Journal Issue
Share
Article

Republic of Slovenia: Selected Issues

Author(s):
International Monetary Fund
Published Date:
May 2007
Share
  • ShareShare
Show Summary Details

II. Efficiency of Slovene Banking Sector in the EU Context1

Summary

  • The Slovene banking sector is dominated by a few large, state-owned banks and has low profitability by regional standards. Its performance will be increasingly tested by deepening EU financial integration and capital market development.

  • To assess Slovene banks’ readiness to face these challenges, the paper analyzes bank efficiency and contestability in Slovenia. It finds that Slovene banks are among the least efficient in Europe, which can reflect the low contestability compared to EU peers. Efficiency and contestability may have been influenced by market concentration and ownership.

  • As state banks are the least efficient, privatization and other measures to increase banks efficiency could have important benefits.

A. Introduction

7. European integration poses new challenges for Slovene banks. The sector remains dominated by domestic, largely state-controlled banks. Competition has increased recently because of the presence of smaller foreign-owned banks and deepening integration with European financial markets since EU accession in 2004 and euro adoption in 2007. As a result, interest margins are declining, which, together with the loss of exchange rate revenues, is squeezing the already regionally low profitability. This raises questions about the efficiency of the Slovene banking sector in the European context, market contestability, and the ability of Slovene banks to function in a more competitive integrated market.

8. To assess Slovene banks’ readiness to face these challenges and identify potential policy measures, the paper analyzes how efficient Slovene banks are in the EU context. Using detailed bank-specific data, the paper develops quantitative and qualitative indicators of bank efficiency and market contestability for Slovenia and compares them with regional competitors.

9. Results show that the performance of the Slovene banking sector is lagging its regional rivals. Slovene banks are on average less efficient and less profitable than those in the EMU and new member states (NMS), and the sector is among the least contested in Europe. Although the paper does not estimate determinants of efficiency and contestability, possible explanations can be the dominance of one large bank and the important share of state ownership in the sector, especially given that the large state-dominated banks are the least efficient. This situation points to potentially large benefits from further privatization and measures to enhance local market competition.

10. The paper is organized as follows.Section B analyzes indicators of bank efficiency by comparing performance indicators for banks in Slovenia, EMU and NMS. Section C presents results from cross-country econometric estimates of banking sector cost efficiency, while Section D reports results from estimates of cross-country banking sector contestability. Finally, Section E discusses determinants of efficiency and contestability.

B. Qualitative Indicators of Bank Efficiency

11. The analysis covers commercial banks in Slovenia and in comparable EMU and NMS countries. The choice of appropriate comparators is crucial, as indicators of bank structure and performance vary systematically with factors such as the size of the market and bank’s core business (ECB, 2006). To address such concerns, the study is limited to commercial banks in the smaller EMU markets and in the NMS.2 Slovene commercial banks are divided into two groups—the three largest state-dominated banks and all other.

12. The study draws on bank-level balance sheet and income statement data from the BankScope database. Given the difficulties in evaluating and comparing bank performance from a single perspective, the study relies on both qualitative assessment of banking sectors’ structure and performance, and quantitative estimates of banking sector cost efficiency and contestability. Because of limited availability of historical data, the main emphasis is on cross-section comparisons rather than time trends. For further details on the sample data see Appendix I.

13. Cost indicators point to Slovenia’s lagging behind comparators in improving its relatively low efficiency. Cost-to-income ratios in the NMS have been falling toward EMU levels in recent years, as banks are increasing efficiency. In Slovenia, these ratios have remained higher, especially in the three largest, state-dominated banks.3 Other cost-efficiency indicators, such as the ratio of noninterest expenses to average assets, show a similar pattern. However, this result may be biased by the large differences in total banking sector asset size between the groups.

Cost-to-Income Ratios

(percent)

Source: Bankscope.

