roe in banks : myth and reality · 2021. 3. 17. · roe in banks : myth and reality christophe...
TRANSCRIPT
RoE in Banks : Myth and Reality
Christophe Moussu
Professeur
ESCP Europe, College of Europe, Labex Refi
Matinale de l’EIFR
Paris, 28 Novembre 2013
This presentation is based on a research conducted with Arthur Petit-Romec, ESCP Europe and Université Paris I Panthéon-Sorbonne, « RoE in Banks: Myth and Reality », available on the Labex Refi Website (working papers)
An addiction to RoE…
A great deal of evidence suggests that RoE is a central measure of performance in the banking industry:
• P. Jenkins: “Still today, most banks around the world use return on equity – RoE – as their main metric of profitability”; Financial Times, 7th of November 2011
• J. Dimon (JP Morgan): “We feel good” about the 17 % RoE of the investment bank”; Financial Times, 16th of January 2013
• H. Schwartz (CFO Goldman Sachs): “The bank did not find the number (RoE of 10.7%) aspirational in the long term” Financial Times, 12th of February 2013
• Jean-Laurent Bonnafé, DG BNP Paribas, “Aujourd’hui, BNP Paribas présente un RoE de 8 %, alors qu’il était le double avant la crise- c’était d’ailleurs plutôt moins que les autres banques à l’époque car nous avions refusé de mener certaines activités spéculatives. En moyenne le secteur bancaire européen offre une rentabilité du capital de 6%... pour un coût du capital qui tourne autour de 10 %. C’est une fragilité pour l’ensemble de l’industrie européenne: cela pèse sur la capacité à financer l’économie, cela rend vulnérable à un nouveau choc, cela conduit les banques se concentrer uniquement sur les sous-activités qui offrent le plus de RoE, au risque de délaisser les autres… Et c’est d’autant plus problématique que les banques américaines, elles, présentent des RoE de 12-13 %, car elles se sont davantage consolidées depuis la crise et plus investies dans les activités de marché, mieux rémunérés qu’en Europe”, Les Echos, 14 octobre 2013
2
Regulation, risk management and RoE
• From a corporate finance perspective, the addiction to RoE makes no sense
• The reliance on RoE has emerged from the risk management and regulation
approaches to the level of bank capital
• If regulation was perfect, risk management strategies efficient/ homogeneous and incentives correct, the level of equity should perfectly be adjusted upward to any incremental risk taken by the banks.
• In this case, higher RoE would signal higher rents or more value creation, and could be an appropriate performance measure.
• In no case it should be associated to higher risks.
3 Prof. Christophe Moussu
Main hypothesis
• Our main hypothesis is that, despite its legitimacy from a risk
management and regulatory perspective, reliance on RoE as a
central measure of performance is responsible for strategies of
excessive risk-taking because of regulatory arbitrage/ inadequate
risk management/ poor incentives/ possibility to hide risks.
• In order to test our main hypothesis, we use the crisis as a risk
revelator event and examine whether the pre-crisis RoE of banks has
an impact on the losses suffered by shareholders in the crisis.
• If RoE is a relevant performance measure, no statistical association
should be found.
4 Prof. Christophe Moussu
Sample construction
• The starting point for the construction of the sample is every financial institution available on the Bankscope database for the year 2006
• First, we exclude from the sample the banks which are not listed, because our main dependent variable is the buy-and-hold stock return of banks during the crisis.
• We only keep banks from those countries for which Djankov, La Porta, Lopez-de-Silanes and Shleifer (2008) provided revised data on anti-director rights and for which the regulation indices based on Barth, Caprio and Levine (2008) are available.
• We focus our analysis on the largest banks, with assets over $10 billion just before the crisis at the end of 2006.
