Is this real? Unbelievable. How could they be so ...?
As 275c mentioned, Andrew Bird and Stephen Karolyi never released their so-called "replication kit," which makes their claims worthless.
Also, at some point after EJW article, SK+AB posted a response to SSRN where they argued that Young made a coding error and that if you fixed it, you could use his data and programs to replicate the SK+AB results.
EJMR found out that SK+AB's "fix" sorted the obs in the wrong order. Yes, you read that right. They actually forgot that Stata sorts in ascending order by default. Shortly after the EJMR posts were made, SK+AB revised their SSRN response by deleting all of their faulty code and tables.
See for yourself, https://www.filedropper.com/byaccountingresponseorigI admit I followed this case only superficially, and I found the whole "replication kit" and "my notebook died" rather unconvincing. But, in this letter - which I have not read until today - they claim the "replication kit" is pretty much the original sample and programs fully recreated. And the results - apparently vetted by people outside of CMU - are nearly identical to those published in the original article. Which I presume includes that table containing the results that were supposed to lose significance based on the EJW article by Alex Young.
If that were the case, and the fully-replicated results are indeed the same: while I see that SK+AB might have completely brushed off Young (which they should not have done), where is the fraud?