). These studies find that depending on the region and time frame chosen the lending behaviour of foreign banks sometimes differs from that of domestic banks and sometimes not. The current study investigates domestic and foreign bank lending behaviour within a single country - Lithuania. Such approach allows showing the wider context which affects lending behaviour, as well as analysing variables that may be unavailable in cross country studies - in this case the distribution of loans. The fact that during the transition to an open market-oriented economy Lithuania’s banking system witnessed expansion and dominance of foreign banks makes it an interesting case to analyze: Lithuania is one of few countries where foreign banks occupy over 80% of bank assets. Additionally, Lithuania’s banking system is highly concentrated as two Swedish sister banks control 60% of bank assets.
This paper analyzes the lending behavior of foreign-owned and domestic banks in Lithuania during the recent global crisis against the background of changes in Lithuania’s financial system. Special attention is devoted to Swedish sister banks in Lithuania (SSBL) as they are the largest banks by market share in Lithuania and have received the greatest criticism in popular media for contraction of credit and thus exacerbating the economic downfall (Diena.LT, 2009; 15min.LT, 2010; Jockus, 2012; Elita, 2012). An analysis based on macroeconomic data and evidence from a survey conducted by Bank of Lithuania reveals the following. Surprisingly, I do not observe significant differences in total lending between foreign and domestic banks. However, the distributions of loans did change significantly between foreign and domestic banks: all foreign banks tended to retain a higher share of issuing housing loans as compared to domestic banks. Therefore within current research, the null hypothesis is that the growth of new loans was uniform across the three types of banks (SSBL, other foreign banks and domestic banks). My analysis shows that SSBL reduced lending to households proportionally less than other foreign banks.
My thesis is structured in a following way: Chapter 2 reflects on existing literature that characterizes the development of Lithuania’s financial system before the external shock. Special emphasis is
4 placed on the development of SSBL as the largest and most influential banks in Lithuania. It also looks at the experience of other foreign banks in host countries and shows that parent banks in Europe were more willing to aid their sister banks in the EU than in other more distant regions, yet domestic banks tended to reduce lending by less than foreign banks. Chapter 3 briefly looks at the macroeconomic and financial situation in Lithuania from 1999 till 2012 and compares it with the one in other European countries to provide the context during which bank behaviour changed. It concludes that Lithuania, together with other Baltic countries, experienced stronger and possibly excessive economic and financial growth prior to the crisis and also experienced a greater shock than Western European states such as Sweden or Germany. This gives evidence that banks faced difficult macroeconomic conditions and tightened lending conditions. Chapter 4 analyses the lending behaviour of SSBL following the initial crisis shocks in 2008-2009 in Lithuania using data obtained from Association of Banks which took its data from balance sheets and income statements of banks in Lithuania. Chapter 5 reviews the main findings and gives recommendations for further studies. Chapter 6 concludes.
5
2. Review of the literature
In this chapter I present evidence taken from the literature that could explain bank lending behaviour in Lithuania during the shock. However, as Lithuania’s banking system was still transitioning to a modern banking system, I first review the literature covering this development from 1990 to 2008 to better understand the state of Lithuania’s financial system in the onset of the crisis. Special attention is paid to foreign banks and their role in the development of Lithuania’s financial system. Next, I review Lithuanian and international literature that focuses on the empirical relationship of parent banks and their sister banks during shocks and recessions with brief reference to growth periods. Then, I summarize the insights on the changes in lending behaviour of banks from 2008 to 2012 in Lithuania. The end of the section summarizes the findings of the chapter and offers a link to further stage of research.
An overview of the financial system development shows that Lithuania had a relatively modern banking system by the time the financial crisis occurred. Čičinskas and Šadžius (2006) describe Lithuania’s banking systems transition from a Soviet mono bank system in 1988 to a modern banking system by 2004. The authors separate the banking history into three stages: (1) nationalization of existing Soviet banks, liberalization which allowed new banks to be established, and early legal framework creation (1989-1991), (2) a period marked by intense legal synchronization with EU standards and the failure of inefficient banks (1991-1997) and (3) the following process of foreign capital and expertise arrival to Lithuania which allowed previously nationalized banks to be privatised. Vilpišauskas (1999) presents evidence that Lithuania’s banking system, especially in legal terms, was being integrated with those that existed in the European Union (EU) while Čiapas (1999) concentrates on the evolution of capital requirements in Lithuania and notes that the first instances of risk weighted-capital were already applied in 1995 after the signing of EU accession treaty. Garbaravičius and Kuodis (2002) present evidence that less efficient banks were exiting the market while new products were being implemented into the traditional banking model (such as life insurance and leasing companies). Povilaitis (2002) looks at the reestablishment of the Bank of Lithuania – Lithuania’s central bank (LCB) - as the main issuer of national currency (litas) and a supervisor of the banking system. Ramonas (2002) looks at the modernizing role of the LCB in providing supervision for commercial banks in Lithuania by adapting new technologies. Leika (2008) finds out that the LCB plays a minimalist role. It means that the LCB can only prepare and form expectation in the market, but has, due to its fixed exchange rate policy, limited tools to tackle large macroeconomic swings. Additionally, detailed analysis of the Lithuanian financial system is provided by the LCB in its two yearly publications (Annual Report launched in 1994 and its Financial Stability Review which started in 2006). They cover the entire financial system and about half of them are
6 dedicated to the banking system and factors that could affect its stability. Factors include macroeconomic variables of Lithuania, the household sector, non-financial corporate sector, and the balance sheets of banks.
Some observers have stated that foreign banks and SSBL in particular, played an important and to some extent unique role in Lithuania as compared to domestic banks. Kuodis and Garbaravičius (2002) indicate that SSBL banks were larger than domestic banks in terms of assets. The authors also note a strong link that was established between parent and sister banks which enhanced the liquidity of Lithuania’s banking system, in addition to increasing quality, competition, transparency and efficiency of financial intermediation. The stability enhancement is supplemented by interviews conducted by Volvonis (2004) who shows that credit risk was controlled at the parent bank level in addition to that of the sister bank thereby gaining additional overall security. Yearly Financial Stability Reviews conducted by CBL show that SSBL are often the most profitable financial institutions (CBL, 2007). Additionally, liquidity and capital requirements are controlled at the parent bank level which allows the sister banks to receive help during turbulent times in the host country (CBL, 2008). Furthermore, the reviews reveal that parent banks had higher internationally acclaimed credit ratings which allowed attaining market funding for a lower cost which was subsequently transferred to sister banks in Lithuania. As a result, sister banks could offer lending at a lower interest rate to borrowers in Lithuania, which meant either higher profitability or being able to accept lower risk borrowers (Leika, 2008: 79). At the same time, Ramanauskas (2005: 92) reveals that the presence of foreign banks reduces the effectiveness of regulation aimed at slowing down credit, especially at times of high lending interest rates and profitability in the host country.