14. The higher cost-to-income ratios in Slovenia are driven by labor costs. The share of labor costs in total costs in both the NMS and Slovenia has increased in recent years. while the share in the NMS remains at 80 percent of EMU levels, in Slovenia it is only 5 percent below the EMU average.4 In general, this is to be expected, because, when income and wages in NMS catch up with those in the EMU, labor costs shares should also converge. However, in Slovenia the large labor cost share points to overstaffing as wages remain regionally low. Personnel expenses per employee in Slovenia in 2005 were about one third of those in the EMU and broadly similar to those in the NMS. This suggests that the Slovene banking sector is overstaffed in a regional comparison - despite low wages, the labor share in operating costs is already at the level of the EMU average. Cost efficiency problems are also more pronounced in the three largest state-dominated banks.

Labor Share in Operating Costs

(percent)

Sources: Bankscope.

Personnel expenses per employee, 2005

Source: Bankscope.

15. Profitability indicators also show that Slovene banks lag regional comparators. The share of pre-tax profits in total operating income and the return on average equity (ROAE) in Slovenia was considerably lower than in the NMS and the EMU in 2005, and showed no increase from previous years. This is partly explained by stricter regulations on loan loss provisions in Slovenia until 2006, when the adoption of International Financial Reporting Standards (IFRS) harmonized provisioning rules. Preliminary estimates show that provisions in Slovene banks decreased from 13 to 8 percent of operating income during 2005–06. However, profitability in 2006 still remained below EMU and NMS levels.

Return on Average Equity

Source: Bankscope.

Operating Costs, Provisions and Profits in Operating Income, share of total operating income

Source: Bankscope.

16. Profits in Slovenia have been affected by declining net interest margins. while these have been broadly flat over the past five years in the EMU and the NMS, they have declined in Slovenia over the same period by around 2 percentage points. This can mean that competition has only recently increased in Slovenia when it joined the EU in 2004 and after interest rates started to converge toward EU levels in anticipation of euro adoption in 2007. By end-2006, net interest margins in Slovenia had decreased further, but remained about 1 percentage point above the EMU average. This suggests that pressure on profitability of Slovene banks will continue.

Net Interest Margins

Source: Bankscope.

Average Net Interest Margin, 2005

Source: Bankscope.

17. In sum, these indicators suggests that the Slovene banking sector is less efficient and profitable than its regional rivals. The problems appear more pronounced in the larger state-dominated banks. The still high interest margins suggest that poor performance may reflect problems with market contestability. To complement these simple comparisons of cost efficiency and profitability, bank efficiency in Slovenia is also analyzed econometrically. This allows for a better control for the impact of size, input costs and business models on performance.

C. Estimates of the Distance of Slovene Banks from a Cost Efficiency Frontier

18. Banking sector efficiency can be assessed by a stochastic “best practices” frontier analysis. This approach estimates indirect levels of costs for a given level of outputs and prices of inputs.5 In line with intermediation approach to banking, assumes that bank’s output, represented by interest earning assets, is produced using labor, capital and funds as inputs. One can then define a total cost frontier, characterized by TCi = f(yi, pik) where pik is a vector of input prices and Yi represents outputs. Total cost equation is used to estimate the cost frontier for the whole sample and distance from the frontier for each of the sample banks.

19. Benchmark estimates are based on the following specification:

where Yi is output captured with total earning assets; pL is personnel expenses divided by the number of employees and is a proxy for the price of labor; pK is other operating and administrative expenses divided by fixed assets and is a proxy for the input price of equipment and fixed capital; pF is interest expenses divided by all funding and proxies the price of funding; and TCi is total costs obtained as the sum of interest expenses and total operating expenses.6 The error term consists of two components: a two-sided random noise component, vi, and nonnegative cost efficiency component, ui. The measure of cost efficiency is provided by CEi = exp{-u}, where CEi = 1 represents the efficiency frontier. The specification imposes homogeneity in prices by normalizing input prices with the price of funds, piF. To minimize heteroscedasticity in the error term, output and total costs are normalized with equity.

20. The results confirm that Slovene banks are on average less efficient than those in the EMU and NMS (left panel in text figure and Table 1). The estimation also shows that there are no clear differences in efficiency between EMU and NMS banks.