• We end up with a sample of 273 banks from 28 countries (245 deposit-taking banks)
5 Prof. Christophe Moussu
Distribution of banks by countries
Prof. Christophe Moussu 6
Countries Number of banks
(Main sample)
Australia
3
Austria 5
Belgium 2
Brazil 2
China 1
Denmark 4
France 14
Germany 11
Great Britain 9
Greece 5
Hong Kong 7
India 15
Ireland 3
Israel 5
Italy 12
Japan 75
Malaysia 1
Netherlands 2
Norway 4
Portugal 3
Russia 2
Singapore 3
South Africa 5
Spain 7
Sweden 4
Switzerland 13
United States 46
Taiwan
10
TOTAL 273
Summary statistics
7
Number of
observations
Mean Median Std. dev Min Max
BHR 273 0.636 0.593 0.323 0.012 1.411
BHR 2006 273 1.185 1.19 0.341 0.343 2.747
RoE (%) 273 20.12 18.49 11.06 0.57 52.39
RoA (%) 273 1.41 1.24 1.08 0.04 9.89
Tangible equity
(%)
273 6.25 5.76 4.01 -0.34 41.65
Tier 1 (%) 212 9.08 8.57 2.39 4.82 22.9
Deposits (%) 273 57.18 61.57 27.11 0.00 93.40
Funding Fragility
(%)
273 24.40 12.83 26.66 0.00 100.00
Non-interest income
(%)
273 38.71 35.60 22.66 0.76 109.80
Density 190 0.60 0.59 0.17 0.15 1.21
Size 273 10.92 10.54 1.40 9.21 14.49
Beta 273 1.02 0.98 0.42 0.23 2.39
ADRI 273 3.82 4 0.94 1 5
Institution 273 1.13 1.24 0.50 -0.76 1.83
Official 273 11.07 12 2.15 5 14
Private monitoring 273 7.12 7 0.80 4 8
Capital 273 6.11 6 1.35 2 9
Restrict 273 9.91 11 2.11 4 15
Prof. Christophe Moussu
Empirical results : regressions of BHR on RoE and
control variables
8
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Constant 0.877***
(0.262)
0.875***
(0.261)
0.829***
(0.313)
0.070
(0.316)
0.289
(0.244)
0.540**
(0.244)
0.640***
(0.237)
1.014**
(0.404)
1.401***
RoE -0.764***
(0.182)
-0.755***
(0.181)
Tangible
equity
-0.012**
(0.005)
Tier 1 0.016*
(0.010)
Deposits 0.496***
(0.089)
Funding
fragility
-0.338***
(0.076)
Non-interest
income
-0.514***
(0.104)
Density -0.479***
(0.124)
BHR2006 -0.282***
(0.072)
Size -0.067***
(0.012)
-0.066***
(0.013)
-0.095***
(0.014)
-0.073***
(0.014)
-0.057***
(0.020)
-0.068***
(0.011)
-0.061***
(0.013)
-0.090***
(0.015)
-0.071***
(0.012)
Beta -0.015
(0.049)
-0.056
(0.049)
-0.076
(0.057)
0.006
(0.045)
-0.013
(0.048)
0.030
(0.050)
-0.133**
(0.062)
-0.063
(0.047)
Country
variables
Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 273 273 273 212 273 273 273 190 273
R² 0.337 0.338 0.306 0.387 0.388 0.351 0.380 0.424 0.335
Prof. Christophe Moussu
Empirical results: regressions of BHR on RoE and
control variables
9
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Constant 0.706**
(0.302)
1.006***
(0.275)
0.953***
(0.278)
0.275
(0.355)
0.648*
(0.351)
0.511
(0.348)
1.073***
(0.358)
1.270**
(0.376)
1.019***
(0.370)
0.544
(0.449)
RoE -0.690***
(0.184)
-0.752***
(0.178)
-0.518***
(0.196)
-0.761***
(0.213)
-0.864***
(0.214)
-0.622***
(0.238)
-0.616***
(0.234)
-0.552**
(0.279)
-0.586***
(0.183)
-0.665***
(0.225)
Tangible
equity
-0.004
(0.005)
-0.006
(0.005)
-0.005
(0.005)
-0.003
(0.005)
Tier1 ratio 0.016**
(0.008)
0.013
(0.008)
0.019**
(0.008)
0.017**
(0.008)
Deposits 0.444***
(0.101)
0.559***
(0.159)
0.557***
(0.167)
0.422***
(0.099)
0.497***
(0.168)
Funding
fragility
-0.295***
(0.081)
-0.226
(0.152)
Non-interest
income
-0.398***
(0.108)
-0.361**
(0.158)
-0.254
(0.172)
Density -0.351***
(0.133)
-0.348***
(0.131)
BHR 2006 -0.123*
