bank performance analysis- methodology and empirical evidence

Bank Performance Analysis: Methodology and Empirical Evidence (Estonian Banking System, 1994-2002)

August Aarma, PhD, Assistant Professor Depertment of Economics at Tallinn Technical University 101 Kopli Street, 11712 Tallinn, Estonia Phone: + 372-6204051; fax: +372-6204051; e-mail: [email protected] Jaan Vainu, PhD, Research Fellow Depertment of Economics at Tallinn Technical University 101 Kopli Street, 11712 Tallinn, Estonia Phone: + 372-6204051; fax: +372-6204051; e-mail: [email protected] Contact person and presenter of the paper: Vello Vensel, DSc, Professor of Statistics Depertment of Economics at Tallinn Technical University 101 Kopli Street, 11712 Tallinn, Estonia Phone: + 372-6204051; fax: +372-6204051; e-mail: [email protected]

Abstract Banks and other financial institutions are a unique set of business firms whose assets and liabilities, regulatory restrictions, economic functions and operating make them an important subject of research, particularly in the conditions of the emerging financial sectors in the EU accession countries from Central and Eastern Europe (CEE). Banks’ performance monitoring, analysis and control needs special analysis in respect to their operation and performance results from the viewpoint of different audiences, like investors/owners, regulators, customers/clients, and management themselves. Some historical notes on the development of the Estonian banking system and the capital structure of banks are presented in this article. Different versions of financial ratio analysis are used for the bank performance analysis using financial statement items as initial data sources. The usage of a modified version of DuPont financial ratio analysis and a novel matrix approach is discussed in the article. Empirical results of the Estonian commercial banking system performance analysis are also presented in the article (1994-2002). EFM classification codes: 510 – Depository Institutions - Management Journal of Economic Literature Classification number: G21 Keywords: banks’ performance analysis, financial ratio analysis, DuPont model, matrix approach and matrix model, banking system restructuring, production function. Acknowledgement This research was undertaken with joint support from the Göran Collert Foundation (Sweden) and from the Estonian Science Foundation Program (contract no. 5185).

Estonian Banking Sector Performance

2

1. Introduction: Theoretical Background and Overview of Related Literature The problem of banking and financial system soundness has become more important in all countries over the recent years. The financial sector, and especially the banking system, is vulnerable to systemic crises which has led to the creation of costly safety nets, as depositor insurance schemes with well-known moral hazard problem. It is argued that there is increasing evidence that banks are “black boxes” due to the week transparency and banks’ unwillingness to disclose information (Hyytinen and Takalo, 2002 and 2003). To measure banks’ creditworthiness and risk exposures is a complicated issue and it is not easy to interpret banks’ accounting data. Kaminsky and Reinhart (1999, p. 476) argued that “Indicators of business failures and nonperforming loans are also usually available only at low frequencies, if at all; the latter are also made less informative by banks desire to hide their problems for as long as possible.” This means that it is needed to use as fully and complexly as possible all available financial information from the official financial statements of banks for making financial analysis of banks’ performance. There are a lot of lessons to learn from earlier financial and banking crises in various countries and recent crises in Asia. Kaminsky and Reinhart mentioned: “The cycle of over-lending is exacerbated by implicit or explicit deposit guarantees, poor supervision, and moral-hazard problems in the banking sector. Crises are accompanied by an overvaluation of the currency, weakening exports, and the bursting of asset price bubbles” (Kaminsky and Reinhart, 1998, p. 444). It seems that this has been written for characterizing the nowadays situation in the EU accession countries It is noted that the most negative factor is the extremely rapid domestic credit expansion during the last years, funded by massive borrowing from abroad due to the sizeable interest rate differential (see, for example, Krzak, 1998). Doubts about the soundness of the banking sector is also one of the greatest threats to the credibility of a currency board arrangement in Estonia. The contemporary banking crises can be classified mostly as “growth crises”, which are characterized by economic deregulation and liberalization, removal of cross-border restrictions on capital flows, and increased competition in the financial sector. Based on a newly constructed cross-country database of financial liberalization, Abiad and Mody (2003) examined the experience of 35 countries over the period 1973-1996 to analyze underlying causes of financial sector reforms. They found that liberalization is a combination of discrete changes in response to economic and political “shocks”, reinforced by a self-sustaining dynamic (called this as “learning”). They draw five specific conclusions about what produce changes (reform): • Countries whose financial sectors are fully repressed (unliberalized) are the ones with the

strongest tendency to maintain their policy stance and hence remain closed and highly regulated. But, initial reforms cause changes that make further reforms necessary.

• Regional diffusion effects appear to be important – the further a country’s stage of liberalization is from that of the regional leader, the greater is the pressure to liberalize.

• Shocks to the economic environment (a new government; decline in US interest rates) play an important role in weakening the status quo and making reforms possible.

• Crises do trigger action, but not always is the direction of reform – balance of payments crises raise the likelihood of reform; banking crises have the opposite effect.

• Among variables representing ideology and structure, only trade openness appears related to the pace of reform. Not important: presidential or parliamentary regimes, right- or left-wing governments, and the legal system proves not to be influential as well.


3

It is evident that to study results of financial sector reform and restructuring, a profound performance analysis is needed. The traditional financial ratio analysis is mainly used for the bank performance analysis. We can find different versions of this approach from various textbooks about banking and financial institutions. Different versions of DuPont financial ratio analysis (see Cole, 1973) seem to be more perspective for banks’ and other financial institutions’ performance analysis (see, for example, Dietrich, 1996). Recent studies of banks’ efficiency and productivity analysis in different countries can be taken as lessons for the Estonian case - see, for example, Hardy and di Patti, 2001 (Pakistan lessons); Spiegel, 1999 (Japaneze experience); Berger and Mester, 1999; Van Greuning and Bratanovic, 2000; and Stiroh, 2000 (US experience); Rebelo and Mendes, 2000 (Portuguese experience); Hasan and Marton, 2000 (Hungarian lessons); Andersen et al., 2000 (Finnish experience); ECB, 1999 and 2000 (EU banks’ experience); Kwan, 2002 and 2003 (Asian countries experience). Different financial ratios are used as predictors of bank failures (Estrella et al., 2000). Berger and Humphrey (1997) presented a review of 122 stud ies in 21 countries about the efficiency and productivity of financial institutions. Mörttinen (2002) demonstrated how the banking sector’s service production can be measured using aggregate financial statement and payment transactions data; she computed banking sector labor productivity Tornqvist indices for six countries (Finland, Sweden, UK, Germany, France and Italy). The focus of financial analysis for the management of any bank (or the banking sector as a whole) should be on the efficiency of performance of the bank measured from the viewpoint of investors/owners’ income maximization. More widely, all stakeholders have to be interested in the performance results of the banks. The concept of a “stakeholder monitor” is useful to take into account in designing performance analysis of any bank. This concept was developed by David Llewellyn – see Llewellyn 2002, Llewellyn and Mayes, 2003. We agree with the suggestion that “Stakeholders, as the name implies, have something in stake in the relative success or failure of the firm. Those who participate in the process of observing the behavior of the firm and forming judgements in the light of it can be described as “monitors”. Such monitors may have access to both market and private information. Combining these ideas, “Stakeholder monitors” are all those agents who have an interest in the outcome of the monitoring process” (Llewellyn and Mayes, 2003, p. 11). The incomplete list of “stakeholders monitors” include: supervisory agencies, rating agencies, market traders, shareholders, board of directors, debt-holders, depositors, managers, borrowers, employees. In carrying out bank performance analysis, it is important to emphasize that banks differ in their corporate governance from firms in other, less regula ted industries. These differences, in turn, present their own challenges for bank managers, regulators, depositors, investors, and other stakeholders. “Bank managers live in a more complex environment than their peers in industry due to bank regulations. In addition to the demands placed on them by shareholders, regulators have strong incentives to influence managerial action, and this may be in conflict with shareholder demands” (Harm, 2002, p. 5). Governance is a set of mechanisms with which the providers of capital and other stakeholders are defending their interests against the firm. The firm is run by managers, and this a point where conflicts of interests starts. An excellent survey of recent literature (both theoretical and empirical) is also presented by Harm (op. cit., pp. 109-128). Macey and O’Hara (2003) argue that bank officers and directors should be held to broader (if not higher) set of standards than their counterparts in less regulated non-financial firms, and banks pose special corporate governance problems. Kose and Qian (2003) consider another important theme in the corporate governance of banks – the effect of the incentive features


4

built into the compensation schemes of bank mangers. Adams and Mehran (2003) focus also on the differences between the corporate governance of banks and manufacturing firms and support the theory that governance structures are industry-specific. In general, the components of firm’s governance structure are determined by various different factors, which all will influence also performance analysis aims and techniques: the nature and structure of firm’s assets and liabilities (leverage, share of financial assets, business risk, cash-flow patterns), firm size, industry, regulations, etc. Various measures of rates of return are used mainly for that purpose. We fully agree with the opinion that “Relaying too heavily on just a few indicators of bank profitability can be misleading. While ROA, ROE, and interest margin (and noninterest expenses) to gross income remain the key measures, they should ideally be supplemented by the analysis of other operating ratios” (Sundararajan, 2002, p.20). In this article, we present one of the possible approaches to such financial analysis using the modified version of DuPont analysis (see Cole, 1973), which is similar to Dietrich’s (1996) approach, and the novel matrix approach which is was firstly presented in Vensel, 1997 (see also Vensel, 2001). We have selected the following years for empirical analysis of the Estonian banking system performance: 1994 – the first year of macroeconomic stabilization after the currency reform in June 2002 and after resolution of the first banking crisis in Estonia; 1997 – extraordinary optimistic year, the first unsuccessful attempts of Estonian banks to expand into Russian and Baltic countries’ markets, Bubble of the Tallinn Stock Exchange; 2000 – the second phase of macroeconomic stabilization and resolution of the second banking crisis in Estonia; 2001-2002 – recent years before accession to the EU The paper is organized as follows. The most important recent developments in the Estonian banking system are presented in Section 2 (some historical notes; banking crises and banks’ rehabilitation; structural developments). DuPont financial ratio analysis modified methodology and empirical results of the Estonian banking system performance are discussed in Section 3. In Section 4, the methodology of a novel matrix model and the usage of this methodological approach are presented. Construction of the Cobb-Douglas production function is discussed in Chapter 5. The paper ends with some concluding remarks. 2. Development of the Estonian Banking System 2.1. Historical Notes The first commercial bank (Tartu Commercial Bank) on the territory of the former Soviet Union was established in Estonia in 1988. This bank went bankrupt and was liquidated in 1992-1993. So, since there was a great demand for banking services by the emerging private sector, the maximum number of commercial banks operating simultaneously in the small Estonian banking market was 42 in 1992. Some of them were liquidated during the banking crises in 1992-1994 and in 1998-1999, and some of them were merged into larger commercial banks. A short history of the Estonian contemporary banking system is presented in Table 1. Up till 1997, the development of the Estonian banking sector was characterized by a rapid nominal growth of total assets and loan portfolios. 1997 was also the beginning of a new stage in the development of the Estonian financial sector, especially in the international context, which is confirmed by investment grade credit ratings assigned to Estonia: Standard and Poor’s BBB+ and Moody’s Investors Service’s Baa1. It has to be added that from 2001-2002