Table 1.Summary of Stochastic Frontier Estimates of Cost Functions
Assumed Bank Output:

Earning Assets
Assumed Bank Output:

Loans, Deposits, Other

Earning Assets
Log likelihood−2.42−74.7
sigmaU/sigmaV2.68 (0.04)2.23(0.08)
sigma20.14(0.02)0.26(.004)
sigmaU2/sigma20.8780.832
Mean efficiency0.8260.756
Sources: Bankscope; and staff calculations.Notes: Log likelihood reports the value of log likelihood function; sigmaU is standard deviation of the inefficiency component of disturbance: sigmaV is standard deviation of the random component of disturbance; sigma2 is variance of the composite disturbance; where applicable standard errors are shown in parentheses.
Sources: Bankscope; and staff calculations.Notes: Log likelihood reports the value of log likelihood function; sigmaU is standard deviation of the inefficiency component of disturbance: sigmaV is standard deviation of the random component of disturbance; sigma2 is variance of the composite disturbance; where applicable standard errors are shown in parentheses.

Average Bank Efficiency Scores by Country, 2005

(Distance from cost frontier, frontier=1)

Source: Staff calculations.

21. The results are robust to alternative model assumptions and regression specifications. The results are broadly unchanged with an alternative specification for the production function, whereby the output of the sector is characterized by loans, deposits, and other earning assets rather than by earning assets alone (right panel of text figure).7 A common cost frontier was also estimated for 2000-05 using time dummies for each year and showed similar results for average banking sector efficiency. Other specifications examined included (i) separate cost frontier estimates for the NMS and the EMU; (ii) the addition of loans/assets, deposits/assets and equity /assets ratios as control variables; (iii) the inclusion/exclusion of commission and fees from interest income and expenses; and (iv) the normalization of the price of capital with fixed or total assets. These all gave broadly similar results about banking sector efficiency in Slovenia.

22. The estimates are also consistent with other efficiency indicators. Cost-to-income ratios already showed that Slovene banks are on average less efficient than banks in the NMS and the EMU, while in general there is no clear difference between the two regions. Also, in line with results from the previous section, the average efficiency in the three biggest Slovene banks is below the banking sector average.

D. Market Contestability

23. Market contestability is analyzed using an index that measures the extent of the pass-through of changes in input prices to revenues in banks. The methodology was developed by Panzar and Rosse (1987) (hereafter PR), who measure contestability based on a concept that, under certain assumptions8, bank revenues under perfect competition increase by the same amount as input prices, with output quantity staying constant and output prices increasing proportionally. With less-than-perfect competition, the pass-through to output prices is less than one-to-one and decreases with lower competition. Thus, market contestability is measured by an estimate of the response of output prices to changes in input prices. This methodology is then used to estimate a competitiveness index for each market in the sample, and markets are ranked according to contestability. In line with benchmark estimates of market efficiency, the study assumes that banks are in the business of producing interest-earning assets using capital, labor, and interest-bearing funds as inputs. Based on the banks’ revenue equation, the contestability of the market can then be inferred from the H−statistic, which measures the extent to which changes in factor prices are reflected in revenues. If the H−statistic assumes a value of 1, the market is perfectly competitive, while, in case of monopolistic competition, the values of the H-statistic are between 0 and 1.

24. The benchmark estimates of market contestability are based on the following specification:

where pY is interest income divided by all earning assets and proxies the price of bank output; pL, pK, and pF are input prices already defined in the previous section; and the H−statistic is defined by H=j={L,K,F,}βj

25. Results show that the Slovene banking sector is among the least contested among the EMU and NMS. for the total sample period, 1995-2005, market contestability was the lowest in Slovenia and several NMS (Lithuania, Hungary, and Slovak Republic), and EMU markets were generally more competitive, with Belgium, Austria, and the Netherlands having the highest scores. Further details are presented in Tables 2 and 3.

Index of Market Contestability

(Benchmark specification)

Source: Staff calculations.