Foreign banks have a history of behaving in some cases similarly and some cases differently from domestic banks before and during episodes of financial distress. In general, any bank behaviour can be pro-cyclical as the supply and demand of loans falls with falling national income (Bernanke and Blinder, 1988). During times of stress banks that operate in international markets have been found to sometimes curtail their operations in the host market (Molyneux and Seth, 1998; Moshirian, 2001; Jeon and Miller, 2006; Choi, Gtierrez and Peria, 2013) although this has not been applied so much in Europe (De Haas and Van Lelyveld, 2006; Choi, Gtierrez and Peria, 2013). Other authors have found that banks in developing countries within South America tend to use the opportunity provided by a crisis to expand their operations (Peek & Rosengren, 2000; Crystal, Dages & Goldberg, 2006). An empirical study using Bank of International Settlements statistics by Hermann and Mihaljek (2010) confirms a reluctance of parent banks to pull out from their sister operations in
7 Europe1 during the 2008-2009 financial crisis if compared to other continents, and suggests that a stronger EU integration can explain these changes. At the same time, De Haas and Van Lelyveld (2011) and De Haas De Haas, et al (2011) in their panel regression using BankScope database show that bank lending has reduced for sister banks (subsidiaries) more than for domestic banks globally and in Europe. Using Dealogic Loan Analytics database, De Haas and Van Horen (2011) also find that bank lending was less reduced in countries where banks had substantial information about those countries. One notable within country analysis is conducted by Puri, Rocholl, and Steffen (2011) who use the difference-in-difference (DID) approach along with a unique German dataset which allows differentiating between supply and demand effects. It has been found that banks in Germany which were exposed to US financial problems tended to give permission for fewer loans as opposed to those banks which were not affected.
The current global financial crisis was also marked by a reduction of lending in Lithuania. All banks in Lithuania suffered losses in 2009 which were greater than profits gained during the economic boom years of 2004-2008 and the outstanding value of loans was reduced by 2.8 billion euro or 14% in 2009 as compared to 2008 (LCB, 2013). Although not a single foreign bank stopped their operations, foreign banks, especially Swedish sister banks, were accused by businesses and business associations for curtailing lending, thereby exacerbating the economic downfall (Diena.LT, 2009; 15min.LT, 2010; Jockus, 2012; Elita, 2012). Arguably, a mixture of supply and demand shocks contributed to this decline. A yearly questionnaire of bank experts that first appeared in 2006 in Lithuania organized by the LCB revealed a clear tightening of lending conditions and a more conservative appetite for risk by the banks, although fewer loans were also handed out due to reduced demand. An investigation into the failure of Lithuanian banks by the LCB as well as an audit by the Swedish National Audit Office (2011) have shown that Lithuanian domestic banks and foreign banks in Lithuania were taking up risks (e.x. liquidity and credit risks) prior to the financial crisis, which could have contributed to supply side problems later on. During the crisis, there was also an observed withdrawal of credit from banks in Lithuania to banks in other countries (Ramanauskas, 2011). This could be explained by lower demand for loans driven by lower economic activity in those countries (LCB, 2009; Ramanauskas, 2011). At the same time, the subsequent repayment of debt to foreign banks has been explained by desire to limit exposure of international banks in host countries (supply side explanation) (Vienna Initiative, 2013). Even if it was a supply side shock that caused lending to be reduced, it seems unlikely to be driven primarily by developments in Sweden. In fact, Financial
8 Stability reports in Sweden have repeatedly provided evidence that Sweden has experienced much milder shocks compared to Lithuania (Sveriges Riksbank, 2012, 2013).
I summarized evidence that Lithuanian banking system was already quite developed when the crisis hit. Foreign banks were seen to increase the efficiency of the system. I find that literature partly explains the reduction in lending during the crisis in Lithuania by a mixture of falling demand and tighter lending conditions but I have not found academic literature in Lithuania discussing differences between banks as a causal factor of shrinking credit. International literature is divided on the actions that foreign banks take during shocks in host countries: some curtail credit, while other see shocks as an opportunity to expand at the expense of other banks. In order to better understand factors that affected lending behaviour of banks in Lithuania I present aggregate statistical evidence on Lithuania’s economic and financial system before and during the shock.
3. Economic background and trends in Lithuanian financial
sector
In this chapter I describe the context in which the crisis took place and review indicators related to changes in bank lending behaviour. This chapter is set up as follows. First, I present the data sources and the countries I used for comparison. A country comparison helps to identify to what extent Lithuania’s economic system differs from other countries. Then, I look at several economic variables and studies that describe the evolution of the banking system in Lithuania. Main findings are summarised in the end of the chapter.
Data choice
I took data from three databases: Eurostat, Statistics Lithuania and from the LCB. Eurostat allows comparing the Lithuanian context to other countries while the LCB and Statistics Lithuania’s databases are more detailed but do not provide statistics on other countries. International comparison is necessary, as it helps to determine the significance of certain measures, such as the level of financial institution’s asset to gross domestic product, which do not have much value on their own.
I compare Lithuania’s variables to five other countries: Sweden, Germany, Poland, Estonia and Latvia. Data on Sweden is presented to show the situation in which the Swedish parent banks operated. Germany is chosen as the largest state within the European Union which also represents “Western European” trends. Poland is chosen as the largest country within “Eastern Europe”. Other Baltic states are chosen to show that these countries have in general experienced similar trends to
9 Lithuania and the change in lending is a regional phenomenon. The majority of the facts covered in this section will focus on the time period of 1999-2012 as the data is more relevant and comparable. As such, the crisis and hyperinflation immediately after the fall of the Soviet Union, the subsequent stabilization and the effects on Lithuania of 1998-1999 Russian crisis which is already well described by Kuodis (2009) are not considered.
Economic context
Lithuania is a relatively small economy in Europe. The yearly Lithuanian economic output (Gross Domestic Product, GDP) was valued at 32.9 billion euro in 2012, which made it the largest of the three Baltic States. Latvia’s GDP of the same year was valued at 22.3 billion euro while Estonia’s GDP has reached 17.4 billion euro. The GDP of Germany equalled to 2,472 billion euro. Sweden’s GDP was 407.7 billion euro.
In terms of GDP growth, the time period of 1999 to 2007 was a dynamic period for all the selected countries but this was especially true for the Baltic countries. All three Baltic countries experienced strong growth (an average of 8% per year as compared to 2% for Germany, 3% in Sweden and 4% in Poland) from 1999 to 2007 (Table 1). It was a period of growing wages, especially in the non-tradable sector (sector that produces goods and services for consumption at a local level, such as housing, public sector services), rather than in the manufacturing sector (Kuodis and Ramanauskas, 2009). Kuodis and Ramanauskas (2009) link this growth with the high credit growth in Lithuania in the period. This was followed by a recession in 2008-2009. In Lithuania, the severe stage of the recession was felt in 2009 when GDP was reduced by 15% as compared to 2008 while unemployment rose during the same period from 5% to 14% (Eurostat, 2013). Moreover, GDP fell by similar per cent in rural and urban areas. A decomposition of GDP growth reveals a fall in domestic demand (Kuodis and Ramanauskas, 2009). The Baltic countries returned to growth in 2010 (except for Latvia which was still in a recession in 2010).