5

Estonia has the fo llowing credit ratings by rating agencies (Leemets and Reedik, 2003, p. 49): Moody’s foreign currency and kroon ratings both A1 (from 12.11.2002); Standard&Poor’s rating both A- (from 20.11.2001); Fitch foreign currency rating A- and kroon rating A+ (from 30.08.2001). The rapidly growing economy (GDP growth rate in 1997 about 11%) boosted credit demand, and also non-banking financial intermediation accelerated. However, implementation of the expected Estonian banks expansion to the other Baltic countries and Russia was only partly realised due to the tightened market situation both in Estonia and internationally. Negative results of the over-optimistic and risky attitude towards the opportunities of the Eastern market and consequences of the bursting of the 1997 stock exchange bubble in Estonia became clearly evident during 1998-1999.

Table 1. History of the Estonian Banking Sector (Only Operating Banks, 2003)

No Bank Established Organizational Changes A. Large Banks 1. Hansapank 01.07.1991 Merged with the Estonian Savings Bank (which was

established 14.04.92 on the basis of former state-owned savings offices and merged with the Estonian Industrial Bank in 1996) in 1998

2. Union Bank of Estonia

15.12.1992 Established on the basis of 11 smaller regional banks, merged with North-Estonia Bank in 1997 and with the Bank of Tallinn (which was established 21.12.92) in 1998

B. Medium-Sized Banks 3. Nordea Bank

Plc, branch 20.06.1995 Established on the basis of merging KOP and SYP

(Finnish banks) offices 4. Sampo Bank 30.06.1992 Previous Optiva Pank, former Forexbank, merged

with Raepank in 1995 and with Estonian Investment Bank (established 30.06.92) in 1998, Sampo-owned since 2000

C. Small Banks 5. Estonian

Credit Bank 10.04.1992 Small niche bank, majority owned by non-resident

legal persons 6. Tallinn

Business Bank 09.12.1991 Small niche bank, majority owned by Estonian legal

persons 7. Preatoni Bank 23.09.1999 Oriented to foreign investments, real estate financing

and asset management The rapid nominal growth both in the real and financial sectors, the deepening dependence on international financial markets and financial problems in the emerging markets in South-East Asia dictated several steps of precaution by the government and the central bank. The most important long-term regulatory measures included raising of the banks’ minimum capital adequacy ratio from 8% to 10%, increasing the risk-weight of local governments’ liabilities from 50% to 100%, and a decision to introduce a market risk component to the capital adequacy ratio. The intermediate steps included the introduction of reserve requirements to the net liabilities of domestic banks vis-à-vis non-resident banks and additional liquidity requirement to restrain capital inflow.


6

Compared to previous years, the growth rate of nominal indicators in the banking sector slowed down during 1998-2000, partly due to the changes in the external environment. With the deterioration of the economic environment in 1998, wrong economic and management decisions that had been made already earlier, surfaced in 1998 and resulted, for example, in the dropout of three banks from the banking market in July-October. Some of the more important interrelated systematic factors behind wrong management decisions were: the expansive development in previous years, lack of experience in doing business in the changing market conditions, insufficient transparency of the market, owners’ weak control over the activities of executive management, tightened competition in the banking market, insufficient risk hedging and management, and external shocks. In 1998, a wave of mergers and restructuring took place in the Estonian banking sector. After the completion of these mergers, Scandinavian banks started to show greater interest in the Estonian banking market. As a result, Swedbank acquired 56% of Hansapank and Skandinaviska Enskilda Banken (SEB) acquired 32% of the Union Bank of Estonia. We may conclude that the Estonian banking sector became healthier when Swedish banks and other Nordic investors joined the circle of bank owners, improving the future outlook of the banking system. So that if during the first banking crises in 1992-1994, Estonia had to resolve the problems by itself, then during the second banking crises in 1998-1999, foreign banks also helped and supported to get over the crises. Smaller banks in Estonia were affected also by the negative developments in Russia. The liquidation of some banks continued in 1999, accompanied by the declaration of the bankruptcy of EVEA Pank and ERA Pank. On the other hand, the first new banking licence issued since 1993 was granted to the new Preatoni Pank in September 1999. Preatoni Pank has focused mainly on intermediation of foreign capital into Estonian economy, real estate financing and asset management. During 1999, Swedish banks - SEB and Swedbank - increased their participation in the equity capital of the Union Bank of Estonia and in Hansapank over 50%. 2.2. Banking Crises and Bank Rehabilitation Estonia has experienced two serious banking crises during the about 12-years period of its banking sector development and restructuring, the first crisis in 1992-1994 and the second in 1998-1999. The first banking crisis occurred during the hard period of starting drastic economic reconstruction when production output was reducing dramatically and the country underwent a period of hyperinflation. The characteristic feature of the first banking crisis in Estonia was that it was caused by internal reasons and it was overcome with Estonia’s own resources and management skills. The main causes of this banking crisis were severe problems in the whole economy, poor bank management and lack of professional skills, weak supervision both from the side of the central bank and owners. The depositors’ losses in the banking crisis were large, the money supply decreased, many loans were depreciated, and the trustworthiness of the banking system fell significantly. The central bank acted quite quickly and resolutely to overcome the banking crisis. The Bank of Estonia brought the prudential requirements into its operation on the basis of international experience for protecting creditors’ and clients’ interests beginning from January 1993. In April 1993, the Bank of Estonia announced a stabilization period in the banking system, during what the issuance of new banking licenses was frozen and for the existing banks it established a schedule of gradual rise in minimum equity capital. After that, the Bank of


7

Estonia did not renew licenses of 8 banks, 10 banks merged into one bigger bank, a moratorium was declared on 3 banks. Looking back, it is possible to establish some signs of a future banking crisis in 1998-1999: (1) Estonian banks took extraordinary high financial risks through investment companies and

their subsidiary companies to get big profits via speculating in securities market – rapid fall in prices on the share market in autumn 1997 reduced significantly banks’ profits and at the end of 1997 and in 1998 almost all banks operated in losses;

(2) Banks hold very high negative level of gap (interest rate sensitive liabilities exceeded significantly rate-sensitive assets) for earning excessive profits in the environment were interest rates steadily decreased during the previous years and they were not able to adjust to changed environment with increasing interest rates from the second half-year of 1997;

(3) Commercial banks absorbed heavily into non-banking business – for example, later bankrupted the Land Bank of Estonia owned seven subordinate establishments and related companies, which dealt with leasing and investing, and with anything else but banking (hotels, processing agricultural products, broadcasting etc), also other banks were absorbed into risky non-banking business;

(4) The decision to expand to the Eastern market (Russia and other Baltic States), where the interest rates and possible profitability seemed to be higher, was also too risky and premature, especially in the framework of the Russian crisis in 1998;

(5) There were various disputes and conflicts if interests between the owners and management which led to wrong (mismanagement) decisions Good examples should be the Land Bank of Estonia and the Estonian Investment Bank – for example, the shareholders of the Investment Bank intended to sell the bank to the German Schleswig-Holstein Bank in autumn 1997, but the top executives threatened to hand in a collective resignation and so the bank was sold to them.

(6) Sometimes there were inadvisable relations between the bank management and political powers, and correspond ing political pressure – a typical “political” bank was the Land Bank of Estonia where almost all financial risks were ignored and later the Government lost its deposits in the bank amounting to more than 800 million Estonian crowns, EEK (more than 50 million euros).

The occasion of starting the second banking crisis was the burst of a market bubble on the Tallinn Stock Exchange in the Autumn 1997, caused partly by the impact of the financial crises in the South-East Asia and supported lately by the Russian crisis in Autumn 1998. In 1998, a wave of mergers and restructuring took place in the Estonian banking sector. We may conclude that Estonian banking sector became healthier when Swedish banks and other Nordic investors joined the circle of owners of banks, improving banking system’ future outlook. So, if during the first banking crisis in 1992-1994 Estonia had to resolve the problems by itself then during the second banking crisis in 1998-1999 foreign banks also assisted and supported to get over the cris is.

The authors are on the opinion that the currency board arrangement helped in Estonia to resolve banking crises rapidly and mostly effectively without remarkable rehabilitation costs. The main instruments for anticipating banking crises are tightening of prudential requirements and strengthening of banking supervision. Recent changes in the operational framework for monetary policy and banks’ prudential ratios in Estonia were aimed at enhancing financial stability and increasing the liquidity buffers of the financial system. In short-term, the priority focused on restoring foreign investors’ confidence in Estonian economic viability.