Table 2.Details of Baseline Regression Results for Market Contestability
CountryβF|t-stat|βL|t-stat|βK|t-stat|Const|t-stat|No. of obs.R2
AUT0.6418.84**0.000.090.010.44−0.823.83**1920.66
BEL0.6711.88**−0.010.290.041.66−0.702.75**1700.46
CZE0.4210.19**−0.284.58**0.000.12−0.633.67**1360.61
EST0.758.07**−0.222.34*0.070.980.642.33*440.82
ESP0.629.83**−0.110.95−0.040.98−0.360.9530.77
FIN0.619.07**−0.040.430.020.43−0.851.85440.68
GRC0.5416.90**−0.213.62**−0.031−0.160.92920.88
HUN0.497.46**−0.223.34**0.041.4−0.160.61710.55
IRL0.7623.89**−0.225.93**−0.063.06**0.442.09*1100.86
LTU0.599.20**−0.253.75**−0.03−0.480.060.33720.82
LVA0.6514.65**−0.112.38*−0.092.27*−0.080.451660.65
NLD0.7428.60**−0.051.09−0.052.98**−0.321.671710.86
POL0.6714.10**−0.112.91**−0.041.25−0.100.851110.81
PRT0.6713.93**−0.255.52**0.051.830.351.691010.75
SVN0.5815.43**−0.315.75**0.031.390.271.531240.76
SVK0.5014.99**−0.102.06*−0.073.62**−0.777.97**1020.86
Sources: Bankscope; and staff calculations.Notes: Dependent variable: interest income to total earning assets. * significant at 5 percent; ** significant at 1 percent.
Sources: Bankscope; and staff calculations.Notes: Dependent variable: interest income to total earning assets. * significant at 5 percent; ** significant at 1 percent.
Table 3.Baseline Competitiveness Index and Tests of Long-Run Equilibrium
CompetitivenessTest of LR Equilibrium, i.e.
CountryindexβF+βL+βK=0
βF+βL+βK|t-stat|βF+βL+βK|t-stat|No. of obs.
AUT0.6511.13**0.0030.63192
BEL0.708.94**−0.0102.70*170
CZE0.151.770.0433.57**136
EST0.613.70**−0.0281.3444
ESP0.472.90**−0.0020.3553
FIN0.596.21**0.0020.4444
GRC0.303.47**0.0222.70*92
HUN0.313.12**−0.0060.8071
IRL0.4810.24**−0.0041.45110
LTU0.312.50*−0.0391.6372
LVA0.454.97**−0.0251.84166
NLD0.6411.43**0.0010.17171
POL0.536.67**0.0030.71111
PRT0.476.36**0.0092.31*101
SVN0.304.04**−0.0030.44124
SVK0.324.38**0.0050.67102
Sources: Bankscope; and staff calculations.Notes: Competitiveness index is calculated from regression results in Table 2. Test for long-run equilibrium estimates the response of pre-tax profits to changes in input prices. A banking sector is in long-run equilibrium if the sum of the response of pre-tax profits to input prices is not significantly different from zero, i.e., βF+βL+βK=0. Since profits can be negative, the dependent variable for estimates of the long-run equilibrium is defined as ln(1+pre-tax profits/total assets). * significant at 5 percent; ** significant at 1%.
Sources: Bankscope; and staff calculations.Notes: Competitiveness index is calculated from regression results in Table 2. Test for long-run equilibrium estimates the response of pre-tax profits to changes in input prices. A banking sector is in long-run equilibrium if the sum of the response of pre-tax profits to input prices is not significantly different from zero, i.e., βF+βL+βK=0. Since profits can be negative, the dependent variable for estimates of the long-run equilibrium is defined as ln(1+pre-tax profits/total assets). * significant at 5 percent; ** significant at 1%.

26. The results are robust to alternative specifications of the revenue function and other alterations. Robustness was tested by (i) adding time dummies and loans/assets deposits/assets and equity/assets ratios as control variables;9 (ii) including commission and fee income in the calculation of the price of funds and output; (iii) changing the normalization of the price of capital from fixed assets to total assets; and (iv) performing estimates with OLS and GLS with fixed bank-specific effects. Each of the resulting 16 specifications has its appeal. Although results for several banking sectors exhibited significant variations, the majority of the specifications showed that the Slovene banking sector was among the least contested in Europe. Also, in line with the baseline results, the average contestability index was higher in the EMU than in the NMS.10 The results were also robust to a shortening of the sample period to 1999-2005.

Robustness of Regression Results for Market Contestability

Source: Staff calculations.