(0.069)
-0.095
(0.083)
Size -0.051***
(0.015)
-0.063***
(0.013)
-0.062***
(0.015)
-0.031**
(0.015)
-0.052***
(0.014)
-0.047***
(0.015)
-0.050**
(0.017)
-0.069***
(0.015)
-0.049***
(0.014)
-0.033**
(0.015)
Beta 0.025
(0.046)
0.008
(0.048)
0.030
(0.052)
-0.028
(0.055)
-0.037
(0.057)
-0.041
(0.058)
-0.062
(0.063)
-0.083
(0.069)
0.012
(0.047)
-0.044
(0.059)
Country
variables
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 273 273 273 212 212 212 190 190 273 212
R² 0.428 0.401 0.399 0.494 0.449 0.464 0.514 0.471 0.435 0.498
Prof. Christophe Moussu
Main results
• RoE has a strong and economically significant impact on the destruction of value for shareholders during the crisis.
• This rejects the null hypothesis that RoE is not be associated with higher risk, and casts serious doubt on the relevance of RoE as a performance measure to drive capital allocation inside banks and to assess bank performance.
• Our results are very robust to the introduction of numerous variables, controlling for risks, both on the asset and the liability side of bank balance sheets, and market measures of risk.
• The fact that a residual effect of RoE remains after the introduction of these variables strongly suggests that, on top of being a poor performance measure, it was responsible for risk-taking in banks.
10 Prof. Christophe Moussu
An incentive to maximise RoE?
• So far, our results strongly suggest that RoE captures risk-taking on top of
various other risk measures and that reliance on this metric is likely to have
induced risk-taking which materialized during the crisis.
• However, they do not reveal whether the incentives to maximize RoE
resulted from direct monetary incentives or from the fact that RoE is
prevalent everywhere, in and across banks.
• We investigate the sensitivity of bank CEO compensation to RoE for a
sub-sample of 64 banks.
11 Prof. Christophe Moussu
Regression of total CEO compensation on RoE and
control variables
12
(1) (3) (4) (5) (6) (7)
Constant
8.267***
(0.313)
8.854***
(0.206)
8.640***
(0.992)
6.379***
(0.642)
6.060***
(0.673)
7.039***
(0.877)
ROE 2.929**
(1.210)
2.289*
(1.182)
2.698**
(1.189)
ROA 9.160
(9.669)
BHR
2006
0.323
(0.801)
-1.393*
(0.746)
Size 0.223***
(0.057)
0.200***
(0.062)
0.252***
(0.072)
N 64 64 64 64 64 64
R² 0.09 0.02 0.003 0.161 0.214 0.253
Prof. Christophe Moussu
Are banks special?
Regressions for non financial firms Table 8: OLS regressions of BHR on RoCE, RoE, RoA for non-financial firms
This table reports the OLS regressions of the buy-and-hold return. In each specification, the dependent variable is the weekly buy-and-hold return from July 2007 to the end of 2008 and we control for
the size and the beta. RoE is the ratio of pre-tax-profit to equity. RoA is the ratio of the pre-tax-profit to total assets. Size refers to the natural logarithm of total assets. Beta is the slope of the regression
of weekly excess stock returns on the MSCI World excess return from 2004 to 2006. Country variables include ADRI, Institution. Below the regression coefficient is reported the standard error in
parentheses and to the right of the regression coefficient its significance (*** significant at 1%, **significant at 5%, * significant at 10%).