8

We may argue that the currency board arrangement practically did not help banks felt into trouble because its resources are intended for guaranteeing the local currency and the central bank is not acting as the lender of the last resort. The currency board is not able to avoid banking crises and can not guarantee a “soft landing” and rehabilitation of banks felt into trouble. At the same time, the currency board arrangement supported and strengthened the discipline and responsibility of the main actors – banks, the central bank, depositors, and the Government. A stable currency and presence of respective financial safety net compensated the absence of classical lender-of- last resort facility and ensured development of in general reliable banking sector. Only Optiva Bank can be mentioned as a successful rehabilitation case in Estonian banking market. Two banks in trouble, Estonian Investment Bank and Forexbank, merged in 1998, the central bank obtained the ownership in merged new bank, re-capitalized the bank and sold it to the Finnish Sampo Group in 2000. So as the bank was relatively small, the rehabilitation costs were not very high. 2.3. Structural Developments The structure of the Estonian banking sector has changed fundamentally during the last years. Today, the banking system is highly concentrated and two Swedish-owned banks dominate in the market (see also Table 1). The consolidation process continued throughout the second banking crisis in 1998-1999 resulting in fundamental bank reorganizations. We can notice all three world-wide trends in the financial consolidation process also in the Estonian market: domestic consolidation, foreign entry and cross-border consolidation, and the formation of financial conglomerates and bancassurances. Some characteristics of the development of the Estonian financial market structure are presented in Table 2. Some more interesting conclusions from Table 2: • The banking market concentration (the share of three largest banks’ assets in total banks’

assets) achieved more than 90% already in 1998; it was 90.4% at the end of 2002; • foreign banks’ share in total assets of Estonian commercial banks increased dramatically

and was 97.5% at the end of 2002; • the Estonian financial sector is clearly bank-oriented – the bank assets to GDP ratio was

75.6% and the banks assets share in total financial assets was 45.2% at the end of 2002; • private credits by banks and other financial institutions increased considerably during the

analyzed period – private credits by banks to GDP ratio was 46.2% and overall private credits to GDP ratio 62% in 2002;

• relatively rapidly have grown leasing and factoring portfolio (about four times during 1997-2002) and stock market capitalization (about 5.5 times); total financial assets ratio to GDP has risen to 167% at the end of 2002.

Table 2. Some Indicators of the Estonian Banking and Financial Sector Development

Indicator 1997 1998 1999 2000 2001 2002 02/97 Number of commercial banks 11 6 7 7 7 7 0.636 Number of private banks 11 5 6 7 7 7 0.636 Number of foreign banks 1 2 2 4 4 4 4.000 Concentration index C3, % 69.7 93.0 92.4 91.1 91.1 90.4 1.297 Concentration index C5, % 83.4 99.4 98.9 98.8 98.9 99.1 1.188 Total assets, EUR m 2594 2620 3008 3695 4372 5221 2.013


9

Total assets/GDP, % 63.4 55.7 61.7 67.7 71.8 75.6 1.192 Foreign ownership in share capital, % 44.2 60.7 61.6 83.6 85.4 86.7 1.962 Foreign ownership in total assets, % 2.3 90.2 89.8 97.4 97.5 97.5 42.39 Private credit by banks, EUR m 1362 1527 1704 2189 2601 3193 2.344 Private credit by banks/GDP, % 33.2 32.6 35.4 40.1 42.7 46.2 1.392 Leasing and factoring portfolio, EUR m 315 399 433 644 893 1232 3.911 Leasing and factoring/GDP, % 8 8 9 12 15 18 2.250 Debt market capitalization, EUR m 258 235 204 231 279 211 0.818 Debt market capitalization/GDP, % 6 5 4 4 5 3 0.500 Stock market capitalization, EUR m 837 531 1913 2095 1999 4570 5.460 Stock market capitalization/GDP, % 20 11 39.8 38.4 32.8 66.2 3.310 Insurance gross collected premiums 70 81 83 98 112 134 1.914 Gross collected premiums/GDP, % 1.7 1.7 1.7 1.8 1.8 1.9 1.118 Investment funds’ assets, EUR m 97 23 73 95 193 280 1.887 Investment funds’ assets/GDP, % 2.4 0.5 1.5 1.7 3.2 4.1 1.708 Total financial assets, EUR m 2458 2912 5550 6727 7748 11551 4.699 Total financial assets/GDP, % 60 62 115 123 127 167 2.783 Total private credit, EUR m n.a. 1902 2106 2777 3395 4308 2.265 Total private credit/GDP, % n.a. 40 43 50 55 62 1.550 GDP, EUR m 4110 4685 4813 5458 6089 6904 1.680 GDP real growth, % 10.6 4.7 -1.1 6.4 5.3 4.7 n.a. Source: Bank of Estonia Notes: (1) Total financial assets consist of the assets of the central bank and other financial institutions, debt securities market, stock market, leasing and factoring portfolio, and insurance gross premiums from; (2) Foreign banks consist of foreign banks’ branches in Estonia and the banks majority owned by foreign banks. At the end of 2002, there were six credit institutions operating in the Estonian banking market, a branch of a non-resident credit institution (Nordea Bank Finland Plc, Estonian Branch) and representative offices of six non-resident financial institutions (Landesbank Schleswig-Holstein and Norddeutsche Landesbank Girozentrales, Svenska Handelsbanken, OKO Osuuspankkien Keskuspanki OYJ, Parekss Banka, Vereins- und Westbank AG,). The ownership structure of Estonian banks is presented in Table 3. The dependence of the Estonian banking system on the developments in international financial markets and on foreign investors’ preferences deepened from year to year. In the course of the restructuring process, foreign banks increased their share in equity capital from 10.3% in 1996 to 79% at the end of 2002. The total share of non-resident owners has risen to 86.7% at the end of 2002.

Table 3. Ownership Structure of Estonian Banks, %

Estonian Owners Non-Resident Owners Year Public Sector

Legal Persons

Individuals Total Banks Legal Persons

Individuals Total

1996 12.0 NA NA 62.8 10.3 NA NA 37.2 1997 4.2 41.6 11.3 57.1 22.7 19.6 0.6 42.9 1998 13.6 22.3 8.6 44.5 45.5 9.5 0.5 55.5 1999 11.6 15.2 11.0 37.6 52.6 8.9 0.7 62.2 2000 0.0 6.8 9.3 16.1 67.0 16.7 0.2 83.9 2001 0.0 5.6 8.5 14.1 63.3 22.3 0.3 85.9 2002 0.0 5.2 8.1 13.3 79.0 7.6 0.1 86.7


10

Equity investments by Swedish banks in two largest Estonian banks (Hansapank and Union bank of Estonia) in 1998 and by Finnish insurance company Sampo Group in Optiva Pank in 2000, increased the share of all non-resident owners from 37.2% to 85.9% during 1996-2001. The public sector (mostly the Bank of Estonia) share in the ownership structure increased in 1998 due to the rescue operation of two smaller banks (the central bank was the core shareholder of the newly established Optiva Pank ), and decreased to zero already at the end of 2000 due to the sale of Optiva Pank to Sampo Group. 3. DuPont Financial Ratio Analysis 3.1. Methodology The starting point of the bank performance analysis is to calculate the book rate of return on equity, ROE

BVE

EATROE

Equity, of ValueBook

Taxes,After Earnings= (1)

which consists of three components: • pull- through, U

EBT

EATU

Taxes, Before Earnings

Taxes,After Earnings= (2)

• financial leverage, LEV

BVE

TALEV

Equity, of ValueBook Assets, Total

= (3)

• return on total assets, ROA

TA

EBTROA

Assets, Total Taxes, Before Earnings

= (4)

These financial ratios form the multiple factor system

BVE

EAT

TA

EBT

BVE

TA

EBT

EATROE =××= (5)

All these financial ratios are widely used for a bank performance analysis. Pull-through (U) shows success of the bank tax management policy as it may be interpreted as one minus the average corporate tax rate. The financial leverage ratio (LEV) measures how many Estonian crowns (EEK) of assets the bank has per EEK of equity and may be interpreted as a bank’s “gearing”. Return on total assets (ROA) is one of the most frequently used financial ratios by financial analysts. ROA measures the ability of bank management to generate income after all financial and non-financial costs and expenses for owners. Changes in ROA are usually the cause of the most important changes in banks’ performance and need a more detailed analysis. The other financial ratios such as components of ROE, pull- through (U) and financial leverage (LEV), reflect tax treatment and capitalization rate, and they usually change less. ROA may be divided into the following components: • bank burden, B

TA

NIENIRTA

NNIRB

−==

Assets, Total Revenue,Interest -NonNet

(6) where NIR - non- interest revenue; NIE - non- interest expense;


11

• earning assets ratio, EAR

TA

EAEAR

Assets, Total

Assets, Earning= (7)

• net interest margin, NIM

EA

IEIR

EA

NIRNIM

−==

Assets, Earning

Revenue,Interest Net (8)

where IR - interest revenue; IE - interest expense, Financial ratios (6-8) form a factor system

TA

EBT

TA

NIRNNIR

EA

NIR

TA

EA

TA

NNIRROA =

+=×+= (9)

Burden (B) measures a bank management’s control of operating expenses. The burden for banks is negative to show the fact that non- interest revenue (fees, earned commissions, other operating income) does not cover labor and other administrative or non- interest expenses. Earning assets ratio (EAR) is usually not an important factor of changes in ROA but it may be interesting to make comparisons between various banks because EAR characterizes different development strategies. Net interest margin (NIM) is a more important and widely used financial ratio in the factor system (9). NIM reflects the interest spread between assets and liabilities, it focuses on the net earnings from investing through borrowed funds and is the major source of profitability for the bank. For a more detailed analysis, NIM may be divided into three following components: • return on earning assets, REA

EA

IRREA

Assets, Earning

Revenue,Interest = (10)

• cost of liabilities, COL

L

IECOL

s,Liabilitie

Expense,Interest = (11)

• liabilities to earning assets ratio, LEA

EA

LLEA

Assets, Earning

s,Liabilitie= (12)

which form the factor system

EA

NIR

EA

IEIR

EA

L

L

IE

EA

IRNIM =

−=×−= (13)

Return on earning assets (REA) connects directly earning assets and interest revenue generated by them. Thus, REA characterizes the average rate of lent funds and earned dividends. The cost of liabilities (COL) may be interpreted as the average price of borrowed capital. Liabilities to earning assets ratio (LEA) measures the intensity of bank investment activity. 3.2. Banking Sector Performance and Profitability - DuPont Analysis It is argued that internationalization, adoption of new banking technologies, deregulation, banking market consolidation and other recent trends in financial intermediation should result in increasing efficiency. On the other hand, since banks are no longer monopoly suppliers of financial services and products and markets are more contestable (increased competition


12

between banks and new competition from non-bank financial institutions and markets), intermediation margins, net interest income and other income should result in decreasing profitability and efficiency. In any case, elimination of inefficiencies and reducing costs would be a challenge for banks’ survival in the rapidly changing market environment. Initial financial information for Estonian banking sector performance analysis (1994-2002) is presented in Tables 4-5 on the basis aggregated consolidated financial statements published by the Bank of Estonia (see Bank of Estonia, Annual Reports).

