Discussion of Altunbas et al. (2016) “Competition and Bank Risk: The Effect of Securitization - PowerPoint Presentation

Slide1

Discussion of Altunbas et al. (2016) “Competition and Bank Risk: The Effect of Securitization and Bank Capital”*

EBA Policy Research Workshop

Philipp Koziol

ECB-UNRESTRICTED

DRAFT

*This presentation represents the

authors' personal opinions and

does

not necessarily

reflect

the views of the

ECB or its staff..

Competition and risk-takingTwo strands of literatureCompetition stimulates risk-taking (focus on

deposit market)

Banks take on more risk to compensate for reduced profits due to competitionCompetition reduces risk-taking (considers deposit and loan market)

Competition lowers the interest rate of loans, reducing the default rate of customersFocus on the influence of securitization and bank capitalDoes securitization amplify the relationship between comp. and bank risk?Does bank capital weaken the relationship between comp. and bank risk?Probit model regressing risk-taking on competition, an interaction term and bank specificsBank risk measured with financial support, systemic risk or idiosyncratic riskCompetition measured with the Boone indicator, bank lending survey or C5Robust findings of securitization amplifying the comp./risk-taking relationshipNo proof for capital to diminish the relationship between comp. and risk-takingAdvising a closer

cooperation between supervisory and competition authoritiesDiscussion of Altunbas et al. (2016)

2

Summary

ECB-UNRESTRICTED

DRAFT

Newness of research and findingsThe paper contributes to the discussion on competition and bank risk and the impact of securitization

The authors find that, all else equal, securitization and bank capital diminish risk-taking

When competition is more intense, securitization reduces risk-taking to a lesser extendContributes to the discussion on how to deal with competition in the financial sector

How to regulate the sector?Who should regulate the sector?The authors argue for better cooperation between banking supervisors and competition authorities

Discussion of Altunbas et al. (2016)3

Value added to the literature

ECB-UNRESTRICTED

DRAFT

Feedback and concerns (1/2)Identification strategy can be streamlinedMeasuring competition

Why is Lerner index not applied as this index measures competition on bank-level rather than country-level? Reasoning in paper is not clear…Boone indicator applied on country level, how is this applied for large internationally active banks? For instance French bank operates worldwide… (weighted average?)

Measuring risk-takingAlthough the materialization of bank risk is most likely correlated to risk-taking, it is not the same and the moment of risk-taking is uncertain

CausalitySecuritization and competition might enhance risk-taking, but risk-taking might also enhance securitizationPartly offset by using different time periods (pre-crisis and crisis periods)Supervisory/Regulatory authorities make risky banks hold more capitalSelf-selection bias: Banks choose their amount of securitization. Risk seeking banks might opt for more securitization, but they are also by definition riskierCausality true for recent period? Securitization exposures have been considerably reduced and the entire markets has been structurally changed

Discussion of Altunbas et al. (2016)

4

Suggestions for improvement

ECB-UNRESTRICTED

DRAFT

Feedback and concerns (2/2)General framework

No statistically significant effect when using idiosyncratic

riskBank specific factors are used. These should explain bank specific risk

More bank capital relates to lower likelihood of state aid. Does this mean they took on less risk? Or were they simply able to withstand more losses?When using interaction terms the likelihood of finding statistically significant results increasesThink of using a Bonferroni correctionRejection of null-hypothesis with statistically insignificant results (p.16)Inconsistent measurement of certain effectsCompetition (Boone: 94-04, bls 03-05; definition of bls, Concentration 95-05)Inconsistent use of the pre-crisis and crisis periodHow can EXLEND (excess lending growth) on average be positive (6.70)?Measuring securitization: trading book vs. banking book (different purpose)At least 75% of banks have no securitization in the pre-crisis periodTreatment of systemic vs. systematic effects vs. idiosyncratic effects at least requires explanations

Discussion of Altunbas et al. (2016)

5

Suggestions for improvement

ECB-UNRESTRICTED

DRAFT

Minor comments

Literature overview is missing key contributions, e.g. Kick/Prieto (2013), papers by Klaus Schaeck

Improve data section:Description of data set lacking important information (for instance I am familiar with most of the sources, I am not able to follow which data exactly you are using) and data manipulation procedures require more details (Did you

winsorize the data set?)Check for typos

Discussion of Altunbas et al. (2016)6

Suggestions for improvement

ECB-UNRESTRICTED

DRAFT