FactSet data problem update

Starting a separate thread now that we know that the problem is on FactSet side.

Interim period 3/31/2023 was deleted sometime around March 4 from the historical dataset for a large number of companies. As to how this happened they are not saying. We think around 2000 companies were affected in the regions we have. We reported the problem yesterday giving NFLX and META as examples, and they only fixed those two.

I was just told that it has been escalated to top priority. I provided them with our full findings so that they don't limit their analysis to the few examples we provided initially. They cannot tell me right now the full scope, like: Is it only the 3/31/2023 interim? Is current data affected? Are other dataset like estimates affected?

As soon as I know more I'll report here.

Thanks

5 Likes

To clearly see if the data is still missing you can run a screen backtest for the SP500 for the past three years with one single rule.

SalesTTM=NA

Or just use my public screen. If you see a big jump like this then the problem persists.

You can also run the screen above with an As-Of Date of 1/1/2024 and see that 145 stocks in the SP500 have NAs for SalesTTM. It should be 2 or less typically.

Seems OK finally. Back to 2 or less NAs for SP500

3 Likes

Do you think the issue is resolved?

No official word from them yet. We can only judge with tests like the one above.

What about European stocks ?

1 Like

CEDER:SWE might be a different situation. Still caused by a missing interim but not in 2023. It's missing 2025-09-30.

Could be isolated to CEDER or could be the same issue.

I will report it. Thanks

1 Like

Please note the we have not received official confirmation that the problems have been fixed. I will change the banner warning to reflect that.

1 Like

Might be unrelated. In our DB snapshot before the fat finger event, the 2025-09-30 interim for CEDRE was missing as well. So it could be the same reoccurring problem we see in Europe where sometime interims are not picked up. Very annoying that they don't have processes to detect that they are creating holes in the data when they added 2025-12-31

1 Like

Latest update

We have not received official confirmation that the problems have been resolved. However, since the missing interims around 2023 seem to have been fixed, we will proceed with Monday's rebalances.

Let us know what you think

It's an old problem

FYI. No official confirmation that all the problems were resolved.

3 Likes

Hi Marco,

To what extent does this affect CompuStat subscribers? I notice the update is still “working” this morning.

Thank you.

Hugh

If we don't get prices (both Compustat and Factset use ICE data) everything grinds to a halt

We reported several European stocks that had gaps in the interims. Looks like they were fixed, including CEDER:SWE

Things seem ok now.

1 Like

When I reran my sim this morning, my annual returns and holdings changed starting from 2019 to present. Prior to 2019 everything was fine. No factors in my sim changed. My 16 year annual CAGR changed by over 7%.

We never got final confirmation that everything was resolved. Like all vendors they are probably fixing big stocks first. Your particular sim is very sensitive to any changes:

  • it has a CAGR of 70+%
  • It only holds 6 stocks
  • Lots of smallish stocks

Any small variation will can have a significant impact.

As soon as I hear official confirmation they are done I will report it.

Thanks

2 Likes

I rerun some of mine in US. No significant move.

1 Like

I have Canadian stocks as well in my sim. Yesterday everything was fine with the sim, then today starting in 2019 the portfolio changed. There are 6 more trades in the system that weren’t there yesterday.

@rtelford did you notice something strange in Cad stocks? Or other users... interesting to see

I also noticed yesterday morning when I downloaded the rank data it was dated March 11. I just downloaded the rank data now, and its also dated March 11, yet its different data.