Is Banks Cost of Equity Capital Different Across Countries? Evidence from G10 Countries Major Banks

Transcription

1 Bocconi University Centennial Conference Risk and Stability in the Financial System: What Role for Regulators, Management and Market Discipline? June, 2002 Is Banks Cost of Equity Capital Different Across Countries? Evidence from G10 Countries Major Banks Aurelio Maccario*, Andrea Sironi** and Cristiano Zazzara** *Unicredit Banca Mobiliare (UBM), **Università Bocconi, ***Università LUISS-Guido Carli

2 Outline Introduction Previous evidence Contributions of this paper Research methodology Data sources Sample features Main empirical results Main conclusions 2

3 Introduction (1) One of the major goal behind the 1988 Basel Capital Accord was that of leveling the playing field internationally Evaluating the Accord contribution to this goal is a difficult task The relative competitiveness of banks is affected by other factors such as: (i) implied penalties for breaching capital ratios, (ii) moral suasion policies, (iii) cost of equity differences If banks cost of equity is different across countries, then banks from lower cost of equity countries enjoy a competitive advantage 3

4 Introduction (2) Potential sources of cross-country cost of equity differences: Capital markets segmentation: exchange controls or taxes to cross border investments Cross-country differences in the safety net : if explicit/conjectural guarantees are in place, then equity investors would face a lower risk Cross-country differences in regulatory discipline: sanctions (PCA type), moral suasion Differences in tax rates 4

5 Previous evidence (1) Zimmer and McCauley (1991) analyze the real cost of equity (measured as real after tax adjusted profit on market price of equity) for 34 international banks from six major countries in the period They find that US, UK and Canadian banks experienced significantly higher cost of equity (measured as real after tax adjusted profit on market price of equity) than their Swiss, German and Japanese counterparts Wagster (1996) argued that the Accord did not eliminate the pricing advantage of Japanese banks 5

6 Previous evidence (2) Zimmer-McCauley Empirical Results Major Banks' Avg Real Adjusted Cost of Equity by Country (avg ) 12,00% 10,00% 8,00% 6,00% 4,00% 2,00% 0,00% USA JNP GER UK CAN SWI AVG STD.DEV. 6

7 Previous evidence (3) Zimmer and McCauley (1991) empirical analysis suffers from two main drawbacks: 1. Actual versus expected profit Cost of equity = after-tax adjusted profit on market price of equity The true cost of equity should be the return stockholders expect on their investment If investors expect a bank s profitability to rise, its current profit rate understates its cost of equity 2. Profits ciclicality Equity investors recognize a temporary downturn of profitability if a bank has a bad year and price the stock in anticipation of higher future profits 7

8 Contributions of this paper This paper compares the cost of equity capital of major banks from twelve economically developed countries over the period The analysis differs from previous studies in three main respects: a larger sample of banks from 12 major countries banks cost of equity capital have been estimated using analysts earnings forecast rather than actual ones a multivariate regression analysis of individual banks costs of equity based on both microeconomic and macroeconomic variables has been performed 8

9 Research Methodology (1) Cost of equity of each single bank in the sample estimated as the ratio between expected EPS and Price ke = 1 PE k e = E( EPS 1 t + 1 E( PE) = P t ) 9

10 Research Methodology (2) Inflation adjustment of expected profits: AdjE( P t + 1, i ) = E( P ) E( π t + 1, i t + 1 ) [ NNA + E( NNA )] t, i 2 t + 1, i where: E(π t+1 ) = poll of inflation forecast produced by The Economist at the end of period t NNA t,i = Net Nominal Asset position of bank i at the end of period t 10

11 Research Methodology (3) Accounting variables used in the regression (Ke dep. var.) Variable Description Expected Sign CAP Tier 1 regulatory capital - LOSS Log growth rate of loan losses - GROWTH Expected growth in earnings + PAYOUT Payout ratio + Ke= a + a a CAP+ a 1 D95+ a 8 2 D96+ a LOSS+ a 9 D97+ a 10 3 GROWTH+ a D98+ a 11 D99 4 PAYOUT+ a 5 D93+ a 6 D94+ 11