Table 4. Aggregated Consolidated Balance Sheet of Estonian Commercial Banks (end of the Year, EEK m)

Item 1994 1997 2000 2001 2002 02/94 02/01 Assets 1. Cash and reserves 1421.0 4247.7 7046.2 5378.3 4954.0 3.486 0.921 Cash 649.2 1212.4 1335.2 1596.4 1394.1 2.147 0.873 Claims on central bank 771.8 3035.3 5711.0 3781.9 3559.9 4.612 0.941 2. Earning Assets 7905.9 32995.4 47102.8 59985.2 73669.9 9.318 1.228 Claims in credit institutions 2681.5 3582.8 5057.6 9040.9 10109.8 3.770 1.118 Claims on customers 4526.6 21315.6 34253.1 40714.9 49976.4 11.04 1.227 Uncollectible claims (-) -154.8 -423.1 -532.2 -598.6 -522.2 3.373 0.872 Securities 852.6 8520.1 8324.3 10828.0 14105.9 16.54 1.303 3. Tangible and other assets 1059.0 3339.4 3670.0 3047.5 3062.3 2.892 1.005 Tangible assets 626.1 1378.1 1226.0 1064.5 1040.4 1.662 0.977 Other assets 432.9 1961.3 2444.0 1983.0 2021.9 4.671 1.020 Total assets (1+2+3) 10385.9 40582.5 57819.0 68411.0 81686.2 7.865 1.194 Liabilities and capital 1. Liabilities 9618.8 36411.6 50540.1 59331.8 71766.8 7.461 1.210 Liabilities to customers 6943.7 21400.7 34774.3 42698.5 48802.2 7.028 1.143 Liabilities to credit institutions and central bank

920.9

6743.4

6271.6

6449.0

11878.6

12.90

1.842

Other liabilities 1754.2 8267.5 9494.2 10184.3 11086.0 6.320 1.089 2. Equity 767,1 4170.9 7278.9 9079.2 9919.4 12.93 1.093 Share capital 712.0 2551.5 5927.9 6149.8 6237.5 8.761 1.014 Reserves 115.6 518.5 730.9 731.7 734.3 6.352 1.004 Profit or loss retained -60.5 1100.9 620.1 2197.7 2947.6 - 1.341 Total liabilities (1+2) 10385.9 40582.5 57819.0 68411.0 81686.2 7.865 1.194 (Source: Bank of Estonia, Annual Reports) The Estonian bank ing system has grown rapidly in nominal terms. The respective growth rates for 2001/1994 and 2001/2000 are also presented in Tables 4-5. In general, we can see high growth rates in almost all balance sheet and income statement items during 1994-2001: for example, both earnings before taxes and after taxes have risen more than 20 times, banks’ equity has risen 14 times, share capital 10 times and total assets about 7 times, etc. A financial ratio analysis is needed for analyzing profitability and efficiency changes in the banking system, using a modified version of DuPont financial ratio analysis technique (see Dietrich, 1996; Vensel, 2001). Initial financial information in the form of simplified consolidated financial statements of the Estonian commercial banking system as a whole in 1994-20012 is presented in Table 6. The respective growth rates for 2002/1994 and 2002/2001 are also presented.


13

Table 5. Aggregated Consolidated Income Statement of Estonian Commercial Banks Indicator 1994 1997 2000 2001 2002 02/94 02/01 1. Interest revenue (IR) 943.6 2658.5 3744.2 4308.1 4253.5 4.508 0.987 Loans 780.2 1956.9 2816.8 3308.2 3309.9 4.242 1.001 Other interest income 163.4 701.6 927.4 999.9 943.6 5.775 0.944 2. Interest expense (IE) 312.8 1217.2 1811.9 2125.7 1883.0 6.020 0.886 Deposits 248.5 690.3 1065.0 1269.8 913.1 3.674 0.719 Other interest expense 64.3 526.9 746.9 855.9 969.9 15.08 1.133 3. Net interest revenue (1-2) 630.8 1441.3 1932.3 2182.4 2370.5 3.758 1.086 4. Non-interest revenue 457.0 3272.0 2065.6 2895.1 2613.4 5.719 0.903 Commission income 217.3 799.1 965.1 1062.6 1202.6 5.534 1.132 Financial operations 115.4 715.1 755.0 690.7 1031.4 8.938 1.493 Other income 124.3 1757.8 345.5 1141.8 379.4 3.052 0.332 5. Non-interest expense 1019.8 3644.4 3384.8 3373.7 3769.1 3.696 1.117 Commission expense 76.2 250.8 255.9 282.9 333.4 4.375 1.179 Administrative expenses 611.5 1244.0 1373.6 1583.9 1757.8 2.875 1.110 Other expenses, losses 332.1 2149.6 1755.3 1506.9 1677.9 5.052 1.113 6. Net non-interest revenue (4-5)

-562.8 -372.4 -1319.2 -478.6 -1155.7 2.053 2.415

7. Earnings before taxes (3+6) 68.0 1068.9 613.1 1703.8 1214.8 17.86 0.713 8. Tax expense 27.1 105.8 - 20.4 61.6 2.273 3.020 9. Earnings after taxes (7-8) 40.9 963.1 613.1 1683.4 1153.2 28.20 0.685 (Source: Bank of Estonia, Annual Reports)

Table 6. Simplified Consolidated Financial Statements of the Estonian Banking System (1994-2002, million kroons)

Items 1994 1997 2000 2001 2002 02/94 02/01 Income Statement Data Interest Revenue, IR 943.6 2658.5 3744.2 4308.1 4253.5 4.508 0.987 Interest Expense, IE 312.8 1217.5 1811.9 2125.7 1883.0 6.020 0.886 Net Interest Revenue, NIR = IR – IE

630.8

1444.1

1932.3

2182.4

2370.5

3.758

1.086

Non-Interest Revenue, NOIR 457.0 3272.0 2065.6 2895.1 2613.4 5.719 0.903 Non-Interest Expense, NOIE 1019.8 3644.4 3384.8 3373.7 3769.1 3.696 1.117 Net Non-Interest Revenue, NNIR = NOIR – NOIE

-562.8

-372.4

-1319.2

-478.6

-1155.7

2.053

2.415

Earnings Before Taxes, EBT = NIR + NNIR

68.0

1068.9

613.1

1703.8

1214.8

17.86

0.713

Earnings After Taxes, EAT 40.9 963.1 613.1 1683.4 1153.2 28.20 0.685 Balance Sheet Data Cash and Reserves, R 1527,8 3203.8 6578.0 6212.3 5166.2 3.381 0.832 Earning Assets, EA 6117.8 25817.0 42019.6 53544.0 66827.5 10.92 1.248 Fixed and Other Assets, FA 742.9 2743.1 3847.3 3358.7 3054.9 4.112 0.910 Total Assets, TA = R+EA+FA

8388.5 31763.9 52444.9 63115.0 75048.6 8.947 1.189

Liabilities, L 7667.3 28562.7 45164.2 54936.0 65549.2 8.549 1.193 Book Value of Equity, BVE 721.2 3201.2 7280.7 8179.0 9499.4 13.17 1.161 Source: The Bank of Estonia Statistical Datasheets and Bulletins


14

Using initial data from Table 6 (the balance sheet data are averaged), results of DuPont financial ratio analysis are presented in Table 7. These results need some comments, focusing on the growth rates of 2002/1994.

Table 7. Financial Ratio Analysis of Estonian Commercial Banks (1994-2002) Financial Ratio 1994 1997 2000 2001 2002 02/94 02/01 Book Rate of Return, %,ROE = EAT/BVE 5.671 30.09 8.59 20.58 12.14 2.141 0.590 Components of ROE, ROE = U×LEV×ROTA Pull-through, %, U = EAT/EBT 60.15 90.10 100.0 98.80 94.93 1.578 0.961 Financial Leverage, LEV = TA/BE 11.63 9.92 7.203 7.717 7.90 0.679 1.024 Return on Total Assets, ROTA = EBT/TA 0.811 3.365 1.192 2.700 1.619 1.996 0.600 Components of ROTA, ROTA = B + EAR×NIM

Burden, %, B = NNIR/TA -6.709 -1.172 -2.493 -0.755 -1.540 0.230 2.040 Earning Assets Ratio, %, EAR = EA/TA 72.93 81.28 80.12 84.84 89.05 1.221 1.050 Net Interest Margin, %, NIM = NIR/EA 10.31 5.594 4.599 4.076 3.547 0.344 0.870 Components of NIM, NIM = REA – COL×LEA

Return on Earning Assets, REA = IR/EA 15.42 10.30 8.921 8.046 6.365 0.413 0.791 Cost of Liabilities, %, COL = IE/L 4.080 4.263 4.012 3.869 2.873 0.704 0.743 Liabilities to Earning Assets Ratio, LEA = L/EA

1.253

1.106

1.075

1.026

0.981

0.783

0.956

Source: Authors’ calculations • The book rate of return on equity (ROE), which is the most widely used and popular

measure of the bank performance results from the viewpoint of owners/investors, increased during the analysed period from 5.67% in 1994 to 12.14% in 2002, i.e. more than two times. We can also mention very high volatility of profitability ratios (both ROE and ROA) during the analysed period. Banks after-tax earnings to earnings before taxes ratio (pull-through, U), which characterises the banks tax management policy efficiency because (1 - U) = t (t - the average tax rate), also increased during this period. Banks were more skilful at finding various “tax shelters” in 1997 compared with 1994, also later. Banks’ financial leverage ratio (LEV) decreased substantially due to the central bank’s new equity requirements, which forced banks to raise equity or to merge. Financial leverage rose again in 2001 and 2002. The main factor of ROE change is the increase of the return on total assets (ROTA), which needs a more detailed analysis.