Table 1: Real GDP growth rates of selected EU countries
1999-2007 2008 2009 2010-2012 Germany 2% 1% -5% 3% Poland 4% 5% 2% 3% Estonia 8% -4% -14% 5% Latvia 8% 3% -15% 4% Lithuania 8% 3% -15% 4% Sweden 3% -1% -5% 3%
10 The size of GDP in Lithuania is reflected by a small and a decreasing population and a lower GDP per capita than Germany or Sweden. After the census in 2010, Lithuania had just over 3.1 million people which are estimated to have dropped to 3.0million in 2012 (Table 2). This makes Lithuania about 27 times smaller than Germany and about 3 times smaller than Sweden. Lithuania has lost over half a million people since 1999, a sixth of its population, largely due to migration. In 2012, Latvia in terms of population had 2.0million, while Estonia – 1.3million. Due to a fast real and nominal growth as well as decreasing population, Lithuania’s GDP per capita has increased 3 times since 1999 and in 2012 reached 35% of GDP per capita of Germany. Since Lithuania’s prices are still below the EU average, GDP per capita in purchasing power parity (PPP) would be even higher.
Table 2: Population, GDP and GDP per capita for selected EU countries
1999 2006 2009 2012 Lithuania GDP (billion euro) 10.3 15.1 21.0 32.4 Population (mln.) 3.5 3.3 3.2 3.0 GDP per capita 2,910.2 4,599.9 6,586.1 10,791.7 Sweden GDP (billion euro) 242.8 266.7 298.4 333.3 Population (mln.) 8.9 9.0 9.3 9.5 GDP per capita 27,421.1 29,481.3 32,232.3 35,143.0 Germany GDP (billion euro) 2,095.2 2,407.9 2,284.0 2,471.8 Population (mln.) 82.0 82.2 82.0 81.8 GDP per capita 25,539.2 29,287.0 27,853.0 30,200.9 Data source: Eurostat. Calculations by the author
Table 3 contains selected indicators at quarterly frequency: GDP per capita, wages and salaries per employed as well as a number of people employed in Lithuania. GDP and wages are deflated by the consumer price index and are seasonally and working day adjusted. The seasonal and working day adjustment is made by Statistics Department of Lithuania. Table 3 describes selected economic indicators where index of 100 represents the values of the above mentioned variables on the first quarter of 2008. The number of employees started falling in the third quarter of 2008, yet GDP per capita and wages were reduced and GDP fell only in the first quarter of 2009. Output in the construction sector was falling in the third quarter of 2008, which coincided with rising unemployment and followed by retail, export and industry (Černiauskas and Černiauskas, 2010).
Table 3: GDP and labour indicators for Lithuania, index (100 = first quarter of 2008)
2008Q1 2008Q2 2008Q3 2008Q4 2009Q1 2009Q2 2009Q3 2009Q4
GDP per capita 100 103 103 103 82 82 80 83
Wages and salaries
per employed 100 103 108 112 90 92 96 98
Number employed 100 100 99 96 94 90 87 84
Unemployment rate
(%) 3.6 3.8 5.9 7.9 10.4 12.9 14.7 16.4
11 Inflation in Lithuania and the rest of the Baltics was pro-cyclical, as the inflation trends were similar to those of GDP. During 1999-2007 inflation in Lithuania was similar to that of Germany or Sweden. Inflation was lower in Lithuania than other Baltic States as a result of falling of its import energy prices from 2001 to 2003 while energy imports were more stable in other Baltic States (IMF, 2005). However, inflation started to converge to the higher Baltic levels in 2006 and in 2008, just before the crisis hit, inflation surpassed 10% in Lithuania. Inflation was especially high in the real estate sector (European Commission, 2010). In 2009 the Baltics experienced deflation before returning to inflation levels that are closer to EU average in 2010-2012.
Figure 1: Harmonized consumer price index of selected EU countries (1999 =100)
Data source: Eurostat
Table 4: Harmonized consumer price index average changes for the selected EU countries (years 1999-2012)
1999-2007 2008 2009 2010-2012 Germany 2% 3% 0% 2% Estonia 4% 11% 0% 4% Latvia 5% 15% 3% 2% Lithuania 2% 11% 4% 3% Poland 3% 4% 4% 3% Sweden 2% 3% 2% 1%
Data source: Eurostat
As the economy grew the current account deficit grew also. In 2008, the net current account balance reached 25% deficit of Lithuania’s GDP which compares to a deficit of 7% in Poland and surpluses of 6% in Germany and 9% in Sweden (Figure 2). This deficit in Lithuania and other Baltic States existed
90 110 130 150 170 190 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Germany Estonia Latvia
12 from 1999 to 2009 the values of both (in absolute values) f exports and imports (which show the openness of Lithuania’s economy) grew from 88% in 1999 to 165% in 2012, according to Eurostat data. Both imports and exports reduced in Lithuania in the fourth quarter of 2008, but the largest fall was observed in first quarter of 2009, when imports and exports were reduced by 28% and 10% respectively (Table 5). This brought Lithuania along with other Baltics countries at a current account surplus. Table 5 shows that exports started to grow again in the third quarter of 2009. Even though exports recovered quickly imports were slower to recover, which meant that current account was more stable also.
Figure 2: Net Current account balance of selected EU countries (% GDP)
Data source: Eurostat
Table 5: Seasonally adjusted quarterly real imports and real exports in Lithuania (millions Euros)
2008Q1 2008Q2 2008Q3 2008Q4 2009Q1 2009Q2 2009Q3 2009Q4 Real export 3,332 3,334 3,368 3,272 2,912 2,713 2,945 3,153 Real import 4,633 4,385 4,304 4,285 3,112 3,129 3,173 3,416 Data source: Statistics Lithuania
To finance the deficit a high financial account surplus was necessary which meant that Lithuania had on net to borrow funds from abroad. However, due to a growing economy Lithuania continued receiving credit (Rodzko, 2005). Funding of the deficit in Lithuania came mainly from private borrowing (LCB Report 2007). The public finances of the Baltic States were quite stable prior to the crisis which allowed the Baltic States to escape large government fiscal imbalances. This can be inferred from consolidated government debt of the Baltic States to GDP ratio which was well below
-30% -25% -20% -15% -10% -5% 0% 5% 10% 15% 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
13 Maastricht’s 60% limit (Table 6). The Baltic Governments even managed to reach a positive primary balance ((government revenue minus expenditure) in 2005 (Table 7). However, the fall in domestic consumption and employment meant that tax revenue declined substantially in all three states and the primary balance in 2009 recorded a deficit of 8.2% in Lithuania and 8.3% in Latvia. Even though Lithuania did not seek aid from the International Monetary Fund as Latvia did, this did create mistrust in the state’s ability to control its finances. The mistrust in Lithuania’s ability to repay its obligations can be inferred from a rise in interest rates on 1 year government bonds. Those rose to over 8% as fears mounted of government collapse or devaluation of the national currency. The government responded by reducing government expenditure - and by safeguarding the litas-euro peg, as well as stabilizing the primary deficit. It managed to proceed without sizable protests or riots (Åslund, 2011).
