The Role of Institutional Investors in Seasoned Equity Offerings Thomas Chemmanur Shan He Gang Hu Di

1
The Role of Institutional Investors in Seasoned
Equity OfferingsThomas ChemmanurShan HeGang
HuDiscussion byMichael R. RobertsThe Wharton
School, University of PennsylvaniaNBER
Microstructure ConferenceOctober 2007
2
Question(s)

• Question(s)
• Do institutions have private information about
  firms?
• Do institutions use this private information to
  obtain allocations in better SEOs?
• Do institutions manipulate prices ex ante via
  pre-offer selling to extract ex post rents?
• Do institutions produce long-lived information ex
  ante and then buy pre- and post-offer?
• Do institutions flip their shares for short term
  profits?
• Do institutions earn abnormal returns after an
  SEO?
• Are these returns a function of the SEO
  allocation?
• How do institutions trade around SEOs?
• Is institutional trading profitable around SEOs?
• Are SEO discounts related to institutional share
  allocations?

3
What do they do and find?

• Empirical Approach
• Examine unconditional and conditional
  relationships between institutional allocations
  and
• Pre-offer trading
• Post-offer trading
• SEO discount
• SEO underpricing
• Trading profits
• Results
• A lot746 hypothesis tests!
• Implications (main economic messages)
• Chemmanur and Jiao (2005) are right!!!
• Institutions produce private information about
  firms that are engaging in SEOs
• Gerard and Nanda (1993) are wrong!
• Institutions are not engaging in price
  manipulation

4
Overview

• There is a lot of information in this paper
• A New Title?
• Everything you ever wanted to know about
  institutional trading around SEOs but were afraid
  to ask
• I did learn a lot about institutional trading
  around SEOs
• I enjoyed reading the paper
• Remainder of the discussion
• A closer look at the results
• Some questions
• Some details

5
Results I

• Question Do institutional investors possess
  private information about SEOs?
• Answer Yes
• Evidence

• Key result ß 0.20 gt 0
• Give institutions more shares, stock price does
  better

6
Results IA Closer Look

• Statistical significance
• When did 10 become the new 5?
• The estimated effect in the full model has a
  p-value of 9.5
• Economic significance
• Dont know the SD of allocation but I do know
• Mean allocation of bottom half of distribution is
  66
• Mean allocation of top half of distribution is
  70
• Tells me SD is pretty small 5? 10?
• What is the marginal effect on annual return of
  increasing the institutional allocation by 10 at
  the mean allocation (68)?
• 27 basis points
• How about a 50 increase in allocation (30 to
  80)?
• 190 basis points

7
Results IA Closer Look

• Regression significance
• The only significant coefficient is the intercept
• Some confusion
• The F-statistics didnt look like they line up
  with p-values
• The p-value on the F-statistics are identical to
  the p-value on the t-test of the coefficient of
  interestevery time!
• What is going on here?
• Authors are performing an F-test on the
  coefficient of interest
• But,
• i.e., youre F-statistic (your t-statistic)2
  and you are repeating the t-test
• Should show the regression F-test
• Given adj. R2 should be significant in raw return
  (due to year dummies)
• Probably insignificant in abnormal return
• In general, econometrics need to be thought
  through more carefully

8
Results II

• Question Do institutions that net buy before
  offering tend get larger allocations?
• Answer Yes
• Evidence
• Key Result ß 2.5 gt 0
• Institutions that buy stock before SEO get larger
  allocations
• Interpretation
• Consistent with information production hypothesis

9
Some Questions

• Where does this information come from?
• What is the mechanism by which mutual fund
  managers obtain private information
• Does someone (management, investment banker) give
  the mutual fund manager this private information?
  If so, what are their incentives to do so?
• Is mutual fund manager just smarter than
  everyone?
• If it is truly private information then is it
  not illegal to trade on this information?
• Results also suggest institutions do not flip. On
  the contrary
• Institutions continue to buy more shares if
• they obtained a bigger SEO allocation, or
• Bought more pre-offer
• When do they make money on their private
  information?
• How long-lived is the private information?

10
Some Further Questions

• What is going on with the firm?
• Do they know about any of this trading?
• Should they care about any of this?
• Do they respond to this behavior
• Are their implications for corporate behavior
• A new strand of research in corporate examines
  the impact of fluctuations in the supply of
  capital on corporate policies
• Microstructure could have a lot to say about this
• What are the frictions preventing a perfectly
  elastic supply of capital
• What are the mechanisms at work

11
Conclusion

• A lot of information in this paper
• I learned a number of new facts
• It might be nice to
• get a better understanding of those facts
• think about other implications of investor
  behavior on corporate behavior