12 Research Methodology (4) Macroeconomic variables used in the regression (Ke dep. var.) Variable Description Expected Sign LINTRATE Real long term interest rate + MKTPM Market premium + GDPGWT Real growth in GDP + Ke = a + a LINTRATE + a MKTPM Ke = a + a GDPGWT + a LINTRATE 12

13 Data sources I/B/E/S information on both current and historical earnings forecast by major brokers Datastream information on both current and historical data on stock prices BankScope/FitchIbca historical and current detailed information on major banks accounting data, ratings and financial ratios 13

14 Sample features (1) Banks from twelve industrialized countries: BEL, CAN, ESP, FRA, GER, ITA, JNP, NET, SWE, SWI, UK, USA Sample period: Total sample of banks ranging from 111 to

15 Sample features (2) Number of banks for each country Country No. Of banks No. Of banks No. Of banks No. Of banks No. Of banks No. Of banks No. Of banks No. Of banks No. Of banks BELGIUM CANADA SWITZERLAND GERMANY SPAIN FRANCE GREAT BRITAIN ITALY JAPAN NETHERLANDS SWEDEN USA TOTAL

16 Main empirical results (1) 1. Since 1996, the G-10 countries major banks average costs of equity have undertaken the way of convergence Std. dev. decreased from 3.66% in 1996 to 1.29% in While the whole period results are similar to Zimmer-McCauley ones, the last four years results appear much better in terms of leveling the playing field 16

17 Main empirical results (2) Major Bank's Average Estimated Cost of Equity by Country 18% 16% 14% 12% 10% 8% % 4% 2% 0% BEL CAN SWI GER ESP FRA UK ITA JNP NET SWE USA 17

18 Main empirical results (3) Zimmer-McCauley Maccario-Sironi-Zazzara USA 11,90% 8,94% 7,56% JNP 3,10% 2,79% 4,46% GER 6,90% 6,98% 6,37% UK 9,80% 8,88% 7,69% CAN 10,30% 12,03% 10,07% SWI 5,30% 8,16% 8,15% AVERAGE 7,88% 7,96% 7,38% STD.DEV. 3,35% 3,04% 1,87% 18

19 Main empirical results (4) Country Averages of Major Banks' Cost of Equity 16% 14% 12% 10% 8% 6% AVERAGE MAX-MIN STD. DEV. 4% 2% 0%

20 Main empirical results (5) 3. Accounting variables based regression: high statistical significance of the coefficients and small differences between actual and predicted values of cost of equity 4. Macroeconomic variables based regression: high statistical significance of the GDP and Long Term Interest Rate variables for all countries included in our sample (on the contrary, the market premium variable has a wrong coefficient sign, although rarely significant) 20

21 Main empirical results (6) Variable Coefficient t-statistic C * (4.4432) CAP ** ( ) LOSS ** ( ) GROWTH (1.3899) PAYOUT * (5.6294) D * (3.8643) D ** (1.9797) D ** (3.0773) D (1.4248) D ( ) D * ( ) D ** ( ) Adj. R F * *, **, and *** denote 1, 5, and 10% level of significance respectively 21

22 Main empirical results (7) 22

23 Main empirical results (8) 23

24 Main conclusions (1) The convergence process in the G-10 countries major banks costs of equity has important policy implications. Lower differences in international active banks costs of equity suggest that the capital adequacy framework is now potentially more effective in pursuing a leveling the playing field goal. The New Capital Accord, as recently proposed by the Basel Committee for Banking Supervision, capturing banks risks more effectively, should definitely improve the international competitive environment in the banking sector. 24

25 Main conclusions (2) Three main regulatory changes might enable to improve competitive equality: 1. Defining uniform standards of regulatory actions PCA type - for those banks that fall below minimum capital requirements 2. Improve the international convergence in the definition of regulatory capital 3. Explicitly forbid bailout policies that include the protection of equity investors and uninsured banks creditors such as sub debt holders 25