Do Auctions induce a winner’s curse? New evidence from the ...
Transcript of Do Auctions induce a winner’s curse? New evidence from the ...
Do Auctions induce a winner’s curse? New evidence from the corporate takeover marketby Audra L. Boone, J. Harold Mulherin
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
Structure
1. Auctions and the Winner’s Curse
2. Impact of the Winner’s Curse on the corporate takeover market
3. Innovations by Boone and Mulherin
4. OLS results and two-step regression models and results
5. Simultaneous equations model and results
6. Contribution
2
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
Auctions and the winner’s curse
3
§ Auction winners pay too much § The magnitude of the winner’s curse depends on the level of:
§ Competition (Kagel and Levin, 1986)§ Uncertainty in asset value (Bazerman and Samuelson, 1983)
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
The impact of the winner’s curse
§ This motivates two opposing theories on market reaction towards the curse§ Information asymmetry => marektwide mistake (Akerlof, 1970)§ Market participants have incentives and the ability to resolve the asymmetry (Barzel, 1977)§ Must be resolved empirically (Thaler, 1988)
§ Winning an auction induces non-positive returns (Thaler, 1988 ; Roll 1986)§ Hubris hypothesis - Overconfident managers overbid
§ Breakeven returns are a result of competitive markets (Mandelker, 1974; Asquith, 1983; Travlos, 1987)
4
Test winner’s curse hypothesis directly in an event study setup -If the winner’s curse holds then bidder returns must be inversely related to:
§ Competition § Uncertainty in the value of the target firm
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
Innovation in testing the winner’s curse for corporate takeovers
Approach of Boone and Mulherin:
§ New Data from the Securities and Exchange Commission§ Measure competition with the # of publicly announced bidders and # of bidders in the private
process § Competition is then categorized into negotiations (one bidder signs conf. agreement) and auctions
(multiple bidders sign conf. agreement) § 145 auctions and 163 negotiations from 1989-2000
§ Different Methodology § Model the endogeneity between takeover competition and bidder returns via two-stage models§ Model the effect of the choice of investment bank
5
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
OLS analysis of bidder returns
§ Standard Cross-sectional OLS Model:
𝑦" = 𝛼 + 𝛽'C) + 𝛽*U) + ∑𝛽-𝑋",- + 𝜖"
§ 𝑦" - Bidder returns in excess of market returns over the event window
§ C) - Dummy for competition (0 if negotiation, 1 if auction)
§ U) - Uncertainty proxied by the ratio of intangible assets
§ 𝑋",- - additional control variables
§ OLS regression is run as a mean to compare results against the following two stage models controlling for endogeneity
6
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
Controlling for Endogeneity 2 Step Regression
§ First Step: Run a probit model to predict sales procedure (negotiation or auction)
P)=Pr(C)=1|X))=Φ(X)′β+ε)
§ The model estimates the probability for the sales procedure§ The greater the uncertainty in the asset being sold, the greater the benefit of competition in
reducing asymmetric information§ States with stronger anti-takeover law provide more bargaining power to targets § The number of potential bidders is directly related to the number of bidders§ Etc.
§ Second Step: Run the bidder return regression with predicted competition (P)∗)
𝑦" = 𝛼 + 𝛽'𝑃)∗ +𝛽*U) + ∑𝛽-𝑋",- + 𝜖"
7
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
First step Probit and second step OLS vs single step OLS
8
Variable First Step Probit(Dep. variable - procedure*)
Second Step OLSDep variable bidder return
(+1,-1)
One Step OLSDep variable bidder retrun
(+1,-1)
Intercept 1.249 0.213*** 0.213***Procedure* -0.007Auction -0.010
Intangible Assets 0.504* 0.033** 0.033**
Target CEO gets job -0.568***Anti-takeover state -0.47*Industry Count 0.15*Unsolicited 0.663*** -0.024* -0.024**Stock -0.350* -0.037*** -0.038***Bidder hires top IB -0.477*** 0.028*** 0.036***Bidder hires no IB 0.037*** -0.016*…Model p-value 0.000 0.000 0.000
*,**,*** indicate significance at the 10%, 5%, 1% level respectively
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
Control for endogeneity via simultaneous equations
§ The choice of sales procedure and bidder returns my be jointly determined§ Two stage simultaneous equation may be a more appropriate framework
First Stage:§ Estimating Bidder Returns
𝑦" = 𝛼 + 𝛽'𝑃) + 𝛽*U) + ∑𝛽-𝑋",- + 𝜖"
§ Estimating Sales Procedure𝑃" =Pr(𝐶"=1|𝑋")=Φ((𝑋"′β+ε))
Second Stage:
I. 𝑦"*= 𝛼 + 𝛽'𝑃"∗ +𝛽*U) + ∑𝛽-𝑋",- + +𝜖"
II. 𝑃"∗=Pr(𝐶"=1|𝑋")=Φ(𝑋"′β+ε) ,𝑦"*∈𝑋"
9
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
Results two stage Simultaneous Regression
10
Variable Bidder Returns Sales ProcedureFirst Stage Second Stage First Stage Second Stage
Intercept 0.257*** 0.192*** 2.135** 0.105Procedure* 0.002Returns* 4.831
Intangible Assets 0.031** 0.031** 0.473* 0.367
Target CEO gets job -0.011 -0.520*** -0.508***
Anti-takeover state 0.003 -0.431* -0.476**
Industry Count -0.007 0.134 0.192**Unsolicited -0.029** -0.027* 0.645*** 0.795***Stock -0.062*** -0.036*** -0.324 -0.204
Bidder hires top IB 0.030** 0.030** -0.370** -0.572***
Bidder hires no IB 0.034*** 0.035*** 0.271…Model p-value 0.000 0.000 0.000 0.000
*,**,*** indicate significance at the 10%, 5%, 1% level respectively
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
Contribution
Contribution:
§ Introduction of a new data set on corporate takeover competition§ Private auctions as previously unobserved competition factor
§ Regression designed to control for endogeneity between competition and bidder returns
§ Strong evidence against the winner’s curse effecting the corporate takeover market § Competition does not effect bidder returns§ Uncertainty is positively correlated with bidder returns
§ Evidence against the hypothesis that prestigious banks encourage overbidding
11
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
Appendix 1 - Sample Statistics
12
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
Appendix 2 - Tests with further Proxies of Takeover Competition
13
OLS: Simultaneous Equation:
Schneider 2017 Do auctions induce a winner’s curse? New evidence from the corporate takeover market Boone and Mulherin (2006)
Appendix 3 - Tests with different event windows
14
2 Step Estimation results: