# comparing total returns from various data vendors

I need to use various data sources to cover all of my data, and I am concerned by the discrepancies in total returns. Data vendors were helpful, but their simple documentation did not help resolve why there can be differences for some months for some ISINs but not for other months. If you are familiar with these issues, please comment on the example data points below.

I am calculating calendar-month returns (in Swedish Kronen) via

• TOT_RETURN_INDEX_GROSS_DVDS from Bloomberg (taking the ratio of consecutive month-end values),
• RI of ThomsonReuters Datastream (again, a ratio),
• FG_TOTAL_RETURNC of FactSet (directly),
• Month End Return of Morningstar (directly),

I would love to see consistent returns, comparable to what MSCI bases their total return indices on. Which vendor corresponds to this?

I think all vendors promised to assume all reinvestments of dividends and capital distribution for this data (though for funds fees and costs might still change things). Meanwhile for

H&M stock (HENNES & MAURITZ B) in April 2000, I see

• Bloomberg: -.0027076
• ThomsonReuters: -.0023833
• FactSet: -.00238335
• Morningstar: .0104167

For an SEB Sverige fund, a difference shows up in June 2000: .0304632 from Bloomberg and .0494242 from Morningstar.

And the fact that sometimes Morningstar produces lower numbers, sometimes higher numbers, and for equities, not stocks, does not suggest a definitional issue about reinvesting or paying dividends, e.g.

For Ericssson stock (ERICSSON (LM) B): Nov 2000: BB: -.1390977; TR DS: -.1390979; FS: -.13909775; MS: -.0912699 but Sept 2001:BB: -.2431374; TR DS: -.2431377; FS: -.2431372; MS: -.2873786;
Oct 2001: BB: .1968915; TR DS: .1968912; FS: .1968913; MS: .2588556.

-
@kristine - I ask to be enlightened, not to argue: It was my impression that data requests were discouraged (the language itself seems to suggest a soft limit). Do you think all questions relating to data and data vendors should be closed? (And I'm willing to close my own question about data vendors, myself.) IMO, a database engineer doesn't necessarily know the business reasons for data discrepancies, and professional quants can benefit from clarity that people in-the-know can provide, where no documentation exists. Is it more that Money.SE would be more appropriate? – Andrew Cheong Jul 19 '13 at 13:23
I also think that this is not a question for a data engineer as the understanding sought is for the business rationale for why the total returns my differ. It is not clear a priori that this might not be due to difference in the quantitative process, e.g. reinvestment at the beginning-of-day vs end-of-day. – snth Jul 24 '13 at 7:36

The best way to answer the question is to look at the data. For example, on H&M in April 2000:

       Close Price    Div
31/03      240
13/04      236
14/04      225        1.35
28/04      238


ThomsonReuters, Bloomberg and Factset do the following calculation for the return (+/- rounding):

r = 236/240 * (225 + 1.35)/236 * 238/225 - 1
= -0.24%


This is the value I get from Bloomberg using the TOT_RETURN_INDEX_GROSS_DVDS field so I'm not sure how you obtained the -0,27%.

You may have got your number from using the PX_LAST field adjusted for dividends. It is calculated based on an adjustment factor for the dividend:

f = 1 - 1.35/236
= 0.9942797


They then adjust the return on the ex-date by that factor:

r = 236/240 * 225/(236*f) * 238/225 - 1
= -0.26%


which is close to the number you got.

-