Skip to content
All library documents

Managing Historical Adjusted Close Data After Corporate Actions

Article Quant Q&A · Author: Kermittfrog
Read the original

Summary

The document describes a data management problem in equity research: historical adjusted close prices change when later dividends or stock splits alter the adjustment factors. As a result, the value stored for an earlier date depends on which corporate actions have occurred since that date, so an adjusted series is not inherently append-only.

It considers two possible workflows: recalculate the full adjusted history internally as new dates arrive, or append unadjusted daily prices and refresh the adjusted history less often, such as around corporate actions. The author raises cost, traffic, and data volume as concerns but provides no answer, recommendation, or comparison of approaches. It therefore serves as a useful statement of the versioning and maintenance issue, while leaving the choice dependent on data vendor behavior, precision requirements, and acceptable update latency.

Key ideas

  • Later dividends and splits can revise adjusted prices for earlier dates.
  • An adjusted historical series may need to be refreshed rather than stored as append-only data.
  • One approach is to recalculate adjusted prices internally from underlying prices and corporate actions.
  • Another approach is to update the full adjusted history periodically, accepting a delay between refreshes.

Tags

EquitiesBacktestingStatistics
Full text
# Adjusted close time series: Frequency of historical updates


# Adjusted close time series: Frequency of historical updates












When working with stock market data for strategy / analysis purposes, I am well aware that I have to distinguish between unadjusted and adjusted prices.

I understand that historical adjusted prices will be re-adjusted whenever a new event (dividend, split) took place; hence the adjusted close data for any given historical date is a function of that day's close level and any corporate action since then until today. I.e.: the adjusted price some historical date varies with calender time of observing the time series.

Hence, adjusted price data cannot be meaningfully stored in an append only fashion, i.e. I cannot load new adjusted price data into my database, but I have to re-create (as in: obtain and store) the whole adjusted price time series each day; or at least on each date where a corporate event took place.

This creates a little bit of a challenge for me: Obviously, downloading the whole time series of adjusted prices on each trading date is quite consuming in terms of data, traffic (and cost) and I am wondering whether it is feasible / economical to:

- re-build the adjusted price series internally (as in: in my database) with each new date, or

- to reload a full adjusted historical close time series at a lower frequency, optimally at or around capital action dates. I.e. by appending un-adjusted prices daily as usual, but then recreating the full adjusted price series only once per month / quarter / year.

I have not found any thoughts on this in this forum - and I'd be very happy to get ideas, hints, or to discuss possible approaches to this. Thanks in advance.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.