Table 6: Consolidated Government Debt of selected EU countries (% GDP, years 1999-2012)
1999 2005 2008 2009 2012 Germany 61.3 68.6 66.8 74.5 81.0 Estonia 6.5 4.6 4.5 7.1 9.8 Latvia 12.5 12.5 19.8 36.9 40.6 Lithuania 22.7 18.3 15.5 29.3 40.5 Poland 39.6 47.1 47.1 50.9 55.6 Sweden 64.3 50.4 38.8 42.6 38.2
Data source: Eurostat
Table 7: Primary government balance of selected EU countries (% GDP, years 1999-2012)
1999 2005 2008 2009 2012 Germany 1.6 -0.5 2.7 -0.4 2.5 Estonia -3.2 1.8 -2.7 -1.8 -0.1 Latvia -3.2 0.1 -3.6 -8.3 0.0 Lithuania -1.4 0.3 -2.6 -8.2 -1.4 Poland 0.0 -1.3 -1.5 -4.8 -1.1 Sweden 4.8 3.8 3.8 0.2 0.5
14
The financial system
The macroeconomic trends occurred in parallel to changes in the domestic financial system. Financial depth as measured by the share of assets of financial corporations to GDP has increased in the Baltic countries, Poland, Sweden and Germany. Commercial banks make up the largest share of financial corporation’s (which include central banks and financial auxiliaries) which makes this is a good estimate for the size of banks as well. The graph below illustrates that the size of Baltic banking systems was much smaller compared to that in Sweden and Germany. This led to a belief that the rapid growth in bank balance sheets and credit in the Baltic States was part of a process of catching up with Germany and Sweden. Hence, it was argued that it was not needed to intervene in the process despite rapid credit expansion (Leika, 2008). Since 1999 Lithuania’s financial corporation’s assets to GDP ratio increased by 217% by 2009 while the financial corporations’ assets within euro area as a whole increased by 60%. In 2010 and 2011 however, the financial corporations’ asset to GDP ratio went into decline in all compared countries. In 2012 lending has started to rebound in Sweden, Germany, Poland and Estonia. Seeing credit growth returning, in 2013 the authors of Financial Stability Report of Sweden became worried about another possible credit bubble and therefore encouraged implementing an extra capital buffer to reduce the chance of and the damage caused by growing credit. Credit did not start rising in Lithuania during 2012, however. Part of the decline of credit in Lithuania can be explained by the failure of 2 Lithuanian domestic banks (Snoras and Ūkio bank) in 2011 and 2012 respectively. The failure of the banks created limited losses and did not spark a greater mistrust of the whole financial system.
Figure 3: Assets of financial corporations (% GDP)
Source: Eurostat 0 100 200 300 400 500 600 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Germany Estonia Latvia
15 Although the financial system in Lithuania consisted of many financial corporations, banks posed a dominant position in terms of assets (LCB website, 2013; Garbaravičius and Kuodis, 2002). As the economy improved and the integration with the euro area proceeded, there was a large rise in assets to GDP ratio of the whole financial sector and in banks in particular. There was a rise of other financial corporations, such as leasing and insurance companies as well as private pension funds which were only developing in 1990s. However, Figure 4 indicates that these institutions in 2007 had four times less assets than banks. Similarly, banks played a greater role than financial markets even though the later were rapidly developing. In 1993 financial markets (where financial markets are defined as the market capitalization of listed equity and obligations in the Vilnius stock exchange) were almost non-existent (Garbaravičius and Kuodis, 2002) and in their peak year (2005) banks had only 1.6 times more assets than financial markets. However, by 2006 the majority of larger companies have already listed themselves, while the slowdown and the eventual crash in 2008 left the equity market weak (Securities Commission of Lithuania, 2008).
Despite the developments in the banking sector up to 2008, the banking situation worsened towards the end of the year and during 2009. Banks assets were reduced by 6.1% in 2009 in Lithuania. Accounts of banks that are published on the LCB’s website show an observed liquidity shortage and a rise in non – performing loans (defined as the sum of non-impaired loans overdue more than 60 days and impaired loans) where business loans and consumer purchase loans were affected heavier than the mortgage loans (25%, 16% and 7% respectively of all loans in the category by 2010). With a high value of impaired loans, in 2009 the return on assets fell to 4% and the return on equity fell to -50% for an average bank.
Figure 4: Financial structure by type of financial institution (Assets as % GDP)
Source: Garbaravičius and Kuodis (2002), LCB 2013 26 29 31 33 38 46 62 71 82 80 92 85 74 65 3 4 4 6 8 10 12 16 20 17 19 18 16 16 14 17 14 14 23 31 39 36 27 11 16 20 14 14 0 20 40 60 80 100 120 140 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Banks Other Equity market capitalization
16 The Lithuanian banking system is characterized by heavy influence of foreign banks. In 1997 the first foreign branch opened up in Lithuania. At the time, there were 12 domestic banks in Lithuania and mentioned pieces of legislation (law on commercial banks and the CBL) were changed to be similar to those of the European Union (Čičinskas and Šadžius, 2006). Since then, more foreign branches started settling in and Lithuania’s largest banks were privatized, and acquired by foreign – largely Swedish – owners (Čičinskas and Šadžius, 2006) which acquired the majority of banking assets (Égert and MacDonald, 2009). In 2009, 85% of assets held in banks in Lithuania were foreign owned and over 60% of them belonged to parent companies whose headquarters are in Sweden. By 2012, 92% of all assets held within banks in Lithuania were stationed in foreign owned banks, as reported by the Bank Association of Lithuania.
The Lithuanian banks differ from that of Sweden in the structure of assets and liabilities2. Data from 2012 balance sheets show that Lithuania’s assets are much more heavily concentrated in loans, while currency and deposits make up 81% of all liabilities. Sweden is different in it relies somewhat less on loans and that 42% of its liabilities come from securities, much of which comes form international markets. Reliance on markets for funding was been seen as a possible threat by Riksbank, especially during the years of the crisis, but the threat has so far has not materialized Sveriges Riksbank (2012). Next, I take a closer look into the assets and liabilities of Lithuanian banks.
2 I was not able to compare Lithuania with EU27 or Germany due to missing data. However, it was possible to look at the breakdown of Germany’s balance sheet when banks were grouped together with assets of the central bank and the result
17
Figure 5: Distribution of assets and liabilities3 of banks in Lithuania and in Sweden in 2012
Source: Eurostat. Calculations by the author
As shown in Figure 5, loans make up the largest share of bank’s assets. The dynamics of loans issued to households and non-financial corporations were different in Lithuania. The LCB provides statistics on the value of new4 loans issued and their interest rates. Figure 6 below shows the value of new loans lent to firms and households. The majority of household loans consisted of mortgages while the majority of lending to firms was under 1 year in duration (short term loans). The value of new loans lent to firms increased from 2004 until 2009 while lending to households started stabilizing by mid 2007 and started decreasing in early 2008. This reflected growing problems in the housing
3
Eurostat uses the same list of labels for all assets and liabilities. For example, loans can be characterised as asset (if a bank provides a loan to a customer) or loans can be characterized as liabilities (if one bank takes out a loan from another bank).
4 New loans refer to loans on the asset side of the balance sheets of banks that are issued in the month of observation. This is in contrast with outstanding loans which are total loans recorded on the asset side of the balance sheets of banks at a point in time (i.e. includes loans that have been issued in earlier months).