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Constant
-0.166
(0.385)
0.140
(0.392)
0.293
(0.395)
0.084
(0.404)
0.142
(0.379)
0.325
(0.377)
0.296
(0.376)
-0.039
(0.386)
0.142
(0.390)
0.132
(0.188)
RoCE
0.600***
(0.167)
0.536***
(0.166)
0.497***
(0.161)
0.469***
(0.163)
0.452***
(0.170)
RoE
0.086
(0.061)
RoA
0.880***
(0.239)
Equity 0.218***
(0.072)
0.193***
(0.071)
Total debt
Net debt
-0.206**
(0.092)
-0.142
(0.091)
-0.167***
(0.062)
-0.107*
(0.063)
Size 0.031
(0.020)
0.027
(0.020)
0.027
(0.020)
0.028
(0.020)
0.030
(0.020)
0.029
(0.020)
0.029
(0.019)
0.031
(0.020)
0.029
(0.019)
0.029
(0.019)
Beta
-0.188***
(0.043)
-0.194***
(0.044)
-0.186***
(0.043)
-0.197***
(0.043)
-0.208***
(0.044)
-0.207***
(0.044)
-0.184***
(0.042)
-0.193***
(0.043)
-0.193***
(0.043)
Country
variables
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Industry dummies
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 460 460 460 460 460 460 460 460 460 460
R² 0.340 0.372 0.356 0.375 0.372 0.362 0.364 0.383 0.377 0.377
Prof. Christophe Moussu 13
BHR and RoE: long term effects
(preliminary results)
Prof. Christophe Moussu 14
BHR 1999-2006 BHR 1999-2008 BHR 1999-2012 RoE 1999-2006 (%)
Q1_RoE 3,02 2,09 2,05 11,22
Q2_RoE 2,54 1,20 1,74 15,23
Q3_RoE 2,18 1,25 1,27 18,67
Q4_RoE 2,56 1,20 1,58 21,93
M1_RoE 2,77 1,63 1,90 13,23
M2_RoE 2,37 1,23 1,43 20,36
Long term average BHR of banks grouped by level of average RoE
pre-crisis (63 U.S. banks with total assets over 10 B$ once over the
period 1998-2006)
Hard facts…
• 1 dollar invested in a equally-weighted portfolio of stocks of the lowest decile of RoE in 1999 (5,2 % average) is worth 3,93 dollars in 2006 and 2,47 dollars in 2008
• The average RoE over the period 1999-2006 is 8,25 % and the average annual stock return 9,46%!
• 1 dollar invested in a equally-weighted portfolio of stocks of the
highest decile of RoE in 1999 (28,81 % average) is worth 2,66 dollars and 1,01 dollar in 2008!
• The average RoE over the period 1999-2006 is 21,97% and the annual average stock return 0,01 %!
Prof. Christophe Moussu 15
Conclusion
• A good performance measure should provide incentives to enhance long term shareholder value and protect from excessive risk-taking, given the explicit and implicit protection of banks
• The specific status of RoE in banks is related to the fact that equity is regulated/ the risk management approach to equity prevails
• In the light of the financial crisis (and before…), measuring performance with RoE appears to be a fallacy…
• In fact, RoE appears as an excellent proxy for risk!
• RoE: “Le Faux Graal de la performance bancaire” Moussu, Ohana, Troege, La Tribune 19 Avril 2011
• Root of Evil? FT Lex team August 18, 2013
16 Prof. Christophe Moussu
Conclusion
• Reference to RoE after the crisis is even stronger as stricter
regulation imposes an increase in E!
• New regulation with no reboot of the way performance is assessed
(and compensation paid) may prove to be ineffective and even
dangerous
• On-going research : effects of annual RoE variations on bank
stock returns… seems to reveal strong short term market bias
• Future research: political process of the performance measure
design in banks, value appropriation in banks, RoE
resistance…
Prof. Christophe Moussu 17