27. Tests detect no significant variation in contestability over time. Given the limited number of observations for each year, estimates of an annual trend in contestability index are not feasible. However, estimates for three subperiods with time dummies - 1995-99, 2000-02 and 2003-05 - show no evidence of a statistically significant variation in the H−statistic. This is somewhat surprising, as one would expect that greater competition from deepening EU integration would have increased contestability in Slovenia. This outcome can, however, reflect the fact that the sample ends in 2005, and the financial integration has accelerated only recently.

28. The findings are consistent with similar estimates in the literature.Bikker and others (2006) report H−statistics for a large set of countries, including the EMU and the NMS, that show that Slovenia ranks low in contestability - 10th, 14th and 15th out of 16, depending on the model specification. Their results also confirm that market contestability in the EMU is greater than in the NMS and that Slovenia lags both country groups.

E. Determinants of Efficiency and Contestability and Policy Issues

29. The results raise questions about the determinants of efficiency and contestability. Data limitation did not allow testing for this formally for Slovenia. However, simple correlations show a negative correlation of-0.74 between the contestability index and the net interest margin of a banking sector, and a positive correlation of 0.61 between the estimates of contestability and cost efficiency. This points to a possible link between efficiency and contestability, in that lack of competition would have led to low efficiency.

30. Market entry and activity restrictions, structure of ownership, bank size, and market concentration have been the main determinants of bank efficiency and contestability in larger cross-country studies.11 Although some of the empirical findings are conflicting, the studies have generally found that market entry and activity restrictions reduce market contestability and, thereby, efficiency. State ownership is also found to have a negative effect on bank efficiency and market contestability, while the opposite holds for foreign ownership. Bank size and market concentration also affect efficiency and contestability negatively, although the empirical evidence here is less clear cut.

31. The Slovene banking sector has faced many of these constraints, which may have contributed to low efficiency and contestability. The share of direct and indirect state ownership of banks in Slovenia is high by regional standards. For example, the government owns directly 35 percent of the largest bank, while it holds indirectly another 16 percent through state-owned investment funds and nonbank corporations. The Slovene banking sector is also highly concentrated, with one large, state-controlled bank accounting for 40 percent of total assets—one of the highest ratios in Europe. The Herfindahl index12 for 2005 also ranks Slovenia as the 6th most concentrated among 16 sample countries (ECB, 2006). This situation may have discouraged bank entry and limited competition in the sector.

32. Improving the efficiency and contestability of banks in Slovenia would enhance growth and financial stability. As growth is becoming increasingly dependent on productivity, a more contestable and efficient banking sector that intermediates finance to the most efficient uses is important for competitiveness and continued convergence toward EMU income levels. Given the underperformance of the sector to date, with greater efficiency its contribution to growth can be substantial.13 A more efficient banking sector can also better deal with a rapidly aging population by providing more sophisticated financial products for pension savings. Greater efficiency and profitability would also help reduce vulnerabilities stemming from credit and market risks.14

33. Going forward, European integration, capital market development, and the role of the state will all be important for the performance of banks. Deeper EU integration and capital market development will no doubt add to competitive pressures and force Slovene banks to improve efficiency. The government is also planning to further privatize banks, which can enhance efficiency and profitability. However, as the government is keen to retain majority shares in the key banks, more active measures may be needed to ensure that efficiency continues to improve. Measures, for example, to raise corporate governance standards to EU levels for board members, define a longer-term growth strategy in banks in which the state remains a dominant shareholder, and to list these banks in the stock exchange, can increase management accountability and transparency of operations. Strong bank supervision should also be maintained to monitor bank performance and risks.

Appendix I. Data Sources and Description of the Sample

The paper uses Bankscope data, which is a comprehensive database with harmonized, detailed balance sheets and income statements of individual banks in various countries. This database allows for a reasonably consistent cross-country comparison of banking systems. The sample in the study only covers commercial banks, and all values are expressed in U.S. dollars. When available, sample data are based on consolidated statements; otherwise, to maximize sample size, unconsolidated statements are used.

To improve data quality, plausible value ranges are defined for some of the key variables. For example, an observation is excluded from the sample if bank balance sheets for a particular year show a negative value for equity or a value that exceeds 50 percent of banks assets. Similarly, an observation is excluded if average yearly personnel expenses per employee in a bank are below US$1,000 or above US$1 million. All the imposed rules are defined in Table A1, and this data filter eliminates 10.2 percent of the sample observations.

