7,612
edits
Changes
Jump to navigation
Jump to search
The old working directory was E:\McNair\Projects\CompAcqs\. There's also a CEO.log file that details results run on the old dataset, for information only.
*Spec and build summary statistics and regression tables==Old Notes==
CEO Compensation and Returns to Public and Private Acquisitions: Testing the Shareholder Value Hypothesis (view source)
Revision as of 18:37, 8 February 2021
, 18:37, 8 February 2021no edit summary
<pdf>File:Brander_Egan_Endl_(2019)_-_Comparing_CEO_Compensation_Effects_of_Public_and_Private_Acquisitions.pdf</pdf>
==Database, dataset and scripts==
MasterV1-1.txt
There's also a CEO.log file that details results run on the old dataset, for information only. ==To Current Notes== The paper was submitted to the [https://www.mdpi.com/journal/jrfm Journal of Risk and Financial Management (JRFM)], an open-access journal with fast turn-around times on January 17th, 2021. On February 8th, 2021, we received an R&R with a 10-day revision window. This section provides notes on the feasibility of the R&R. A key issue concerns requests for us to update the data, which:*Isn't a live dataset on my current [[Research Computing Hardware]]. *Requires Cumulative Abnormal Returns (CARs) that were previously calculated as a part of the [[Winner's Curse in Acquisitions (Academic Paper)]] that was accepted for publication in 2017.**Uses both public and private acquisitions data, that were loaded as a part of [[VCDB20]] but may need more variables.*Needs ExecComp, Compustat, and CRSP data from WRDS (I have access to all until March). I didn't notice any requests from the reviewers to handle multiple acquisitions on the same day, which doesn't appear to have been fixed in the previous version. However, I can see we are aggregating acquisitions over the year, so this is somewhat moot in the averages and totals. ===Estimating a new build=== ====Variables from the Winners dataset==== The biggest hurdle to updating the data is that we pulled many of the variables from the Winner's dataset: --Builds the Acqbase table where the unit of observation is an acquirer year DROP TABLE Acqbase; CREATE TABLE Acqbase AS SELECT acusip, year, sum(car5) as car5tot, avg(car5) as car5avg, sum(car5*tpublic) as car5totpub, avg(car5*tpublic) as car5avgpub, sum(car5*(1-tpublic)) as car5totpriv, avg(car5*(1-tpublic)) as car5avgpriv, avg(sharesoutstanding) as sharesoutstanding, sum(car5*sharesoutstanding) as valctot, avg(car5*sharesoutstanding) as valcav, sum(tpublic) as numacqpub, sum(1-tpublic) as numacqpriv, count(tname) as numacq, sum(transactionvalue) as transactionvalue, sum(transactionvalue*tpublic) as transactionvaluepub, sum(transactionvalue*(1-tpublic)) as transactionvaluepriv, sum(CASE WHEN transactionvalue IS NOT NULL THEN 1::int ELSE 0::int END) as numwtv, sum(CASE WHEN transactionvalue IS NOT NULL THEN (1::int)*tpublic ELSE (0::int)*tpublic END) as numwtvpub, sum(CASE WHEN transactionvalue IS NOT NULL THEN (1::int)*(1-tpublic) ELSE (0::int)*(1-tpublic) END) as numwtvpriv, sum(CASE WHEN transactionvalue IS NULL THEN 1::int ELSE 0::int END) as numwotv, sum(CASE WHEN transactionvalue IS NULL THEN (1::int)*tpublic ELSE (0::int)*tpublic END) as numwotvpub, sum(CASE WHEN transactionvalue IS NULL THEN (1::int)*(1-tpublic) ELSE (0::int)*(1-tpublic) END) as numwotvpriv, sum(coalesce(transactionvalue,enterprisevalue,equityvalue)) as targetvalue, sum(coalesce(transactionvalue,enterprisevalue,equityvalue)*tpublic) as targetvaluepub, sum(coalesce(transactionvalue,enterprisevalue,equityvalue)*(1-tpublic)) as targetvaluepriv, min(anaic::int) as anaic, sum(tit) as numacqit, sum(tvc) as numacqvc, sum (stockswap) as numstockswap, sum (terminationfeeacquiror) as numterminationfeeacquiror, sum (terminationfeetarget) as numterminationfeetarget, sum (tbalancesheetprivate) as numtbalancesheetprivate, sum (tincomestatementprivate) as numtincomestatementprivate, avg (agovgoldenparachute) as agovgoldenparachute, avg (agovlimitabilityactbywritconsent) as agovlimitabilityactbywritconsent, avg (agovmajvoterequirement) as agovmajvoterequirement, avg (agovresignrequiredonmajorityvote) as agovresignrequiredonmajorityvote, avg (pcstock) as pcstock, avg (pccash) as pccash, avg (drivingdur) as drivingdur, avg (distance) as distance, sum (CASE WHEN pcstock=100 THEN 1::int ELSE 0::int END) as numallstock, sum (CASE WHEN pccash=100 THEN 1::int ELSE 0::int END) as numallcash FROM Winners GROUP BY acusip, year; --18341 The variables, along with their new sources and comments, are as follows:*acusip, year, tname, tpublic, anaic, transactionvalue, enterprisevalue, pcstock, pccash -- A part of the [[VCDB20]] build*tit, tvc -- Added with lookup tables based on NAICS (Est. 30 mins)*car5, sharesoutstanding -- Would need to be calculated used GVKEY based data pulled from CRSP, some pre-processing, and a STATA script, then a load. (Est. 6hrs, assuming I can find the old code.) *equityvalue, stockswap, terminationfeeacquiror, terminationfeetarget, agovgoldenparachute, agovlimitabilityactbywritconsen, agovmajvoterequirement, agovresignrequiredonmajorityvote -- Could be pulled from SDC but I don't think they are used*tbalancesheetprivate, tincomestatementprivate -- Built from a COMPUSTAT pull (Add 30 mins)*drivingdur, distance -- Determined using a Google Maps API script. (Est 2hrs + $100 and 5hrs to run). ====Other things==== On top of this. we'd need to dothe following:*The ExecComp pull (1hr)*The COMPUSTAT pull (1hr)*Rework the SQL (2 hrs) So, my best guess is: 1/2 + 6 + 1/2 + 2 + 1 + 1 + 2 ==13hrs...
===Same Day Acquisitions===
