US equities daily (+ adjustments)

Discussion in 'Data Sets and Feeds' started by laplacian, Aug 28, 2019.

  1. Hi all,

    I am looking for US equities daily price data (history to present), with adjustments for corporate actions (adjusted close price taking into account for splits, dividends,...), ideally with corporate actions data itself too but not essential. The aim is to get a history of the top stocks on a rolling basis since the 1990s, in total a few 1000s stocks.

    I believe that Reuters (now Refinitiv) is a standard feed for trading firms that only care about daily resolution. Bloomberg seems less popular in this category (more geared towards intraday and their better data API is much more expensive). Quandl has cheap datasets but some I used are bad quality (esp the adjustment factors) so I'm looking for more established.

    Any other suggestions? My expectation is that a good quality feed (and history) in this category is roughly $20k/year.
  2. Robert Morse

    Robert Morse Sponsor

    Those would be the two. Refinitiv cost less. Ask for a WebEx demo of each to view with a salesperson.
  3. drm7


    I don't know if Portfolio123 provides you the raw data, but you can code pretty much whatever you want on their platform to analyze it. Price-related and fundamental data. Uses the Compustat database and is fully adjusted. It even has data on discontinued tickers, which helps mitigate survivorship bias.

    (I have no affiliation with Portfolio123.)
  4. drm7


    And I saw on your other thread about order execution. Portfolio123 has an API with Interactive Brokers which will automatically send orders based on your criteria. I have not used it and don't know how it works, but it is available.
  5. dinn13


    QA Direct through Refinitiv might be a good option. Used it 10 years ago and they did have corporate action adjusted total returns. It came in the form of a sql database which I didn't like. Pricing likely has changed since then but we were paying much more than 20k/year.

    Currently I use Bloomberg back office files which is provided end of day in a pipe delimited file. Also have access to datascope from refinitiv which has a lot of history and appears to be as good as bloomberg. Neither provides a total return calculation so I do that myself. Bloomberg with history from the 1990's would cost multiples of 20k so probably a no go.

    Are you going to have a risk model? If you are and are considering Barra then they also provide the market data they use and if I recall correctly do provide the adjusted returns.
  6. Metamega


    If you just need historic EOD and updated in evening of current day, check out Norgate Premium Data.

    They have a delisted stocks package aswell.

    I know their plugin for Amibroker even has a historical index constituent watchlist. Not sure if this is a function in their downloaded. Refinitiv and Bloomberg might be overkill depending on your current needs.

    Norgate historical and IQFeed for live would be a more affordable approach.
  7. %%
    Most any good charts have splits/dividends history. BUT many of them still lie; not many or any report C reverse split 10 times so that dip to $48 was really $4.80. Earnings are still the main driver+ sometimes analysts; not BS like press reports.Best also to keep your own notebook - i do..................................................................................................................
  8. For a risk model I'd rather use free risk factor data (Kenneth French data lab seems good quality) and fit my own model - I'm not aware of an advantage of using Barra (which costs a lot) for a portfolio of a few hundred stocks of US equities, which is a standard scenario already covered by the mass of risk model literature.
  9. dinn13


    Yeah totally, I use my own risk model for calculating residuals for forecasts and optimization. Only use barra for tracking factor exposures since our investor also uses it for tracking our risk. I find it to be fairly cheap but I guess that's just relative to other expenses and pnl.
  10. djames


    Hi @laplacian, i'm also interested in fitting my own model to 'open source' risk factor data, could you provide a bit of direction on this? I'm trying to build up my risk modelling experience.
    #10     Aug 29, 2019