Table A1.Rules for Data Filtering
VariableLower BoundUpper BoundFallout (percent)
Commisions and fee income/interest income0103.7
Equity/total assets0.010.53.4
Other administrative and operating expenditures/fixed assets0152.7
Interest income/total earning assets00.251.1
Interest, commision and fee income/total earning assets00.350.9
Personnel expenses/employment110000.2

The remaining dataset covers 16 countries over an eleven-year period from 1995 to 2005. It includes a total of 579 banks. Table A2 lists the number of banks included in the sample for each year and each country. Selected sample statistics for 2005 are summarized in Table A3.

Table A2.Number of Banks Included in the Sample, 1995–2005
All years19951996199719981999200020012002200320042005
AUT733435353938424244434543
BEL523639383029282529322923
CZE311316181317181918172017
EST1168934555556
ESP932939373426272525182750
FIN1065666422245
GRC271113131287510121816
HUN281111151113171614151715
IRL348898881213141922
LTU13371010101099999
LVA291618211819191820222222
NLD522325211817141817183027
POL551922242722231723253324
PRT301819191918151210101312
SVN201112141515161513141512
SVK21812151211121314141415
Total579252289304275261265253266270320318
Source: Bankscope.
Source: Bankscope.
Table A3.Sample Statistics, 2005
Total assets (bn

USD)
Total earning assets

(bn USD)
Fixed assets (ml

USD)
Deposits, short term

and other funding (bn

USD)
Equity (ml USD)Employment (’000)
No.

of

obs.
MeanSt.

Dev.
No.

of

obs.
MeanSt.

Dev.
No.

of

obs.
MeanSt.

Dev.
No.

of

obs.
MeanSt.

Dev.
No.

of

obs.
MeanSt.

Dev.
No.

of

obs.
MeanSt.

Dev.
AUT438.830.3438.328.23971.5231.4437.927.143386.91,359.1291.65.7
BEL2366.6165.02361.2150.022578.61,559.82358.8146.0231,967.84,500.3113.55.4
CZE173.54.9173.34.61735.172.4173.04.217291.6151.1151.11.9
EST63.45.963.15.3632.562.962.95.16293.3507.651.93.0
ESP5038.9148.05035.3131.049491.21,826.35033.2125.0502,308.28,621.61716.036.8
FIN540.660.2537.558.4564.746.4564.752.753,613.76,097.152.93.7
GRC1617.322.71615.520.016385.0606.51615.119.5161,187.71,627.1113.03.6
HUN154.76.3154.25.51598.9161.9154.15.415410.2648.3112.65.3
IRL2238.772.62237.170.519129.6306.42235.267.6221,451.42,835.2211.13.6
LTU91.71.891.51.7941.342.291.51.69130.4145.370.90.8
LVA220.81.1220.81.02214.018.7220.71.02265.483.7210.50.7
NLD2794.8268.02788.4254.026707.32,348.62787.8253.0272,577.96,382.9225.620.5
POL244.75.6244.145.12478.1113.2244.14.824501.3671.8143.54.5
PRT1214.629.01213.225.912129.9296.41213.025.912850.01,579.093.16.8
SVN122.93.8122.73.51259.793.5122.63.512238.7272.8121.32.3
SVK152.22.6152.02.51544.960.8152.02.315172.8206.6131.31.6
Total31821.551.731819.948.0,308185.1489.631819.246.63181,028.02,249.32233.16.6
Source: Bankscope.
Table A3.Sample Statistics, 2005 (concluded)
Interest income (ml

USD)
Interest expenses (ml

USD)
Personnel

expenditures (ml

USD)
Other administrative

and operating

expenses (ml USD)
Total operating

expenses (ml USD)
Pre-tax profits (ml

USD)
No.

of

obs.
MeanSt.

Dev.
No.

of

obs.
MeanSt.

Dev.
No.

of

obs.
MeanSt.

Dev.
No.

of

obs.
MeanSt.

Dev.
No.

of

obs.
MeanSt.