• ROTA rose from 0.81% to 1.62% between 1994 and 2002 was caused by the significant decrease of the Estonian banks’ burden (B) due to the improvement of the banks’ cost control and services pricing, also due to the substantial increase in the share of interest-earning assets in total assets. However, the net interest margin level (NIM), which reflects the interest rate spread between assets and liabilities for deposit-taking financial institutions and is the major source for the profitability of banks, has decreased substantially, from 10.31% to 3.55 %, i.e. about three times. This phenomenon also needs further analysis.

• We may draw some important and interesting conclusions from the component analysis of the substantial decrease of the NIM level:


15

(a) The average return on earning assets (REA) has fallen substantially over the recent years due to the overall falling of interest rates in the Estonian banking market, the average cost of liabilities (COL) increased slightly and fell in 2001 and in 2002 compared with 2000;

(b) REA has fallen much faster than COL, i.e. the interest spread decreased considerably over the analysed period ((15.42% - 4.08%) - (6.37% - 2.87%) = 11.34% - 3.50% = 7.84%), - this change reflects the sharpened competition between banks themselves and with other financial institutions, as for example insurance and investment funds;

(c) liabilities to earning assets ratio (LEA) has also fallen substantially, i.e. Estonian commercial banks intensified their lending and investment activities, and almost all available resources (in 2002, also a part of the equity) have been invested in the earning assets.

4. Matrix approach for banks’ performance analysis 4.1. Methodology It is possible to use the matrix model to present and analyse interrelations between various economic and financial indicators. On the basis of n quantitative indicators Yi (i = 1, 2, …, n; n - the number on initial quantitative financial indicators) it is possible to define n(n - 1) qualitative indicators, for example financial ratios

) ; ..., 2, ,1,( jinjijYiY

ijx ≠== (14)

which form the (n × n) square matrix

{ }ijx

nnxnxnxnxxxnxxx

==

21

2222111211

X (15)

and which we call the matrix model of the studied phenomenon. We can draw two important conclusions from these features of the matrix model (15): (1) as the symmetric square matrix (15) consists of two triangular matrices which are “mirror reflections” of each other (contain elements which are reciprocal to each other), the financial (or other economic) information needed for the analysis is presented only in one triangular matrix; (2) as the square matrix (15) consists of row and column vectors in linear dependence, it is sufficient for a generalized estimate of the studied phenomenon, to take into consideration only one vector. The first of these conclusions is important for the component or decomposition analysis of bank performance, and we take under observation only one triangular matrix focusing on the study and analysis of interrelations between the elements/financial ratios of the matrix model. Further steps of analysis depend on how many and which initial quantitative indicators from the balance sheet and income statements of the bank to choose for the formation of the matrix model, and which sequence to follow when including the initial indicators into the model. If we choose to start with the output or results indicators according to their degree of “finality”, and to end with the input or resource indicators by their “preliminarity”, we receive the matrix model which consists of two triangular matrices: • matrix of effectiveness, where the elements/financial ratios reflect various aspects of the

efficiency of the bank performance; • inverse matrix of effectiveness, which consists of the reciprocals of the efficiency

indicators/financial ratios.


16

A financial analysis of any bank starts with obtaining the financial data. If to follow the principles discussed here, we get the following sequence of the most important initial quantitative financial indicators: (1) Y1 - earnings after taxes, EAT; (2) Y2 - earnings before taxes, EBT; (3) Y3 - net interest revenue, NIR; (4) Y4 - interest revenue, IR; (5) Y5 - total operating income, TOI; (6) Y6 - earning assets, EA; (7) Y7 - book value of equity, BE; (8) Y8 - total assets, TA. The matrix model for a bank performance analysis is presented in Table 8.

Table 8. The Matrix Model for Banks Performance Analysis Earnings

After Taxes, Y1 , EAT

Earnings Before Taxes, Y2, EBT

Net Interest Revenue, Y3, NIR

Interest Revenue, Y4, IR

Total Operating Income, Y5, TOI

Earning Assets, Y6, EA

Book Value of Equity, Y7

Earnings Before Taxes, Y2, EBT

x21=Y1/Y2 Net Earnings to Earnings, NEER *

Net Interest Revenue, Y3, NIR

X31=Y1/Y3 Net Ear-nings to Net Interest NENIR

x32=Y2/Y3 Earnings to Net Interest Ratio, ENIR

Interest Revenue, Y4, IR

x41=Y1/Y4 Net Earnings to Interest, NEIR

x42=Y2/Y4 Earnings to Interest Ratio, EIR

x43=Y3/Y4 Net Interest to Interest Ratio, NIIR

Total Operating Income, Y5 TOI

x51=Y1/Y5 Net Earnings to Total Inco-me Ratio, NETIR *

x52=Y2/Y5 Earnings to Total Income Ratio, ETIR

x53=Y3/Y5 Net Interest to Total Income Ratio, NITIR

x54=Y4/Y5 Interest to Total Income Ratio, ITIR

Earning Assets, Y6, EA

x61=Y1/Y6 Net Return on Earning Assets, NREA

x62=Y2/Y6 Return on Earning Assets, REA

x63=Y3/Y6 Net Interest on Earning Assets, NIEA (NIM) *

x64=Y4/Y6 Interest on Earning Assets,IEA (Return on EA) *

x65=Y5/Y6 Total Income on Earning Assets, TIEA

Book Value of Equity, Y7, BE

x71=Y1/Y7 Net Return on Equity, NROE *

x72=Y2/Y7 Return on Equity, ROE *

x73=Y3/Y7 Net Interest on Equity, NIOE

x74=Y4/Y7 Interest on Equity, IOE

x75=Y5/Y7 Total Income on Equity, TIOE

x76=Y6/Y7 Earning Assets to Equity Ratio,EAER

Total Assets, Y8, TA

x81=Y1/Y8 Net Return on Assets, NREA *

x82=Y2/Y8 Return on Assets, ROA *

x83=Y3/Y8 Net Interest on Assets, NIOA

x84=Y4/Y8 Interest on Assets, IOA

x85=Y5/Y8 Total Income on Assets, TIOA *

x86=Y6/Y8 Earning Assets Ratio, EAR *

X87=Y7/Y8 Equity Multiplier, EM = 1/LEV *


17

We have to say some words about the terminology. The fact is that different authors use different definitions and terms of financial ratios for banks’ (or firms) performance analysis. Some of the most frequently used and well-known ratios are exceptions, for example, return on assets (ROA), return on equity (ROE), net interest margin (NIM). But even ROA and ROE may be calculated using the net income/earnings after taxes (EAT) or earnings before taxes (EAT). For organizing the terminology, we used the following terminological principles: (1) all qualitative indicators/financial ratios which reflect proportions among the output quantitative indicators are defined as (net) earnings or (net) interest ratios to another corresponding quantitative input indicator, for example, net interest to interest ratio, NIIR

IR

NIRNIIR

Revenue,Interest

Revenue,Interest Net = (16)

(2) all qualitative indicators/financial ratios which reflect proportions between output and input quantitative indicators, and are traditional indicators of the efficiency, are defined as (net) return, (net) interest or total income to corresponding input quantitative indicator, for example, interest on earning assets, IEA

EA

IRIEA

Assets, Earning

Revenue,Interest = (17)

(3) qualitative indicators/financial ratios which reflect proportions among input quantitative indicators are defined more traditionally, for example, as equity multiplier (EM) and earning assets ratios. More widely used financial ratios in Table 8 are marked with asterisk (*) and some of the traditionally used terms are also presented in parentheses, for example,

PMNETIRTOI

EATx ===

Income, Operating Total

Taxes,After Earnings51 (18)

4.2. General Results of Using the Matrix Model

The initial quantitative financial indicators of the Estonian commercial banking system needed for using the matrix model are presented in Table 9. The effectiveness matrix of Estonian commercial banks with all the above-described financial ratios is presented in Table 10. Actually, six different matrixes are presented in Table 10. The key for reading the financial information in Table 10 is as follows: (1) acronym of the corresponding financial ratio; (2) - (4) levels of respective financial ratios in 1994, in 2001 and in 2002; (4) relative changes in the corresponding financial ratio as the growth rates 2002/1994 and 2002/2001;

Table 9. Initial Financial Indicators for the Matrix Model (EEK m) Financial Indicators 1994 1997 2000 2001 2002 02/94 02/01 Y1, Earnings After Taxes, EAT 40.9 963.1 613.1 1683.4 1153.2 28.20 0.685 Y2, Earnings Before Taxes, EBT 68.0 1068.9 613.1 1703.8 1214.8 17.86 0.713 Y3, Net Interest Revenue, NIR 630.8 1444.1 1932.3 2182.4 2370.5 3.758 1.086 Y4, Interest Revenue, IR 943.6 2658.5 3744.2 4308.1 4253.5 4.508 0.987 Y5, Total Operating Income, TOI 1401.4 5930.5 5809.8 7203.1 6866.9 5.229 1.226 Y6, Earning Assets, EA 6117.8 25817.0 42019.6 53544.0 66827.5 10.92 1.248 Y7, Book Value of Equity, BVE 721.2 3201.2 7280.7 8179.0 9499.4 13.17 1.161 Y8, Total Assets, TA 8388.5 31763.9 52444.9 63115.0 75048.6 8.947 1.189


18

The most important element/financial ratio of the effectiveness matrix is x81 , which forms the following multiple factor system: 8776655443322181 xxxxxxxx ××××××= (19) or substituting with the definitions of corresponding financial ratios

TA

EAT

TA

BVE

BVE

EA

EA

TOI

TOI

IR

IR

NIR

NIR

EBT

EBT

EAT

EMEAERTIEAITIRNIIRENIRUNROA

=××××××=

=××××××=

(20)

So far, we have demonstrated only one possibility of using the matrix model for the banks performance analysis. There are a number of other possibilities to develop various other factor systems, to determine the absolute influence of changes in the respective financial ratios on the change of different quantitative financial indicators, to compose multi- factor aggregate index-numbers in the case of de-aggregated initial information, etc. The initial quant itative financial indicators needed for the development of the banks performance efficiency matrix, may be divided into two groups by their economic substance: • results or output indicators of the bank activities, which one may take from the income

statement: earnings after taxes (EAT), earnings before taxes (EBT), net interest revenue (NIR), interest revenue (IR), total operating income (TOI);

• resource or input indicators of the bank operating, which one may take from the balance sheet: earning assets (EA), book value of equity (BVE), total assets (TA).

We may compare the bank with any other business firm, which uses available resources/production factors/inputs (equity and borrowed external funds) for producing something useful, i.e. during the bank operating, certain inputs are transformed into certain outputs. The bank’s operating result is a production of specific financial services during the financial intermediation: credit services, securities services, transaction proceeding services, asset management services, information and financial advice offering services. All these financial services in money terms are expressed in the bank management revenues and income. Analogously with the classification of the initial financial indicators into two groups, the efficiency matrix of the bank performance analysis consists of three partial matrices: • triangular matrix, where the elements are financial ratios characterising proportions

among the quantitative output indicators, and which are by their nature co-ordination ratios: NEER, NENIR, ENIR, NEIR, EIR, NIIR, NETIR, ETIR, NITIR, and ITIR - we named this triangular matrix as “output matrix”;

• triangular matrix, where the elements are financial ratios reflecting proportions among the quantitative input indicators, and which are also typical co-ordination ratios: EAER, EAR, EM - we named this matrix as “input matrix”;

• Quadrate matrix, where the elements are financial ratios characterising proportions among different quantitative output and input indicators, i.e. these are typical intensity ratios, or traditional output/input-type efficiency indicators: NREA, REA, NIEA, IEA, TIEA, NROE, ROE, NIOE, IOE, TIOE, NROA, ROA, NIOA, IOA, and TIOA - we named this matrix as “output- input matrix” (do not confuse with Leontieff input-output type matrices).


19

Table 10. The Effectiveness Matrix of the Estonian Commercial Banking System (1994-02)

EAT 40.9 1683.4 1153.2

28.20 0.685

EBT 68.0 1703.8 1214.8

17.860.713

NIR 630.8 2182.4 2370.5

3.758 1.086

IR 943.6 4308.1 4253.5

4.5080.987

TOI 1401.4 7203.1 6866.9

4.900 0.953

EA 6117.8 53544.0 66827.5

10.92 1.248

BVE 721.2 8179.0 9499.4

13.17 1.161

EBT 68.0 1703.8 1214.8

17.86 0.713

NEER 0.6015 0.9880 0.9493

1.578 0.961

NIR 630.6 2182.4 2370.5

3.758 1.086

NENIR 0.0648 0.7714 0.4865

7.507 0.6307

ENIR 0.1078 0.7807 0.5125

4.7540.6565

IR 943.6 4308.1 4253.5

4.508 0.987

NEIR 0.0433 0.3907 0.2711

6.261 0.6829

EIR 0.0721 0.3955 0.2856

3.9610.7221

NIIR 0.6685 0.5066 0.5573

0.8337 1.1001

TOI 1401.4 7203.1 6866.9

4.900 0.953

NETIR 0.0291 0.2337 0.1679

5.771 0.7184

ETIR 0.0485 0.2365 0.1769

3.6480.7480

NITIR 0.4501 0.3029 0.3452

0.7670 1.1387

ITIR 0.6733 0.5981 0.6194

0.92001.0356

EA 6117.8 53544.0 66827.5

10.92 1.248

NREA 0.0067 0.0314 0.0173

2.5756 0.5510

REA 0.0111 0.0318 0.0182

1.63960.5723

NIEA 0.1031 0.04076 0.03547

0.3441 0.8702

IEA 0.1542 0.08046 0.06365

0.41280.7911

TIEA 0.2291 0.1345 0.1028

0.4485 0.7643

BE 721.2 8179.0 9499.4

13.17 1.161

NROE 0.0567 0.2058 0.1214

2.141 0.5889

ROE 0.0943 0.2083 0.1279

1.35610.6140

NIOE 0.8747 0.2668 0.2495

0.2853 0.9352

IOE 1.3084 0.5267 0.4478

0.34220.8502

TIOE 1.9432 0.8807 0.7229

0.3720 0.8208

EAER 8.4828 6.5465 7.0349

0.8293 1.0746

TA 8388.5 63115.0 75048.6

8.947 1.189

NROA 0.00488 0.02667 0.01537

3.1488 0.5763

ROA 0.00811 0.02700 0.01619

1.99630.5996

NIOA 0.07520 0.03458 0.03159

0.4200 0.9135

IOA 0.11249 0.06826 0.05668

0.50380.8304

TIOA 0.1671 0.1141 0.0915

0.5476 0.8019

EAR 0.7293 0.8484 0.8905

1.2210 1.0496

EM 0.0860 0.1296 0.1266

1.4723 0.9769

Source: Authors' calculations


20

4.3. Analysis of the Partial Matrices Output Matrix: Initial quantitative financial indicators, needed for building up the output matrix, are sequenced by the degree of their “finality”. The most ultimate output indicator of the bank operating is earnings after taxes (EAT) which is also the ultimate indicator in the income statement because it reflects disposable for owners generated by the bank management income. Then follows earnings before taxes (EBT), then net interest revenue (NIR), interest revenue (IR), and total operating income (TOI). Net earnings to total income ratio (NETIR = EAT/TOI) is the leading element in the output matrix. All other elements/financial ratios of the output matrix reflect in this ratio. These other financial ratios form various multiple factor systems, as for example

NETIR NEER ENIR NIIR ITIR

EATEBT

EBTNIR

NIRIR

IRTOI

EATTOI

= × × × =

= × × × = (21)

or NETIR NEER ETIR NEIR ITIR NENIR NITIR

EATEBT

EBTTOI

EATIR

IRTOI

EATNIR

NIRTOI

EATTOI

= × = × = × =

= × = × = × = (22)

One may check these and other interrelations between corresponding financial ratios. For example, using data from Table 7, growth rates (2002/1994) of financial ratios of the factor system (21) are related as follows, 5.771 = 1.578 × 4.754 × 0.834 × 0.920, i.e. earnings to net interest revenue (ENIR = EBT/NIR) very significant increase was the major factor of the level of NETIR substantial growth. It needs more exact explanation. ENIR may be expressed as

ENIREBTNIR

NIR NNIRNIR

NNIRNIR

= =+

= +1 (23)

The ratio NNIR/NIR is negative and thus ENIR shows the bank management success in the management of bank’s non- interest activities, which is in general a burden for banks operating: effectiveness of the bank cost control, success designing the price policy, success of various other financial transactions. Growth of ENIR about five times in Estonian commercial banking system was the main factor of the rising of ROA, ROE, and other rate of return indicators. One may check that the negative ratio NNIR/NIR decreased substantially during analyzed years, from 89.2% in 1994 to 48.8% in 2002. It means, that if in 1994 the most of earned net interest revenue was spent to cover net non- interest revenue, then in 2002 remarkable part of generated net interest revenue (NIR) remained also for growing shareholders profits. It is also possible to carry out component analysis of NETIR on the basis of an additive factor system. So as

NETIREATTOI

EBT TTOI

NIR NNIR TTOI

NITIRNNIRTOI

TTOI

= =−

=+ −

= + − (24)

and NNIR = EBT - NIR, and T = EBT - EAT. From Tables 4 and 7 we can obtain the absolute change of NETIR under the influence of these three factors: 1994: 0.0291 = 0.4501 + (-0.4016) - 0.0194 2002: 0.1679 = 0.3452 + (-0.1683) - 0.0090


21

absolute change: 0.1388 = -0.1049 + 0.2333 + 0.0104 And so, net earnings (earnings after taxes, EAT) have grown 13.89 cents from every generated kroon of total operating income under the influence of the following factors: (1) negative influence of the decrease of net interest revenue to total income ratio (NITIR),

- 10.49 cents; (2) substantial decrease of the negative net non- interest revenue to total income ratio

(NNIR/TOI), + 23.33 cents; (3) decrease of the tax to total income ratio (T/TOI), +1.04 cents. Expense indicators (interest expense, IE, non- interest expense, NOIE) may also be included into the factor system (24)

NETITR

EATTOI

IR IE NOIR NOIE TTOI

ITIRIE

TOINOIRTOI

NOIETOI

TTOI

= =− + − −

=

= − + − −

( ) ( )

(25)

All elements/financial ratios of the efficiency matrix (or the inverse matrix of efficiency) are interrelated with each other. There exist eight different possibilities depending on the location of related financial ratios (left, right, above, or below of each other, or one ratio is in the inverse matrix of efficiency). For example, if the question is, how net interest to interest ratio (NIIR = NIR/IR) from the output matrix, and earning assets ratio (EAR = EA/TA) from the input matrix, are related, we may build up the following multiple factor system

NIIRNIRIR

EATA

TAIR

NIREA

EAR NIEAIOA

= = × × =×

(26)

and using the factual data from Table 7: 1994: 0.6685 = 0.7293 × 0.1031 : 0.1125 2002: 0.5573 = 0.8905 × 0.03547 : 0.05668 growth rates: 0.8337 = 1.2210 × 0.3441 : 0.5038 What can we conclude from these numbers? The rise of earning assets ratio (EAR) and the decline of interest on total assets ratio (IOA) influenced positively to the relative change of net interest to interest ratio, but substantial negative relative change of net interest to earning assets ratio (NIEA) level had more relevant negative influence. Input Matrix: Only three financial ratios (earning assets to equity ratio, EAER = EA/BVE; equity multiplier, EM = BVE/TA; earning assets ratio, EAR = EA/TA) are presented in the output matrix of our example, see Table 10. All these financial ratios characterize proportions between input quantitative indicators. Total assets are the “bottom” of that matrix, and others characterize the structure of the bank’s balance sheet: earning assets from the assets side, and book value of equity from the liabilities’ side. One may discuss the sequence of including earning assets (EA) and equity (BVE) into the matrix model, but if the main task is to study and analyze efficiency of banks performance, then earning assets to equity ratio (EAER) is just the one of the efficiency partial indicators. Earning assets ratio is the leading element of the output matrix, as this is the final indicator of the multiple factor system


22

EMEAERTA

BVEBVEEA

TAEA

EAR ×=×== (27)

Using data from Tables 7 we may obtain 1994: 0.7293 = 8.4828 × 0.0860 2002: 0.8905 = 7.0349 × 0.1266 growth rates: 1.2210 = 0.8293 × 1.4723 It is possible to develop other factor systems using quantitative financial indicators from Table 7. For example

EARTA R FA

TAR

TAFATA

=− −

= − −1 (28)

where R - cash and reserves in the bank assets; FA - fixed and other assets. Using data from Table 7 we may study influence of these factors (cash and reserves ratio to total assets, and fixed and other assets ratio to total assets) to the absolute change of EAR level 1994: 0.7293 = 1 - 0.1821 - 0.0886 2002: 0.8905 = 1 - 0.0688 - 0.0407 absolute changes: 0.1612 = 0.1133 + 0.0479 It means that the increase of EAR level was achieved due to the decrease of the share of cash and reserves in total assets (+ 0.075), and also due to decrease of the share of fixed and other assets in the total assets (+ 0.002). We may once more call the readers’ attention to the fact that the development and usage of concrete financial analysis methods depend on posted tasks of the analysis. For example, if one must study the structure of the bank balance sheet assets side, analysis methods depend on, what is actually needed to study, where these weak sides of the bank activities and operating which need more detailed analysis are, etc. For example: (1) if the main task is to study general efficiency of the banks performance and the banks

management income generation possibilities, one may use approach discussed in this paper to study the structure of banks assets and liabilities, emphasizing the role of EAER, EAR, and EM;

(2) if the bank has liquidity problems and the main task of the financial analyst is to study bank’s liquidity position and related with this problems, one may start from calculating the ratio R/TA, and thereupon to develop corresponding analysis scheme/factor system for detailed analysis of factors influencing bank’s liquidity position;

(3) if credibility and confidence in the bank’s activities is weak and hurt, or the “brand name” of the bank is unknown, among other issues it may be useful to study the share of fixed and other assets in the total assets, etc.

Output-Input Matrix: The quadratic output- input matrix is the most important partial matrix of the bank performance matrix model because elements of that matrix are typical output/input-type financial ratios, i.e. efficiency indicators. There are at all 3×5=15 financial ratios (from which almost half are frequently used) in the output-input matrix presented in Table 10.


23

Net return on assets (NROA = EAT/TA) is the leading element of the whole efficiency matrix, which forms the following multiple factor system from all elements of the main diagonal

NROA NEER ENIR NIIR ITIR TIEA EAER EM

EATEBT

EBTNIR

NIRIR

IRTOI

TOIEA

EABVE

BVETA

EATTA

= × × × × × × =

= × × × × × × = (29)

We may check easily the factor system (29) 1994: 0.6015 × 0.1078 × 0.6685 × 0.6733 × 0.2291 × 8.4828 × 0.0860 = 0.0048 2002: 0.9493 × 0.5125 × 0.5573 × 0.6194 × 0.1028 × 7.0349 × 0.1266 = 0.0154 growth rates: 1.578 × 4.574 × 0.8337 × 0.920 × 0.4485 × 0.8293 × 1.4723 = 3.15 Naturally, it is possible to aggregate or de-aggregate elements/financial ratios included in the multiple factor system (29), to include elements from the inverse matrix of efficiency etc. I.e. in principle it is possible to develop a number of different factor systems depending on posted tasks of analysis and established bottlenecks in the bank activities, operating and performance. It is also possible to start with the analysis of other financial ratios in the output-input matrix. For example, from Table 10 we can see that net interest on assets ratio (NIOA = NIR/TA) in Estonian commercial banking system is fallen substantially (more than twice) during the analysed period, which phenomenon needs further analysis. We may develop the following factor system from the elements of the main diagonal of the matrix model

NIOA NIIR ITIR TIEA EAER EM

NIRIR

IRTOI

TOIEA

EABVE

BVETA

NIRTA

= × × × × =

= × × × × = (30)

Using data from Table 7 we may obtain 1994: 0.07520 = 0.6685 × 0.6733 × 0.2291 × 8.4828 × 0.0860 2002: 0.03159 = 0.5573 × 0.6194 × 0.1028 × 7.0349 × 0.1266 growth rate. 0.4200 = 0.8337 × 0.9200 × 0.4485 × 0.8293 × 1.4723 Consequently, all factors from the output matrix and from the output-input matrix negatively influenced the change of NIOA: net interest to interest ratio (NIIR=NIR/IR) decreased 16.63%, interest to total income ratio (ITIR = IR/TOI) decreased 8.00%, and total income to earning assets ratio (TIEA = TOI/EA) decreased 55.15%. Financial ratios, which characterize proportions between input indicators, changed as follows: earning assets to equity ratio (EAER = EA/BE) decreased 17.07%, and only equity multiplier (EM = BVE/TA) increased substantially, 47.23%. It is suitable to aggregate some financial ratios. For example, as NIIR × ITIR × TIEA = NIEA =NIR/EA, and EAER × EM = EAR = EA/TA, we receive the following factor system:

NIOA NIEA EARNIREA

EATA

NIRTA

= × = × = (31)

Using data from Table 10: 1994: 0.0752 = 0.1031 × 0.7293 2002: 0.03159 = 0.03547 × 0.8905 growth rates: 0.4200 = 0.3441 × 1.2210


24

Consequently, efficiency of Estonian commercial banking system from the point of view earned interests has fallen substantially during the last four years. Positive changes in the proportions of input indicators slowed down this falling a little. Now we turn back to the analysis of the leading element of the output-input matrix, i.e. to the analysis of NROA dynamics. It is suitable to aggregate financial ratios from the output matrix and from the input matrix correspondingly to formulae (21) and (27)

NROA NETIR TIEA EAREATTOI

TOIEA

EATA

EATTA

= × × = × × = (32)

Only one output/input-type traditional efficiency indicator, TIEA = TOI/EA, is added to the leading elements of the output matrix and input matrix. Results are as follows:

1994: 0.0049 = 0.0291 × 0.2291 × 0.7293 2002: 0.0154 = 0.1679 × 0.1028 × 0.8905 growth rates: 3.15 = 5.771 × 0.4485 × 1.221

Hence, the growth of net return on assets (NROA = EAT/TA) more than three times was caused by the influence of following changes: (1) favorable changes in the proportions of banks output indicators, which are expressed in the

5.771 times rise of the level of NETIR, were the main “engine” of the very substantial rise of NROA;

(2) there was also small positive change in the proportions of banks input indicators, which became apparent in the 22.1% rise of EAR;

(3) the only typical and traditional output/input-type efficiency indicator, TIEA = TOI/EA, in the factor system (42), decreased substantially (growth rate 0.4485).

Rise of the general level of the efficiency of Estonian commercial banking system performance during the analyzed period (1994-2002) seems at first very efficacious. This reflects in the increase of NROA, but it was mainly caused by changes in the proportions of output indicators (mainly ENIR about 4.75 times rise). We are of the opinion that such phenomenon is not relevant in the long-term perspective. Traditional output/input-type efficiency ratios, as NIEA, IEA, NIOA and others, have been decreased substantially and there are other major sources for rising banks operating efficiency in the future. Now we present the analysis of the absolute change of NROA from entirely other aspect. It is possible to develop the following additive factor system

NROA

EATTA

EBT TTA

NIR NNIR TTA

IR IE NNIR TTA

IOAIETA

NNIRTA

TTA

= =−

=+ −

=− + −

=

= − + −

( ) ( )

(33)

Using data from Tables 6 and 10 we may find: 1994: 0.00488 = 0.11249 - 0.03729 + (-0.06709) - 0.00323

2002: 0.01537 = 0.05668 - 0.02509 + (-0.01540) - 0.00082 absolute changes: 0.01049 = -0.05581 + 0.01220 + 0.05169 + 0.00241 Expressing these results in Cents for one Estonian crown (EEK) of total assets and rounding these results a little, we get the absolute change of NROA to 1.05 cents from every EEK of assets in the result of following changes: (1) decrease of interests on assets ratio (IOA = IR/TA), - 5.58 cents;


25

(2) decrease of interest expense to total assets ratio (IE/TA), + 1.22 cents; (3) decrease of net non- interest revenue to total assets ratio (NNIR/TA), + 5.17 cents; (4) decrease of taxes to total assets ratio (T/TA), + 0.24 cents. As it was mentioned already earlier, the main source of the rise of net return on assets was decrease of banks burden. The burden for banks is their non-interest activities what has always negative results. The measure of banks burden was in the last discussed case expressed as net non-interest revenue to total assets ratio, NNIR/TA. Estonian commercial banks major activities (lending and investing) results have been worsen during the last two years. Now let us look how traditional financial ratio analysis indicators are represented in the matrix model. First of all, more widely used rate of return type financial ratios, as return on assets (ROA) and return on equity (ROE), also related with them financial leverage (LEV) and equity multiplier (EM), are directly presented there (LEV = 1/EM). There is also all needed information for the usage of DuPont financial ratio analysis different versions. ROE decomposition analysis indicators are all directly in the matrix model, see formulae (1)-(5). There is also information for carrying out component analysis of ROTA, see formulae (6)-(9). As net non- interest revenue NNIR = EBT - NIR, it is possible to calculate banks burden, B = NNIR/TA. Also it is not difficult to carry out NIM component analysis, see formulae (10)-(13) because all needed initial information is in the efficiency matrix. As interest expense IE = NIR - IR and liabilities L = TA - BE, we may calculate cost of liabilities COL = IE/L and liabilities to earning assets ratio LEA = L/EA. We may conclude that the list of initial quantitative financial indicators in Table 9 and the efficiency matrix in Table 10 contain enough information for carrying out different versions of traditional financial ratio analysis, including DuPont analysis versions. But, in addition, there is much more information for making more profound bank performance analysis, from which possibilities we illustrated in this paper only partly. And so, most important financial information, which is possible to obtain from the balance sheet and the income statement of the bank, is in more complex and aggregated form presented in the matrix model (Table 10). 5. Production functions of the Estonian banking It is possible to use the existing information to construct production functions of banking, treating banking as a separate sector with its own inputs and outputs. We selected total income of the banks (y) as the output variable, and used fixed assets ( 1x ), earning assets ( 2x ), liabilities ( 3x ), and equity ( 4x ) as factors. We used financial statements’ quarterly data for individual Estonian banks, from the fist quarter 1995 to the second quarter 2003, as these were published in the Bulletins of the Bank of Estonia. The strength of the relationships between the selected variables is presented in Table 11, mainly due to existing of trends in all observed variables. We chose the Cobb-Douglas function as the type of the production function. .1, ≠+= βαβα zaxy (34)

To estimate the parameters α and β of the function (34), using ordinary least square method (OLSM) estimators, it is first necessary to find logarithms of the primary data. Thereafter we analyzed the trends of the logarithms. The following proved to be the best.