17% 8% 72% 3% 0%
Lithuania's banks
(assets)
Currency and deposits Securities other than shares Loans Shares and other equity Other accounts receivable/paya ble 18% 12% 63% 5% 2%Sweden bank's (assets)
81% 2% 0% 12% 5%
Lithuania's banks
(liabilities)
Currency and deposits Securities other than shares Loans Shares and other equity Other accounts receivable/paya ble 42% 42% 0% 10% 6%Sweden's banks
(liabilities)
18 sector. These problems were spotted by independent authors such as Kasputytė-Šarkauskienė (2005) who found out that banks were gradually accumulating more risk as the liquidity ratio reduced and loans to deposits ratio increased. Ramanauskas (2005) showed that bank credit started to increase rapidly and warned that Lithuania could be heading for a soft landing. Šimkus and Mendelevičius (2006) carried out a loan portfolio analysis and saw a concentration of loans within the real estate sector. Leika and Valeitinaitė (2007) saw that banks, facing competition for clients, gave housing loans at more favourable conditions (such as larger share of a loan that could be backed by the value of the house and lower interest rates charged) which also contributed to rising real estate market.
Figure 6: Value of new loans granted to firms and households (billions of litas)
Data source: LCB
Some of the changes in loans can be explained by interest rates charged for the two types of loans. The interest rate charged to households has increased at a faster rate than the interest charged to non-financial corporations from December 2005 until December 2012 as shown in Figure 7. As a result the spread between the interest rates charged to households and the interest rate charged to firms increased from 3.95% to 9.29%. Figure 7 also reveals that the main refinancing operations rate (which is the rate the European Central Bank uses to target the medium term interest rate levels and conduct monetary policy) can only partly explain the rise in interest before and the reduction in interest after the crisis.
0 0.5 1 1.5 2 2.5 3 3.5 4 1 0 /2 0 04 0 2 /2 0 05 06 /2 00 5 1 0 /2 0 05 0 2 /2 0 06 0 6 /2 0 06 1 0 /2 0 06 0 2 /2 0 07 0 6 /2 0 07 1 0 /2 0 07 0 2 /2 0 08 0 6 /2 0 08 10 /2 00 8 0 2 /2 0 09 0 6 /2 0 09 1 0 /2 0 09 0 2 /2 0 10 0 6 /2 0 10 1 0 /2 0 10 0 2 /2 0 11 0 6 /2 0 11 10 /2 01 1 02 /2 01 2 0 6 /2 0 12 1 0 /2 0 12
19
Figure 7: Nominal interest rates charged by banks to firms and households and the European Central Bank’s main refinancing operation rate (%)
Data source: LCB
As SSBL borrow intensely from their parent banks and parent banks are affected by the economic condition in their home countries, the parent banks may transmit these effects to SSBL. Swedish parent banks undertook operations, including open market operations, with the Swedish central bank – the Riksbank. While Riksbank can set its own policy rates – including the repo rate that Riksbank uses to target inflation levels – it has often been close to the ECB MRO rate (see Figure 8). Even if this factor, along with the rest of the factors that identified within the transmission mechanism literature (Bernanke and Gertler, 1995), could have affected SSBL, data on the interest rates paid on loans coming from parent banks is not publicly available for this study.
Figure 8: Monetary policy rates of ECB (MRO) and the Riksbank (repo) , (%)
Source: Riksbank, ECB 0 5 10 15 20 10 /2 00 4 02 /2 00 5 06 /2 00 5 10 /2 00 5 02 /2 00 6 06 /2 00 6 10 /2 00 6 02 /2 00 7 06 /2 00 7 10 /2 00 7 02 /2 00 8 06 /2 00 8 10 /2 00 8 02 /2 00 9 06 /2 00 9 10 /2 00 9 02 /2 01 0 06 /2 01 0 10 /2 01 0 02 /2 01 1 06 /2 01 1 10 /2 01 1 02 /2 01 2 06 /2 01 2 10 /2 01 2
Interest on loans to nonfinancial corproations Interest rates charged to households
ECB MRO 0 1 2 3 4 5 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Repo ECB MRO
20 Additional information on lending behaviour was taken from LCB’s questioners. Experts from different banks in Lithuania were asked whether the credit terms and demand for credit changed as compared to previous six months. Experts were asked to express their opinion about two markets: households and nonfinancial corporations. The answers to the questions are presented in Figure 9. The values of 0 to 100 indicate that the demand for loans has increased as compared to the previous six months while any number below zero illustrates falling demand. Similarly, 0 to 100 indicate a tightening of lending conditions by the banks as compared to the previous six months while any number below zero illustrates a loosening of lending conditions. The red colour represents the demand and credit conditions towards household’s intent to buy a house (HH) while the blue colour corresponds to the behaviour towards non-financial corporation (NC). As can be seen in Figure 9, credit conditions were tightened for both HH and NC from mid 2006 to 2010. In 2011 the credit conditions eased but were tightened in 2012 once again. The demand for credit kept increasing until 2008, although at a falling rate. Notice, that demand for household credit was the first to fall in early 2008 while lending to non-financial corporations’ was reduced in late 2008.
Figure 9: Survey of bank experts on credit demand and lending conditions (see above for interpretation of the y axis)
Data Source: LCB survey. In the survey a respondent could have wrote a number between -100 and 100. The values of 0 to 100 indicate that the demand for loans has increased as compared to the previous six months while any number below zero illustrates falling demand. Similarly, 0 to 100 indicate a tightening of lending conditions by the banks as compared to the previous six months while any number below zero illustrates a loosening of lending conditions.
Data is available for new deposits from 2004. Since Lithuania carried out operations in euro and in litas, banks had to charge two different interest rates. The interest rate in Figure 10 was calculated as the weighted average interest rate (interest paid on deposits in a given currency multiplied by the
-100 -75 -50 -25 0 25 50 75 100 125 2006 04 2006 10 2007 04 2007 10 2008 04 2008 10 2009 04 2009 10 2010 04 2010 10 2011 04 2011 10 2012 04 2012 10 2013 04 2013 10 HH Demand HH credit terms NC Demand NC credit terms
21 share of loans in that currency in the currency basket). Notice that the below deposits only account for fixed-term (not on demand) deposits made by nonfinancial corporations and households that are residents in Lithuania. The value of new deposits and the interest rate charged for placing deposits started to increase only in mid 2006, after that lending growth had increased and kept increasing until early 2009 with a spike of interest rates in late 2008. Since banks increase deposit rate when they wish to get more deposits, a rise in interest rates could indicate a liquidity shortage..
Figure 10: Value of new fixed-term deposits held within financial intermediaries (banks) other than the Central Bank of Lithuania in billions of euro (left hand side) and weighted average interest paid to resident households and non-financial corporations (right hand side)
Data source: LCB. Calculations by the author
While the deposits described in Figure 10 made up a large share of liabilities, the source of remaining part of liabilities can be detected looking at outstanding assets and liabilities (see Figure 11). This data series was not presented initially, as it does not contain information on interest rate charged on loans, but it does have a longer time span. LCB provides evidence that there was a constant gradual increase in outstanding deposits from 1999 to 2003. From 2004 onwards, which coincided with Lithuania’s membership of the EuropeanUnion, there was an acceleration of resident deposits being accepted by banks. Also, there was a rise in a new type of funding – external liabilities. These were mainly long term (over 3 years) deposits of foreign parent banks in sister banks in Lithuania (as explained in the Financial Stability Review 2013 by LCB) which is in stark contrast to the short term terminated deposits and on-demand deposits (deposits that could be withdrawn upon request without losses) provided by residents. Both type of liabilities increased rapidly until 2008. Since then, external liabilities were dropping while deposits were returning to growth. Also, there was an increase in capital and reserves within banks.