Dev.
No.

of

obs.
MeanSt.

Dev.
AUT43277.5939.043172.6530.34362.7266.94351.5219.443140.0598.14352.9232.8
BEL233,519.49,270.8232,878.98,239.723382.8899.7223375.8888.123763.61,831.923379.5838.1
CZE17137.3198.41761.665.31728.648.11740.363.71772.2118.81755.9111.8
EST6123.5219.7645.978.2632.756.2636.367.3668.8123.5665.6120.0
ESP501,527.76,099.150910.24,039.750310.71,116.050272.11,051.050699.42,530.450460.41,637.5
FIN51,070.91,461.95662.3884.45190.3226.65186.4195.05387.8442.05317.1442.6
GRC16820.81,088.916331.6479.716233.2294.816162.0191.116488.6590.916220.1353.7
HUN15343.0540.315156.5196.91570.8112.91593.7144.415188.4292.115107.2225.1
IRL221,282.12,543.422945.21,979.922163.6430.822111.6278.722288.1745.422259.9580.6
LTU955.358.0919.819.3915.816.0919.218.6939.640.4918.022.5
LVA2231.643.12211.516.6228.711.12211.414.82220.826.12216.323.2
NLD273,312.99,343.2272,335.96,492.327617.01,894.227526.61,658.9271,186.33,691.027602.31,660.9
POL24264.8307.424121.4127.72475.395.82495.2111.824178.3220.32499.5145.6
PRT12597.81,191.012342.7666.112180.6412.112137.6282.212364.3763.912149.7322.8
SVN12129.4176.11256.473.81236.553.41235.148.61289.8124.71232.236.7
SVK1589.1115.91531.636.51524.031.51536.247.71564.384.51529.346.9
Total318848.92,099.8318567.81,495.4318152.1372.9318136.9330.1318315.0764.0318179.1425.0
Source: Bankscope.
Source: Bankscope.
References

Prepared by Rudolfs Bems (EUR).

The excluded larger EMU markets are Germany, Italy, France, and Luxembourg. NMS include the Czech Republic, Hungary, Poland, the Slovak Republic, Estonia, Latvia, and Lithuania. Slovenia is excluded from both groups. Unless noted otherwise, these definitions of EMU and NMS are followed throughout the paper.

Measured by asset size, these banks constitute 60 percent of the Slovene banking sector.

Labor cost share and its trend for Slovenia in Bankscope data are very similar to what is reported in Bank of Slovenia (2006). The latter is based on the aggregated banking sector balance sheet for Slovenia.

The method builds on the assumption that all banks in the sample face a common production function. for further details on stochastic frontier analysis methodology, see Kumbhakar and Lovell (2000).

These definitions of regression variables are standard in the literature. The only notable deviation is the definition of the price of labor. General practice in the literature, due to lack of data on employment, has been to express it as personnel expenses over total assets. Although using data on employment meant a somewhat smaller number of observations, correlations between the price of labor expressed using the two methods were close to zero, and, therefore, the series with a more appealing economic interpretation was chosen.

With more than one output variable, the regression equation takes a translog form.

PR methodology assumes, among other things, that bank cost structure is homogenous and banks operate in a long-run equilibrium with exogenous input prices. For a more detailed discussion of underlying assumptions, see, e. g., Bikker and Haaf (2002). This paper tests for the presence of long-run equilibrium in each market with an approach previously used in the literature (see, e.g., Claessens and Laeven (2004). Results of the long-run equilibrium test are reported in Table 3.

Our baseline specification already includes the absolute size of assets as a control, since the price of bank output is expressed as the ratio of revenues to all earning assets.

In calculating the average index values and rankings, we have ignored long-run equilibrium test results. for Slovenia, existence of a long-run equilibrium was not rejected in any of the specifications.

Calculated as H = Σ(Sij · 100)2 , where sij represents total assets of bank i in country j as a share of country j total bank assets.

As a reference, further financial integration within EMU is estimated to add 1 percentage point to GDP growth over the next 10 years (Giannetti and others 2002).

See Chapter III on “Bank Risks from Cross-Border Lending and Borrowing in Slovenia.”

Other Resources Citing This Publication