Darryl Laws

 What robustness checks were conducted? Malmendier and Tate discuss the robustness of their results to the changes in the empirical model. Simply stated they focused upon the baseline estimates of binary regression equation provided in regression method use (above). First, they considered; Is the Option in the Money? Their CEO stock option long holder measure of overconfidence is they classify a CEO as overconfident if he ever holds his company stock option(s) until expiration. The less an option is in the money, the less delayed exercise indicates likely overconfidence. As a robustness check of their measure, they require that the option that is held until expiration be at least x% in the money at the beginning of its final year. They vary x between 0 and 100 by increments of 10. As they increase x, the classification as overconfident becomes more restrictive. Concurrently, they hold the definition of rational option exercise behavior constant. (Example: they require that the CEO never holds an option until the final year.) This restriction keeps the comparison group the same across all regressions. Their exhibits in their appendices presents the coefficients on those modified proxies for overconfidence in estimates of the general regression equation. In the logit and random effects logit specifications, the overconfidence coefficient is roughly constant since they vary x. In the fixed effects logit specification, the coefficient appears to modestly increase. They conclude that the effect of long holder on acquisitiveness is not driven by CEOs with options. A natural alternative is to require that CEO always hold their option packages until expiration. Malmendier and Tate’s (2008) rational CEO model requires the non-overconfident CEOs to be habitual be early exercisers of their personal options. 

Strengths and weaknesses of the article.  

Weakness. –

• Assumption that there is symmetrical in information amongst the buyer and seller. This is not always true. It is more often in a public company context but certainly not in a transaction comprising private companies. I have spent thirty-five years in private equity buyouts making sure that I had the dominant strategy via better information and understanding of the target’s synergies with portfolio companies, when it comes to negotiating a cheaper price for an acquisition target. 

• Assumption that the capital markets are always efficient.    

• The goodness of fit method is not described. Log likelihood may have been a better measure of fit.

Strengths.

• Research design – explanatory mixed methods convergent research design.

• Data collection – Forbes 500 database, archival, regulatory filings, CEO data,

• Independent and dependent variable selections – were well depicted. 

• Impact - on organizational design and CEO agency relationship to shareholders and owner.

• Impact – mergers and acquisitions decision-making.

• Impact – CEO stock option compensation. 

• Impact – on the use and deployment of internal company resources (cash, leverage debt). 

Conclusion. Because of the restrictions that Malmendier and Tate’s tests impose on sample size are so severe I am not confident that their results hold true. It is my opinion that the authors’ explanation of the regression methodology used was poorly communicated and was insufficiently numerically depicted. For instance, the goodness of fit was not narratively discussed openly but appeared out of nowhere in the correlation tables that the authors presented in the appendices.  On the other hand, their presentation of the correlation with company characteristics and CEO characteristics was definitively presented.