OWNERSHIP STRUCTURE AND FIRM PERFORMANCE:

Transcription

1 OWNERSHIP STRUCTURE AND FIRM PERFORMANCE: EVIDENCE FROM THE BANKING INDUSTRY 1 By Simon Kwan Discussion by Lamont Black CAREFIN Universita Bocconi Business Models in Banking Conference September 21, 2009 The opinions of this discussion are solely those of the author and do not necessarily reflect those of the Board of Governors of the Federal Reserve System.

2 Main Idea 2 Question: How does public ownership affect firm performance? Agency costs of separating ownership and control Does market discipline mitigate these costs? Compares the performance of public and private BHCs using three measures: profitability, operating efficiency, i and risk-taking ki ( public = publicly traded equity)

3 Data 3 U.S. Top-tier Bank Holding Company Y-9Cs 1986 to 2008 Match to Call Reports of bank subsidiaries Exclude those with less than 90% banking assets Use CRSP to identify publicly-traded BHCs

4 Performance Measures 4 Y-9C accounting measures Profitability Return on assets (ROA) Return on equity (ROE) Operating efficiency Ratio of total noninterest operating costs to total assets (OPCOST) Components of operating costs

5 Risk-Taking Measures 5 Risk-taking Loan portfolio composition Ratio of commercial real estate loans to total loans (CRELOAN) Ratio of construction loans to total loans (CONSTRUCT) Leverage/Capitalization Ratio of equity (including common and perferred stock, and retained earnings) to total assets

6 Hypothesis 1 6 H1: Public BHCs are less efficient. [bad] Public BHCs are less profitable and less efficient than similar private BHCs. Assertion: public BHCs have greater separation of ownership and control Agency theory: manager has incentive to consume perks rather than maximize i shareholder h value

7 Hypothesis 2 7 H2: Public BHCs are less risky. [good] Public BHCs take less risks than similar private BHCs. A) Agency theory: Manager of a public BHC with nondiversifiable human capital will take less risk than desired by shareholders B) Market discipline theory: Public BHC is disciplined by public market signals to take less risk

8 Tests 8 Nonparametric Wilcoxon Rank Sum Tests Compare distributions of public and private BHC measures within size classes OLS Regressions Control for firm characteristics, time, and location Uses size classes because banks of different sizes have different business models Size classes are based on private BHCs Largest public BHCs excluded

9 Findings and Interpretations 1 9 H1 (public BHCs are less efficient) Support: All public BHCs are less efficient. Public BHCs other than largest are less profitable. Against: Largest public BHCs are more profitable. Interpretation: Publicly held BHCs have greater separation of ownership and control. Some results support the predictions of agency theory.

10 Findings and Interpretations 2 10 H2 (public BHCs are less risky) Against Small publicly traded BHCs tend to make more risky loans than small privately held BHCs Mixed Mixed evidence for largest BHCs Interpretation: Results do not support agency theory Interpretation: Results do not support agency theory or market discipline.

11 Comments Initial Response 11 Interesting idea, difficult to execute Limited data on private firms Remind the reader: Cannot use market data for performance or risk (e.g. stock performance or volatility) (Another paper: degrees of ownership within public BHCs)

12 Comments about Measures 1 12 Performance Efficiency: What about the banking literature (e.g. Berger and Mester JBF 1997)? Interpreting cost components should be done more carefully (e.g., could reflect different use of labor) Should consider other measures of capital (like TCE)

13 Comments about Measures 2 13 Risk-taking Ex-ante?: Portfolio composition (CRE concentration) Community banks have focused on CRE Concentration in CRE may depend on securitization (CMBS) Against ex-post like non-performing loans (NPL): reflects external circumstances not under manager s control What about ratio of NPL relative to average NPL of the size class within that year? Time dummies help with this

14 Comments about Hypothesis 2 14 Risk-taking is not performance Is taking less risk a good thing? Maybe for regulator and debt holders Maybe not for equity holders. (e.g., Morgan Stanley) Tighten theoretical framework Opposite predictions Nondiversified wealth: In private BHC, less separation of Nondiversified wealth: In private BHC, less separation of ownership and control. Less appetite for risk. Alternative agency theory: taking on more risk than the shareholders want (compensation reform)

15 Comments about Main Idea 15 Premise: Agency costs are greater for public BHCs Find that public BHCs underperform private BHCs Alternative stories to agency costs? Ownership structure does not cause performance, performance causes ownership structure Why do BHCs go public?

16 Specific Comments 1 16 Differentiate diffuse ownership vs. manager ownership Private BHC may still have separate owner and manager Merger adjustment? Remove first and last quarter What about measures of acquirer? Profitability and cost Have you considered a matching approach rather than size classes?

17 Specific Comments 2 17 Y9C data (BHC) vs. Call Report data (bank subs) Can do same exercise using Call Report data Example: Capitalization can be different for BHC and bank subsidiaries Percentage of assets in banking likely differs across public and private BHCs Interesting topic. Keep digging!

18 18

19 Small Comments 19 Y9-C data is quarterly, but paper talks about firmyears. Please clarify.