0 2 4 6 8 0 0.5 1 1.5 2 2.5 2004 -10 2005 -01 2005 -04 2005 -07 2005 -10 2006 -01 2006 -04 2006 -07 2006 -10 2007 -01 2007 -04 2007 -07 2007 -10 2008 -01 2008 -04 2008 -07 2008 -10 2009 -01 2009 -04 2009 -07 2009 -10 Dep o si t in te res t ra te s B ill io n s o f Eu ro
22
Figure 11: Outstanding value of liabilities held in monetary financial institutions in Lithuania other than the central bank (billions of euros)
Data source: LCB
A look at the financial account of Lithuania also shows this fact. There was a high net inflow of euros under the category of “Other investments”. By the definition set by the International Monetary Fund (IMF) adopted by the European Commission, “Other investments” are largely currency and deposits that enter Lithuania from abroad (European Commission, 2013). The net size of these flows has increased from less than 1 billion euro (already 3% of Lithuania’s GDP) to 3.5 billion euros in 2007, before turning negative in 2009 and stabilizing after the crisis. Importantly, the change of net other investments were attributed to a fall in liabilities. This means that banks in Lithuania were returning their existed debts to foreign institutions.
Figure 12: Net assets of Lithuania (billions of euros)
Data source: Eurostat 0 2 4 6 8 10 12 14 16 1999 -01 1999 -08 2000 -03 2000 -10 2001 -05 2001 -12 2002 -07 2003 -02 2003 -09 2004 -04 2004 -11 2005 -06 2006 -01 2006 -08 2007 -03 2007 -10 2008 -05 2008 -12 2009 -07 2010 -02 2010 -09 2011 -04 2011 -11 2012 -06 2013 -01
Rezident deposits Capital and Reserves External liabilities Other liabilities
-4 -3 -2 -1 0 1 2 3 4 5
Financial account, Direct investment
Financial account, Portfolio investment
Financial account, Other investment
Financial account, Official reserve assets
23
Banks and financial institutions
Lithuania has a modern two tiered banking system with an independent (Povilaitis, 1998; 2002) central bank established in 1990. The LCB is the issuer of the national currency litas and the supervisor of the Lithuanian banking system (Povilaitis, 2002). LCB conducts reviews based on presented bank documents (external evaluation) and yearly internal evaluations of bank performance using the CAMEL methodology, collects statistical data (Povilaitis, 2002) and since 1997 requested mandatory yearly independent audit from every bank (Ramonas, 2002) (even though the independence of the auditors is questioned by the author). Since 2001 LCB recommended commercial banks to use stress test to look at micro (bank) level while LCB uses stress tests at a macro level (Ramonas, 2002). By 2004 Lithuania had its own credit registry, although historic data was insufficient to put the technology to estimate portfolio risks yet (Kamienas, 2004).
LCB’s main policy objective was price stability and there was a tendency of alignments of the LCB’s policy and tools to those of the euro area. In order to hold prices steady in the long term, the LCB fixed the exchange rate of the national currency litas to the dollar (from 1994 till 2002) and then to the euro till the time of writing (LCB, 2013). The pegging of the currency has helped to stabilize inflation from triple digit growth to single digit inflation by 1998 and also meant that the foreign currency inflows were the only parameter which could affect the change in the value of litas (Povilaitis, 2002). In 1999 the board of LCB publicly announced that they would convert euro to litas in unlimited amounts. Furthermore, the LCB was allowed to exercise modern open market operations and provided overnight lending to commercial banks as done by central banks in the euro area (Povilaitis, 2002) as long as the value of money absorbed was greater than the money placed in circulation (in order to not infringe the currency convertibility principle). By 2001 the LCB could no longer lend to the Government of the Republic of Lithuania and had to focus on price stability as its primary goal by retaining a fixed exchange rate (Povilaitis, 2002). Lithuania joined the European Exchange Rate Mechanism I (ERM I) in 2002 and ERM II in 2004 which meant that the 1 euro would be worth 3.45280 litas (ECB press release, 2004) and was on its way to fully integrate into the euro area. As of 2014, Lithuania still does not have the euro. Even though ERM mechanism allows litas to fluctuate 15% from the fixed euro rate, the LCB has established a currency board which is a special version of a hard currency peg (Alonso, 2002) and which prevented any fluctuation (Piffer, 2012). The currency board is a monetary authority that allows a country to explicitly target the exchange rate by assuring “convertibility of reserve money on demand at a fixed rate and the full backing of reserve money by international reserves” (Alonso, 2002: 5). Short term interest rates are set by the European Central Bank and then applied to Lithuania as well. This means that Lithuania’s monetary
24 policy is restricted in this respect. However, evidence suggests that the interest rates set by the European Central Bank were not able to affect Lithuania’s market rates and hence lending rates efficiently (Ramanauskas, 2011: 10).
In sum, though Lithuania’s economy and financial system were still lagging behind Germany and Sweden, certain financial infrastructure (Bank’s supervision tools) was quite well developed which provided a number of options for monitoring the financial system.For instance, the aggregate bank balance sheet data lets me conclude that Lithuania was highly reliant on external funding from parent banks. The questioner given to bank managers helps me reason that banks were providing fewer loans due to both reduced borrower demand and tighter lending conditions set by banks. However, neither source helps me see if foreign banks acted differently from domestic banks. I do not find evidence that problems in the home country (Sweden) were greater than those in the Baltics which would require fast withdrawal of funds to help the home market. Yet, there is evidence that Swedish households and firms were continuing to take out loans at a faster pace than those in Lithuania which would encourage foreign banks to redirect their lending to Sweden. The next chapter compares domestic banks to foreign banks to see if foreign banks reacted to the crisis differently.
25
4. Analysis of the banking behaviour in Lithuania
In this chapter I analyse bank lending behaviour by looking at bank level data. This section is structured as follows. First, I present data sources and explain how I grouped the data. This is followed by a cross sectional analysis to show how foreign banks and especially SSBL differed from domestic before the crisis. Time series analysis looks at lending trends between domestic and foreign banks when the crisis hit. Then, I conduct several statistical assessments of differences in lending behaviour between domestic and foreign banks. Each analytical subchapter starts by summarizing key results.
Data and grouping
This section deals with data taken from individual bank and income statement balance sheets in Lithuania from 2007 to 2012. The data are provided from theAssociation of Lithuanian Banks (ALB). Data on balance sheets are available on a monthly basis while data on income statements are available on a quarterly basis. Table 8 lists the banks for which data is available:
Table 8: Banks for which data are available from the ALB
Bayerische Hypo-und Vereinsbank AG Vilniaus Skyrius. Since 2007: AS UniCredit Bank Lietuania Nordea Bank Finland Plc Lietuvos skyrius
AB Sampo bankas. Since 2007: Danske Bank A/S Lietuvos filialas AB bankas „Hansabankas“. Since 2009: „Swedbank“, AB
AB DnB NORD bankas
AB Parex bankas. Since 2010: AB "Citadele" bankas AB SEB Vilniaus bankas. Since 2007: AB SEB Bankas UAB Medicinos bankas
AB “Šiaulių” bankas
AB “Ūkio” bankas. Ended operations in 2012 AB bankas “Snoras”. Ended operations in 2011 Source: ALB
The banks shown in Table 8 have different legal status. Some of them have changed their names only (Swedbank and AB SEB Vilniaus bankas), while Bayerische Hypo-und Vereinsbank AG and AB Sampo bankas underwent M&A and two banks (AB Ūkio bankas and AB bankas „Snoras“) have ended their operations. For example, AB SEB Vilniaus bankas, AB Parex bankas changed its name to AB SEB Bankas in 2007, while Ūkio bankas ended its operations in 2012. Table 9 organizes the official bank titles by segments (short name, type, whether it is listed on the stock exchange (AB) or not (UAB) and its shareholder) based on the state of the bank in the fourth quarter of 2007 – a year preceding the crisis. This is a simplification which allows avoiding issues associated with changing bank names or bank types during the period 2007-2012.
