In another thread Jerome posted his work comparing Sales numbers between the two data providers. I noticed when looking at the list it seemed like there were quite a few Financial companies (which we know FactSet treats differently) and REITs. One of the REITs that stuck out in my own analysis was VNO. If we look at the Trailing 24 months of Sales between the two engines we get the following results
Legacy = $3810
FactSet = $7675
So FactSet shows approximately 2x the Revenue of Compustat. Who is correct? If I look at the 10Qs and 10Ks for VNO I get a Trailing 24 months of Sales of $3454. Clearly, Compustat is much closer to this number, so where does the discrepancy come from? Digging in a little deeper to the VNO Financial Statements it seems clear that FactSet is including the following line items as top-line Revenue
Net gain on transfer to Fifth Avenue and Times Square JV
Net gains on disposition of wholly owned and partially owned assets
This appears to account for the difference in the Revenue values between the two data providers. It is pretty clear from the Financial Statements that VNO does not consider these asset sales to be Revenue, nor given all I know of accounting should they. The question is why does FactSet do this, and is there anyway that P123 can correct for this? It is reasonable to assume that FactSet is doing similar things with the other REIT’s which is why a disproportionate number of them seemed to show up on Jerome’s list. Note that because these asset sales are “lumpy” they will completely screw up Revenue Growth calculations, which is why VNO was flagged in my analysis.
Hope someone finds this useful.
Cheers,
Daniel