Why Cointegration Tests Should Use Adjusted Prices

Summary

The note addresses whether cointegration tests on dividend-paying stocks should use closing prices or adjusted closing prices. Its answer recommends adjusted prices, and ideally a series reconstructed from total returns, so corporate actions do not create artificial discontinuities in the data.

A stock split illustrates the issue: a ten-for-one split sharply changes the quoted share price even though the investor’s economic exposure is continuous. A relationship tested on unadjusted prices can therefore reflect the mechanics of the split rather than a genuine long-run link between assets. The same concern applies, less dramatically, to dividends paid on different dates. The answer is brief and gives no empirical comparison or details about adjustment conventions, so researchers still need to ensure the adjusted series consistently reflects their intended return measure.

Key ideas

• Use adjusted price series when testing cointegration between stocks affected by dividends or splits.
• A split can create a large artificial price discontinuity in unadjusted data.
• Total-return-based series are an ideal input when constructing price histories for such tests.
• Adjustment conventions should be consistent with the research question.

Tags

# Should cointegration be tested using close or adjusted close prices?


# Should cointegration be tested using close or adjusted close prices?












When doing cointegration tests should I use the adjusted close price or just close price for the time series? The dividend of each stock is on different dates and can cause jumps in the data.

## Answer by Tal Fishman (score 6, accepted)

https://quant.stackexchange.com/a/2233

Such tests should always be done using adjusted prices. In fact, ideally, you should reconstruct your own price series using the total returns series. To see this, suppose you have a 10:1 split rather than a relatively small cash dividend. Then it is clear that the cointegration relationship can only hold with respect to the adjusted series.

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.