26
Table 9: Breakdown of official bank titles by name, type, join stock company and its shareholders
Short
name Type
Join stock
Company Shareholders
Unicredit Cross border
Nordea Branch Nordea (Finland)
Danske Bank to
branch Listed Danske Bank (Denmark)
Hansabank
Bank Listed
Hansabank (Estonia) and Hansabank (Sweden)
DnB Nord Bank Listed DnB (Norway)
Parex Bank Listed Parex banka (Latvia)
SEB Bank Listed Skandinaviska Enskilda Banken (Sweden)
Medicinos Bank Not listed Lithuanian shareholders
Šiaulių Bank Listed EBRD (20%), Lithuanian shareholders
Ūkio Bank Listed Lithuanian shareholders
Snoras Bank Listed Lithuanian shareholders
Source: ALB
In Lithuania a “bank” can be a commercial bank (bank), a foreign bank branch (branch), or banks operating in the Republic of Lithuania without a physical branch operating in the country (cross border). It will be assumed that branches are similar to banks but different from cross border banks, even though they all offer lending services. Unicredit will be removed from the static and dynamic analyses. There is some legal difference whether a company is a listed joint stock company or a not listed joint stock company as shown in Table 9. A listed company can trade it shares in the stock exchange while an unlisted company cannot do that. In general, listed companies tend to be larger than non listed ones, as larger companies require more funding which open markets can provide. However, this difference will be assumed insignificant for this analysis. To a large extent, this is an innocuous assumption, especially since all but Medicinos and Nordea banks are listed banks.
To compare SSBL banks with the rest of banks, banks are divided into three groups: SSBL, Lithuanian banks (LB) and other foreign banks (OFB). Each bank/branch was placed in the group affiliated with the country where its main shareholder was located. If the main shareholder was a parent bank in Sweden, the bank belonged to “SSBL” group. SEB was owned by Skandinaviska Enskilda Banken in Sweden while Hansabank was firstly owned by Estonian Hansabank which in turn was owned by Hansabank in Sweden. Four banks/branches were grouped into LB (Medicinos, Šiaulių, Ūkio and Snoras) as their ultimate shareholders were individuals in Lithuania and they did not have a parent company. However, the largest single shareholder of Šiaulių bank was European Bank of Reconstruction and Development which meant that this bank was heavily influenced from outside. Therefore, the data on Šiaulių bank will always be included in the LB group but sometimes shown also as an individual bank (ŠB). Five other banks (Nordea, Danske Bank, Hansabank (renamed Swedbank in 2009), DnB Nord, Parex) were sister banks of parent banks operating in other foreign countries and were thus included in the OFB group.
27
Cross-section analysis
This subchapter identifies the variables which reveal differences across the banks. It is found that Swedish sister banks in Lithuania (SSBL) are larger in terms of assets, the number of clients served and are more concentrated than domestic Lithuanian banks (LB) or other foreign banks (OFB). SSBL and OFB both rely heavily on their parent banks for funding which allows devoting more resources to lending than domestic banks. The majority of loans are lent out to firms by all banks, with 45% to 18% of loans directed to households. A breakdown of loans to households shows that SSBL and OFB were dominant in providing mortgage loans.
When comparing SSBL to OFB and LB, besides many similarities several differences emerge in terms of the clients who were served and services which were offered in 2007. Bank groups are part of wider conglomerates each of which provide such services as insurance and leasing besides commercial banking services. Also, all of them sell their services to households as well as legal entities (Table 10). All banks use various client oriented technologies such as payment cards, ATM machines and Cards POS terminals. However, SSBL have a larger customer base and more resources than the OFB in Lithuania. SSBL banks had 4.8 times more clients than LB and 3.7 times more clients than OFB at the end of fourth quarter of 2007. Even though all banks served household and legal entities, SSBL served more clients in each category. The SSBL offer more client technologies. The SSBL also employ labour force which is roughly equal to the sum of employees of LB and OFB. Also, SSBL have fewer numbers of branches (buildings) than LB but more banks bank branches than OFB. This could be explained by a wider geographical distribution of LB while SSBL and OFB tend to be established within cities.
Table 10: General information on banking sector (year 2007)
SSBL LB OFB
Number of branches (buildings) 200 399 143
Number of employees (conglomerates) 5,647 3,859 2,596
- o/w: number of employees in the bank 5,078 2,696 2,450
Payment cards 2,545,825 833,371 502,055
ATMs 676 375 281
Cards POS-terminals 22,014 1,501 6,892
Clients 2,347,318 486,819 682,894
- o/w: Households 2,249,230 465,740 634,457
- o/w: Legal entity 98,088 21,079 48,437
Source: ALB. Calculations by the author
While the above distribution helps seeing the absolute size of the banks, additional information can be gained by using relative measures. Below the above variables were normalized by the number of clients that the bank served.
28
Table 11: Normalized indicators about the Lithuanian banking sector (year 2007)
SSBL LB OFB
Number of branches per 100,000 clients 9 82 21
Number of employees (group) per 100,000 clients 241 793 380 - o/w: number of employees in the bank per 100,000 clients 216 554 359
Payment cards per 10 clients 11 17 7
ATMs per 100,000 clients 29 77 41
Cards POS-terminals per 100,000 clients 938 308 1,009
Clients 1 1 1
- o/w: Households (per cent of clients) 96% 96% 93%
- o/w: Legal entity (per cent of clients) 4% 4% 7%
Source: ALB. Calculations by the author
Data show that the banks are not homogenous. Table 11 reinforces the argument that SSBL are more concentrated than OFB and LB in particular (judging from the number of branches per 100,000 clients and the number of employees per 100,000 clients, and ATMs per 100 000 clients). At the same time SSBL are similar to LB in terms of the share of legal entities and households while OFB were somewhat more focused on legal entities.
Table 12: Breakdown of assets (2007, thousands LTL)
LB SSBL OFB ŠB
Cash in banks and other financial institutions 2,445,554 2,835,229 916,878 40,490
Loans granted 6,967,218 35,153,209 17,898,818 1,540,031
Debt securities 1,619,463 4,588,990 1,139,778 182,696
Equity securities 284,669 188,296 21,694 6,867
Total assets 12,493,840 46,899,446 21,482,760 2,013,146 Source: ALB. Calculations by the author
SSBL were market leaders in terms of assets and were followed by OFB and LB as shown in Table 12. LB had the most equity securities and came very close to SSBL in the value of cash held. Table 13 presents the same figures but normalized by the value of assets.
