| Index | Recent Threads | Who's Online | User List | Search |
|
|
 |
Portfolio123 » List all forums » Forum: Data related issues » Thread: Possible Errors in P123 database |
|
Total posts in this thread: 23
|
[Request a Feature]
[Post new Thread] [Add To My Favorites] [Watch this Thread] |
| Author |
|
|
dodge1664
Advanced Member
|
Hi Marco, I think I may have found some errors in the P123 database. After getting some rather odd results, I tried to calculate the built in stock factor PEExclXorTTM myself. I believe this formula is correct PEExclXorTTM = Price/(ItemQ(SDBF,0)+ItemQ(SDBF,1)+ItemQ(SDBF,2)+ItemQ(SDBF,3)) For most stocks this formula generates exactly the same number as the built in P123 Stock Factor, however there are exceptions, which may be having a significant impact on my results. An example is Avalon Bay communities Inc (AVB). Portfolio123 returns the following data ItemQ(SDBF,0) 0.25 ItemQ(SDBF,1) -0.09 ItemQ(SDBF,2) 0.52 ItemQ(SDBF,3) 0.23 But according to MSN money, which also claims to use the Reuters data source, the corresponding entries for Diluted EPS Excluding Extraordinary Items are 0.25,-0.09,0.38,0.22. I have checked these numbers with the SEC filings by this company, and they are correct. So it looks like the P123 entries for ItemQ(SDBF,2) and ItemQ(SDBF,3) are returning the wrong numbers. The second problem is that P123 returns 153.42 for PEExclXorTTM using the data from 06/12/10. If I use the numbers P123 quotes for SDBF above I get 113.84, with the last quoted share price being 103.23. If I use the MSN Money entries I get 135.83 as the P/E ratio. So I don't understand how P123 arrives at the result of 153.42. Are you able to shed any light on what's going wrong here? regards |
||
|
|
dcnelson
Advanced Member
|
Are there any other names other than AVB. Some systems and databases report FFO as earnings when looking at REITS. |
||
|
|
marco
|
This is due to restatements that have 6 month values instead of 3 month values. When this happens we need to look at prior quarters to normalize back to 3 months. In this case however there are two 6-month restatements in a row and the logic gives up, spitting out a 6 month value. We'll correct it to look further down the quarters to get a three month value. Thank you for reporting this. We'll announce the fix on this thread. ---------------------------------------- Portfolio123 Staff |
||
|
|
marco
|
Should be ok now. Thanks again. ---------------------------------------- Portfolio123 Staff |
||
|
|
dodge1664
Advanced Member
|
Hi Marco, OK thanks. For AVB, I'm now getting ItemQ(SDBF,0) 0.25 ItemQ(SDBF,1) -0.09 ItemQ(SDBF,2) 0.28 ItemQ(SDBF,3) 0.23 Whereas MSN money and the SEC filings have 0.38 and 0.22 for the last 2 entries. Is that due to the re-statement? At least my formula for PEExclXorTTM now gives the same answer as the built-in stock factor (154). Another company which produces a difference between my formula and the built-in stock factor is Hartford Financial Services (HFI). I get ItemQ(SDBF,0) -0.42 ItemQ(SDBF,1) 1.19 ItemQ(SDBF,2) -0.79 ItemQ(SDBF,3) -0.06 These numbers look correct, however if I put them in my formula, I get a P/E of -311. I know that a negative P/E doesn't make any sense, but why does P123 return PEExclXorTTM = 839.12? Could it be that there are small rounding errors, and that with earnings close to zero, a small difference could produce a wildly different P/E? thanks Dodge |
||
|
|
marco
|
AVB
Looks like they are reporting the original numbers: 0.37671 and 0.22098 I think you mean HIG PEExclXorTTM is still calculated by reuters (we plan to do our own calculations from the line items ourselves to have more control). Obviously reuters has the same line items as us, but we'll need to investigate why they are not returning NA since the sum of the SDBF items is clearly -ve. ---------------------------------------- Portfolio123 Staff |
|||
|
|
marco
|
I've submitted the question for HIG to reuters. ---------------------------------------- Portfolio123 Staff |
||
|
|
mgerstein
Advanced Member
|
Reuters responded by explaining a new approach they are taking to calculating PE. Usually, it's price divided by diluted EPS. They revised this to substitute Basic EPS if use of diluted EPS would not produce a PE (i.e. if diluted EPS is in the red). HIG is an example of a situation where diluted EPS was less than zero and would have shown NM in the PE field. But for Basic EPS, HIG as a profit of a couple of pennies per share. As a result, Reuters computed a PE and that it what was supplied to us. A related issue is the magnitude of the PE they supplied. It's above 800 because the EPS figure was nominal. Traditionally, they, and others, had imposed a cap on PEs and reported anything above the cap as NM. They no longer do that. We are in the process of enhancing our database in a way that will allow us to calculate all ratios in house. There will still always be debatable matters. (Should there be a PE ceiling? If so, what should the maximum be? Etc.) But at least we will be in control and able to make our own decisions, helped by feedback from Portfolio123 members. |
||
|
|
o806
Advanced Member
|
Hi. I know as a data provider you will have to continue supplying PE numbers because that is what so many are used to using. But many problems related to strange PE behavior disappear by using Earnings Yield instead ( EPS/price instead of price/EPS). If you are going to the effort to provide in house calculations for PE, could you do the same for EY (earnings yield)? There are several PE variants (with and without extraordinary, forward PE, etc). It would be nice to have EY with the same variants. Regards, Brian |
|||
|
|
dodge1664
Advanced Member
|
Hi Marco, ok, thanks for investigating that for me. Looks like I'll need to be careful when dealing with extreme values of P/E. regards, Roger |
||
|
|
[Show Thread Printable Version] [Post new Thread] |