26

Table 11. The strength of the relationships between the selected variables Y X1 X2 X3 X4 Y 1 X1 0.460073 1 X2 0.771868 0.70662 1 X3 0.76014 0.73396 0.998335 1 X4 0.694385 0.843456 0.951603 0.95453 1

Source: The authors’ calculations ;054,01515,6ln ty += (35) ;ln3461,01060,7ln 1 tx += (36) ;0632,03456,9ln 2 tx += (37) ;0676,04188,9ln 3 tx += (38) .8623,02014,6ln 4 tx += (39)

Deviations from the trends were used as new primary data in constructing production functions. The parameters of the production functions were found applying Kramer’s method. The following results were obtained: ;4025,0 )ln3332,00279,0(0392,0

39628,0

1ttexxy −= (40)

;0427,0 0092,00176,03

9324,02

texxy −= (41)

;1801,0 )ln3987,00034,0(4624,04

5376,02

ttexxy −= (42) .5029,0 )ln7227,00427,0(8381,0

41689,0

3ttexxy −= (43)

All equations are statistically significant according to the F-criterion; however, we

regard equation (42) as the best as it has the smallest standard error. As the models show, earning assets are the most important factor for the total income. This is quite logical as the earning assets are the resource for lending, and that is from which the banks earn their income. The exponential functions in the models indicate that the growth rate of incomes is decelerating. The above analysis shows that it is possible to treat banking as an independent sector of the economy of a country and that econometric modeling yields interesting results (possible combinations of factors). 6. Conclusions (1) The development of the Estonian banking sector can be described by a quite rapid

nominal growth of total assets, loan portfolios, net income and other quantitative financial indicators. Although the Estonian banking market was already quite concentrated, the consolidation process continued. The capitalization of Estonian banks improved, and the share of non-residents in the share capital increased significantly.

(2) The banks performance analysis is an important issue in the conditions of transition economies in the Central and Eastern European countries because the financial sector could play the key role in a successful transition. The balance sheet and income statement of the bank is the major source for carrying out the bank performance analysis. A modified version of the DuPont financial ratio analysis and the matrix approach, which enable to follow interrelations between different financial indicators, seem to be more perspective.


27

(3) Some empirical results of the usage of both the DuPont financial ratio analysis and the matrix model for the Estonian commercial banking system in 1994-2002 are presented in the article. As the Estonian banking system is developing rapidly, both input and output quantitative financial indicators have increased substantially during the recent years. There was an overall falling of the market-determined interest rates in the Estonian banking market, the interest spread decreased substantially, which influenced the dynamics of various discussed financial ratios.

(4) The rise of the Estonian commercial banking system performance efficiency, which is revealed in the increase of the rate of return indicators such as return on assets (ROA) and return on equity (ROE), was caused mainly by the changes in the proportions between output indicators (for example, the banks’ burden has decreased substantially). The traditional output/input-type efficiency ratios (interest or income on assets or on equity ratios) however, decreased substantially during the analyzed period.

(5) The initial quantitative financial indicators and the respective matrix model used in this article contain all relevant information for carrying out different versions and modifications of the traditional financial ratio analysis, including DuPont financial ratio analysis. In addition, there is much more information for carrying out a more profound banks performance analysis, the possibilities of which we partly illustrated in this article.

(6) Application of econometric models (production functions) in the analysis of banking information provides new opportunities to combine factors in analyzing and forecasting output variables.

References

Abiad, Abdul and Ashoka Mody. 2003. Financial Reform: What Shakes It? What Shapes It? IMF Working Paper No. 03/70. Washington, DC: IMF.

Andersen, Atso, Ari Hyytinen and Jussi Snellman. 2000. Recent Developments in the Finnish Banking Sector. Bank of Finland Discussion Papers, 15/2000. Helsinki: Bank of Finland.

Bank of Estonia: Annual Reports. Tallinn: Bank of Estonia, 1994-2003. Berger, Allen N. and David B. Humphrey. 1997. Efficiency of Financial Institutions:

International Survey and Directions for Future Research. - European Journal of Operational Research, 98, 175-212.

Cole, David. 1973. Return on Equity Model for Banks. - The Bankers Magazine, Spring.

Dietrich, J. Kimball. 1996. Financial Services and Financial Institutions. Prentice Hall. Estrella, Arturo, Sangkyun Park, and Stavros Peristiani. 2000. Capital Ratios as

Predictors of Bank Failure. Economic Policy Review, Federal Reserve Bank of New York, 6, 2, 33-52.

European Central Bank. 1999. Possible Effects of EMU on the EU Banking Systems in the Medium to Long Term. Frankfurt: ECB, February.

European Central Bank. 2000. EU Banks’ Income Structure. Frankfurt: ECB, April. Hardy, D. C. and E. Bonaccorsi di Patti. 2001. Bank Reform and Bank Efficiency in

Pakistan. IMF Working Papers 01/138. Washington DC: IMF. Harm, Christian. 2002. Bank Management between Shareholders and Regulators.

SUERF Studies No. 21. Vienna: SUERF. Hasan, Iftekhar and Katherin Marton. 2000. Development and Efficiency of the

Banking Sector in a Transitional Economy: Hungarian Experience, BOFIT Discussion Paper 7/2000. Helsinki: Bank of Finland.


28

Hyytinen, Ari and Tuomas Takalo. 2003. Preventing Systemic Crises through Bank Transparency. Bank of Finland Discussion Papers No. 25/2003. Helsinki: Bank of Finland. Hyytinen, Ari and Tuomas Takalo. 2002.Enhancing Bank Transparency: A Reassessment. – European Finance Review, 6, 429-445.

Kaminsky, Graciela L. and Carmen M. Reinhart. 1998. Financial Crises in Asia and Latin America: Then and Now. - American Economic Review, 88, 2, 444-448.

Kaminsky, Graciela L. and Carmen M. Reinhart. 1999. The Twin Crises: The Causes of Banking and Balance-of-Payments Problems. - American Economic Review, 89, 2, 473-500.

Kose, John and Yiming Qian. 2003. Incentive Features in CEO Compensation in the Banking Industry. – Federal Reserve Bank of New York Economic Policy Review, 9, 1, 107-121.

Krzak, Maciej. 1998. Large Current Account Deficits – The Case of Central Europe and the Baltic States. - Focus on Transition, 1/1998, Oesterreichische Nationalbank, 22-46.

Kwan, Simon H. 2003. Operating Performance of Banks among Asian Economies: An International and Time Series Comparison. - Journal of Banking and Finance, 27, 3, 471-489.

Kwan, Simon H. 2002. The X-Efficiency of Commercial Banks in Hong Kong. Working Papers of the Federal Reserve Bank of San Francisco, WP 02-14.

Leemets, Maris and Reet Reedik. 2003. Rating Agencies’ Assessments of Estonia and Other Acceding States. Eesti Pank Quaterly, Kroon&Economy, No. 2, 48-53.

Llewellyn, David T. Comment. 2002. In G. G. Kaufmann, ed. Prompt Corrective Actions: Ten Years Later. Amsterdam: JAI , pp. 321-333.

Llewellyn, David T. and David G. Mayes. 2003. The Role of Market Discipline in Handling Problem Banks. Bank of Finland Discussion Papers 21/2003. Helsinki: Bank of Finland.

Macay, Jonathan R. and Maureen O’Hara. 2003. The Corporate Governance of Banks. – Federal Reserve Bank of New York Economic Policy Review, 9, 1, 91-107.

Mörttinen, Leena. 2002. Banking Sector Output and Labour Productivity in Six European Countries. Bank of Finland Discussion Papers 12/2002. Helsinki: Bank of Finland.

Rebelo, Joao and Victor Mendes. 2000. Malmquist Indices of Productivity Change in Portuguese Banking: The Deregulation Period. - International Advances in Economic Research, 6, 3, 531-543.

Spiegel, Mark M. 1999. Moral Hazard under the Japaneze “Convoy” Banking System. - Economic Review, Federal Reserve Bank of San Franc isco, no. 3, 3-13.

Stiroh, Kevin J. 2000. Compositional Dynamics and the Performance of the U.S. Banking Industry. Federal Reserve Bank of New York Staff Reports, No. 98.

Sundararajan, V., Charles Enoch, Armida San José, Paul Hilbers, Russell Krueger, Marina Moretti and Graham Slack. 2002. Financial Soundness Indicators: Analytical Aspects and Country Practices. IMF Occasional Paper No. 212. Washington DC: IMF.

Van Greuning, Hennie and Sonja Brajovic Bratanovic. 2000. Analyzing Banking Risk: A Framework for Assessing Corporate Governance and Financial Risk Management. Washington D.C.: World Bank.

Vensel, Vello. 2001. Estonian Banking System Performance, 1994-2000. In: Vensel, Vello and Clas Wihlborg, eds. Estonia on the Threshold of the European Union: Financial Sector and Enterprise Restructuring in the Changing Economic Environment. Tallinn: Tallinn Technical University, 23-43.

Vensel, Vello. 1997. Banking System Development in Estonia. In: T. Kowalski, ed. Financial Reform in Emerging Market Economies, Poznan: Poznan University of Economics, 141-164.

bank performance analysis- methodology and empirical evidence

Documents