Table 13: Breakdown of assets (2007, % of assets)
LB SSBL OFB ŠB
Cash in banks and other financial institutions 20% 6% 4% 2%
Loans granted 56% 75% 83% 76%
Debt securities 13% 10% 5% 9%
Equity securities 2% 0% 0% 0%
Total assets 100% 100% 100% 100%
Source: ALB. Calculations by the author
SSBL are quite similar to OFB in the distribution of assets but different from LB (excluding ŠB). Table 13 shows that in 2007 SSBL, OFB and ŠB held a relatively low share of assets as cash and equity securities when compared to LB. This allows SSBL to dedicate 75% of their assets to loans as compared to 56% for LB. ŠB and OFB are even more focused on loan provision (76% and 83%
29 respectively). Since loans grant interest as opposed to cash, SSBL could be considered more efficient than LB. At the same time, this exposes SSBL, ŠB and OFB to a larger liquidity risk and credit risk. Liquidity risk refers to the ability of banks to finance short term liabilities with short term assets. Since cash is a highly liquid asset, LB seem to be more liquid. In section 3 it was stated that liquidity was managed at a parent bank level which allowed their sister banks to focus on lending. Possibly the relation with ERBD provided such opportunities for the ŠB as well. This was not the case for Snoras, Ūkio and Medicinos banks which had no parent bank support. Credit risk refers to the possibility of lenders failing to repay loans. This is especially relevant for macroeconomic shocks, when there is an increase in failure to repay all types of loans. More loans does not necessarily mean that credit risk is higher as some loans are less risky than others, yet this would be a first indication of the risk being greater in SSBL than in LB. Access to parent funding can help protect against this risk.
As loans constitute the largest share of assets, Table 14 provides information on the distribution of loans lent to different entities.
Table 14: Loans lent to different entities (2007, % of total loans)
LB SSBL OFB ŠB
General government institutions 1% 1% 2% 1%
Enterprises of state and municipalities 0% 0% 0% 0%
Financial institutions 15% 5% 3% 10%
Private legal entities 63% 53% 51% 71%
Households 20% 40% 45% 18%
Source: ALB. Calculations by the author
A look at the loan distribution in Table 14 reveals several similarities and differences among the bank groups. None of the banks lend significant amounts to public institutions, such as government institutions or state enterprises. Instead, the banks focus on private entities. However, this is where yet more differences between SSBL and LB, and similarities between SSBL and OFB emerge. SSBL as well as OFB allocate quite a high share of their loans to households (40% and 45% respectively) while LB provide only 20% (ŠB 18%). Instead, LB provide relatively more funding to private legal entities (firms) and financial institutions. Therefore, LB are more exposed to shocks affecting business than shocks affecting households. Financial contagion - problems in other (non-parent) banks spreading to banks in Lithuania - would be a greater problem for LB and ŠB than SSBL and OFB.
The difference is further clarified when looking at the distribution of loans granted to households as a share of total assets (Table 15). SSBL and OFB have been especially active in providing housing loans. In the end of 2007, 21% and 26% of assets held on the balance sheets of SSBL and OFB
30 respectively went to finance housing. Furthermore, SSBL were issuing more new loans. Loans worth 0.6% of total assets were granted by SSBL as new housing loans in December alone. LB dedicated only 2% of their assets for housing loans and issued 0.1% of their assets as housing loans in December 2007. ŠB gave out relatively more loans than other LB. LB provided 2% of its assets as overdraft loans whereas the two other bank groups hardly participated in this market. All banks issued a similarly low share of consumer loans and other loans. Therefore, comparison of dynamics of consumer loans in particular and to a lesser extent a comparison of other loans can be more helpful in identifying whether it was changes in lending behaviour or external shocks that influenced changes in lending.
Table 15: Distribution of loans to households by purpose (2007, % of total assets)
Total loans LB ŠB SSBL OFB
Housing loans 2% 6% 21% 26%
Consumer loans 3% 3% 3% 3%
Overdrafts in accounts and cards 2% 0% 0% 0%
Other loans 4% 3% 5% 8%
New loans LB ŠB SSBL OFB
Housing loans 0.120% 0.161% 0.611% 1.145%
Consumer loans 0.117% 0.112% 0.115% 0.114%
Overdrafts in accounts and cards 0.217% 0.065% 0.041% 0.033%
Other loans 0.289% 0.025% 0.117% 0.576%
Source: ALB. Calculations by the author
The liability side of the balance sheet shows several similarities and differences between the different bank groups. None of the banks receive significant loans from international organizations and relatively few debt securities are issued by the banks (Table 16). One significant difference is that SSBL and OFB receive large amounts of funding from banks and other credit institutions and parent banks in particular.
Table 16: Breakdown of liabilities (2007, million LTL)
LB SSBL OFB ŠB
Liabilities to banks and other credit institutions 1,154 16,282 10,844 313 - o/w: liabilities to parent banks and other
financial institutions 0 11,251 5,167 0
Loans from international organizations 84 0 79 84
Deposits 8,907 24,046 7,153 1,300
Issued debt securities 711 875 1,300 29
Total equity and minority interest 1,292 3,337 1,304 270 Total liabilities and minority interest 12,494 46,899 21,483 2,013 Source: ALB. Calculations by the author
31
Table 17: Breakdown of liabilities (2007, % of total assets)
LB SSBL OFB ŠB
Liabilities to banks and other credit institutions 9% 35% 50% 16% - o/w: liabilities to parent banks and other financial
institutions 0% 24% 24% 0%
Loans from international organizations 1% 0% 0% 4%
Deposits 71% 51% 33% 65%
Issued debt securities 6% 2% 6% 1%
Total equity and minority interest 10% 7% 6% 13%
Total liabilities and minority interest 100% 100% 100% 100% Source: ALB. Calculations by the author
24% of the liabilities in SSBL and OFB in 2007 came from parent banks (Table 17). Additionally, OFB gained a significant share of liabilities from other credit institutions (non-parent banks). As a result, SSBL and OFB could rely relatively less on deposit funding (51% and 33% respectively) as compared to LB (71% and 65% for ŠB). Additionally, SSBL and OFB held relatively less equity and minority interest (capital). On the one hand, this could be considered risky behaviour, as capital is the prime way to finance losses. On the other hand, access to parent banks in Sweden and other countries meant that these losses could be financed by them.
Table 18: Deposits by the depositors (2007, % of total deposits)
LB SSBL OFB ŠB
General government institutions 2% 3% 3% 7%
Enterprises state and municipalities 5% 5% 10% 7%
Financial institutions 5% 2% 3% 4%
Private legal entities 34% 30% 43% 20%
Households 55% 61% 42% 61%
Source: ALB. Calculations by the author
The distribution of deposits shows more similarities than differences (Table 18). The banks are similar in terms of their relatively low dependence on public institution deposits and few deposits of financial institutions. The main difference is that SSBL hold deposits of households and fewer deposits of legal entities than LB and OFB do. Income statement information supports the facts found in the literature review that the SSBL had the highest profit, income, and expenditure in 2007 (Table 19).
Table 19: Income, expenses and profit of banks (2007, million of Litas)
LB ŠB SSBL OFB
Income 871,705 129,446 2,703,599 1,191,629
Expenses 653,486 100,245 1,746,054 992,539
Profit (loss) before income tax 218,219 29,201 957,545 199,090 Source: ALB. Calculations by the author
