Comments on Chhaochharia

1
Comments onChhaochharia LaevenCorporate
Governance Norms and Practices

• Prof. Bernard Black
• University of Texas
• Law School McCombs School of Business
• Istanbul Emerging Markets Conf, Nov. 2007

2
Three groups of governance studies

• Broad, shallow cross-country studies
• Narrow, deep country studies
• Endogeneity skeptics
• This paper is in the first group, which is fine
• It slights the other two, which I think is not

3
Cross-country studies (which I don't do)

• Limited set of indices
• CLSA (2001, not repeated, partly subjective
  index)
• Klapper Love (2004)
• Mitton (2004)
• SP (2002, not repeated)
• Durnev Kim (2005)
• Doidge, Karolyi Stulz (2007)
• Alliance Bernstein (partly subjective index, no
  details)
• Baker, Godridge, Gottesman Morey (2007)
• ISS (2003-2005, developed markets)
• Aggarwal, Erel, Stulz Williamson (2006)
• Bruno Claessens (2007)
• Durnev Fauver (2007)
• this paper (Chhaochharia Laeven, 2007)

4
Country studies (the stuff I do)

• Published studies on
• Brazil (Leal and Carvalhal-da-Silva, 2007)
• Germany (Drobetz, Schillhofer and Zimmerman
  (2004)
• Hong Kong (Cheung, Connelly, Limpaphayom and
  Zhou, 2007a)
• Korea (Black, Jang and Kim, 2006a, 2006b)
• Russia (Black, 2001 Black, Love and Rachinsky,
  2006)
• Switzerland (Beiner, Drobetz, Schmid and
  Zimmermann, 2006)
• Turkey (Orbay Yurtoglu, 2006)
• Also working papers on
• China (Cheung et al., 2007b)
• India (Balasubramanian, Black Khanna, 2007)
• Japan (Aman Nguyen, 2007)
• Ukraine (Zheka, 2007).

5
Endogeneity skeptics

• Need to rule out
• reverse causation
• omitted variables
• optimal differences
• Some recent skeptical views
• Hermalin Weisbach (2003), Board of Directors as
  an Endogenously Determined Institution
• Chidambaran, Palia Zheng (2006), Does Better
  Corporate Governance Cause Better Firm
  Performance?
• Lehn, Patro Zhao (2006), Governance Indices and
  Valuation Which Causes Which? (2006)
• Listokin (2007), Interpreting Empirical Estimates
  of the Effect of Corporate Governance (2007)
• Wintoki, Linck Netter (2007), Endogeneity and
  the Dynamics of Corporate Governance

6
Why does this matter?

• Nice to have a new data set, even if for limited
  set of countries
• But . . .
• Literature review focuses on cross-country papers
• Cites almost none of the second group
• Cites none of the third group
• Why does this matter?
• Guide to what matters in corporate governance
• How you build your index
• Guide to robustness checks
• Tobin's q vs. market/book
• effect of outliers
• Demirguc-Kunt-Maksimovic vs. Rajan-Zingales
  measure of financing constraints
• different ADR levels
• How to study individual elements

7
Black, Kim, Jang Park (2006)
Lesson 1 Aspects of governance matter Lesson 2
When study one aspect, control for others Lesson
3 Firm fixed effects vs. OLS
8
Black, Love Rachinsky (2006)
Lesson 1 Not all indices are created
equal. Lesson 2 Firm fixed effects vs. OLS
Can be very different
9
Black, Jang Kim (2006)

• Lesson 1 Coefficients change if control for
  rest of governance
• Lesson 2 Some results survive, but many don't
• Not just higher standard errors due to
  colinearity

10
Country Fixed Effects

• How different than competing papers?
• Aggarwal et al., Bruno Claessens, Durnev
  Fauver use same data set, total variation
• This paper variation above country minimum
• Why is variation above minimum different than
  total variation, with country fixed effects.
• Aggarwal et al use country fixed effects
• Bruno Claessens don't but should
• Authors say country fixed effects don't allow
  for variation in country minimum across time.
• But how large is this variation? (I bet it is
  small.)
• Similar results for total variation with country
  fixed effects? (I bet they are.).

11
Firm fixed effects next level down

• Good instruments for governance are hard to find
• I know of no good ones in cross country work
• Legal origin as country level instrument . . .
  doesn't predict only governance
• I like your instrument for Tobin's q, which you
  use to address reverse causation (oil
  priceindustry price sensitivity)
• Without good instruments, we need firm fixed
  effects
• not perfect, but better than OLS
• In five years, you will be able to do this
• your results will survive or not (I'm agnostic)
• Meanwhile, want to see more on the next level
  down
• groups of similar elements
• individual elements

12
More CG Isn't Always Good

• One can read this paper and never get a clue that
  more CG isn't always good
• Yet most evidence on reaction to SOX by investors
  and cross-listed firms is negative
• U.S. (Zhang, 2008)
• cross-listed firms (Litvak, 2007 others)
• delisting studies
• So, SOX raises US minimum score from 1 to 3, and
  that is apparently bad!
• Canada has min score of 0 U.S. used to have 1
• What then to make of your governance index?