7,649
edits
Changes
Jump to navigation
Jump to search
CEO Compensation and Returns to Public and Private Acquisitions: Testing the Shareholder Value Hypothesis (view source)
Revision as of 19:45, 9 February 2021
, 19:45, 9 February 2021→Restoring the old data
*The CARs appear to have been processed with E:\mcnair\Projects\Winner's Curse\Stata\EdsRegs.do, which takes Estimation5K.txt as an input.
I made compacqs folders in E:\projects and /bulk/ for use, created a new '''compacqs''' database, and restore the tables from the last backup. The pg_restore threw some errors (the new server doesn't have a plpythonu extension control file) but was otherwise fine. I put the latest version of everything I could find in an old subfolder on E.
The last version of the dataset produced by BuildCompAcqs.sql was 2-3, which was processed by CEOAug22.do to build CEOAugustv2.log. However, I don't see any outregs or other automated table generation and most of the regressions are missing.
There's a do file in the dropbox, '''CEOOct.do''', that is identical to CEOAug22.do up until around line 100. It looks like the last version of the STATA code. However, it loads AcqCEOSep, which must have been a .dta produced by Jim.
I built the following:
*MasterV3-0.txt from the master table of the reloaded dbase, with the code in Revision.sql
*Revision.do, based on CEOOct.do, and loading MasterV3-0.txt
I also copied them to the dropbox.
===Estimating a clean new build===
