1
$\begingroup$

I am merging two datasets:

1) One which includes Reuters US companies' RICs and the number of news linked to each company;

2) Compustat US, taking data from 1996 to 2018.

I have few questions:

  • In order to merge them, I consider that Reuters RICs without the stock exchange (e.g. AAPL for Apple) correspond to Compustat Company Ticker, is this correct? Looking through the Reuters RICs of companies mentioned in news during 2019, I found three exceptions to this rule, one of them is Moog Inc, whose Reuters' RIC is MOOGa, while Compustat Ticker is MOOG.A. The other two cases are different in the same way (JWa - JW.A and BRKb - BRK.B). I would like to know whether this is actually true.

  • Do the Reuters RICs change over time? Are they re-used or are they uniquely assigned to firms from 1996? For example, if I find a company's RIC in 1996, is it sure that this RIC will be always attached to the same company and will not be assigned to another company in the following years? In case RICs actually change, is there a guide with all the changes (possibly a guide that does not require to login in any website)?

Thanks!

$\endgroup$
  • $\begingroup$ Please accept the response if it answers your question. If not, please respond or clarify your question. $\endgroup$ – Chris Jun 24 at 6:20
1
$\begingroup$

You've stumbled upon one of the fundamental challenges to working with historical equity data.

The quick answers are that (1) yes, identifiers can and do change over time; (2) there isn't a clear/obvious/'right' method for mapping securities in creating a security master, at least in part due to differing treatment of corporate actions; (3) the general approach involves creating a single unique security identifier and then mapping various sources to it, after which you're able to move back and forth between sources using the unique ID.

In your case, linking on RIC/ticker should be good enough, and it'll become clear soon enough where there are gaps that need to be accounted for. I'd also generally recommend using something non-text-based (eg, SEDOL, ISIN, CUSIP) as matching on text-based IDs can be a little tricky (as you'd noted accounting for different share classes of the same company).

$\endgroup$
  • $\begingroup$ I did not accept the answer because it seems a bit general to me (maybe my question was also general). For example: "it will become clear soon enough where there are gaps that need to be accounted for", what do you mean by this? Actually what I am a bit worried about is that I won't notice wrong links. Also, you say that I should link RIC with ticker, but that it would be better to use something non text based, so how do I combine the two? Thanks for the reply anyway! $\endgroup$ – Esperanta Jun 26 at 13:00
  • $\begingroup$ @Esperanta, that's probably in part because there isn't a 'right' answer for what you're encountering. What I mean is you'll see which securities don't map over time. Using tickers as primary keys can be challenging because different vendors may have different versions (eg, ticker, ticker+exchange, ticker+assorted info) which can make mapping a headache. A CUSIP is always a CUSIP. Even with SEDOLs, the biggest complication is whether the check digit is included or not, which is easy enough to account for. $\endgroup$ – Chris Jun 30 at 3:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.