Stocks / ETFs

All stock and ETF datasets include both unadjusted data, split-only adjusted data, and split and dividend adjusted data.

Split Adjustments

For stock splits (and reverse splits) the prior data is adjusted by the split ratio. For example, on Aug 31 2020 Apple's stock was split 4:1 resulting in the below price data:

{Timestamp, Open, High, Low, Close, Volume}
2020-08-28 19:59:00,501.8,502.0,501.71,501.98,5786
2020-08-31 04:00:00,128.0,127.29,125.6,126.0,57640

To avoid the price drop from the split the prior data is adjusted by the split ratio:

{Timestamp, Open, High, Low, Close, Volume}
2020-08-28 19:59:00,125.45,125.5,125.4275,125.495,23144
2020-08-31 04:00:00,128,127.29,125.6,126,57640

Note the volume is also adjusted by the split ratio (except volumes are adjusted for by multiplying by the split ratio).

Dividend Adjustments

Dividends also have a distorting effect on prices and are adjusted for. On the ex-dividend date, stock holders are no longer entitiled to receive the dividend and as such the price should be adjusted since it no longer includes the cash payment of the dividend.

The adjustment for dividends is a little more complex since the dividend is a cash amount and not a ratio as with splits. The adjustment factor is:

1 - (Dividend Amount / Prior Day's Close Price)

For example, on Nov-6-2020 Apple stock went ex-dividend and shareholders were no longer entitled to the $0.205 per share dividend (note this is the ex-dividend date, the actual payment date of Nov-12-2020 has no price impact).

The unadjusted price data is :

{Timestamp, Open, High, Low, Close, Volume}
2020-11-05 19:59:00,118.0500,118.0500,118.0000,118.0000,15203
2020-11-06 04:00:00,117.7700,117.7700,117.0700,117.0800,4692

The adjustment factor is :

1 - (.205 / 118.00) = 0.99826

This factor is then multiplied to all the prior price data. Note, for dividend adjustments the volumes are not adjusted. Thus the dividend adjusted data is:

{Timestamp, Open, High, Low, Close, Volume}
2020-11-05 19:59:00,117.8449,117.8449,117.795,117.795,15203
2020-11-06 04:00:00,117.7700,117.7700,117.0700,117.0800,4692

The dividend adjustment is applied to all prior prices in the series from the first of the series to the last date before the ex-dividend date.
Note that if the price has been adjusted for a split then the dividend should be adjusted by the same ratio in the adjustment factor calcuation.

Dividend adjustments can lead to differences between prices between different data vendors. The most common cause is some data vendors use the price at the close of regular trading hours to calculate the adjustment factor. At FirstRate Data we use the last available traded price on the day prior to the ex-dividend date to calculate the adjustment factor. This gives a slightly more accurate estimate of the price impact of the stock going ex-dividend.

---

Futures

All futures datasets contain continuous adjusted and continuous unadjusted data (individual contracts are also included for completeness).

Both continuous series are created by chaining together front-month contracts for the future, with the roll date being the end of the final trading day for the expiring front-month contract.

Futures Adjustment

An issue with continuous futures series is the artificial price jumps on contract roll dates (these jumps do not usually create trading profits/losses as the contracts are rolled at a minimal cost).

The 'adjusted' series attempts to address this issue and create a consistent series free from price jumps on roll dates.
The adjustment is done by taking the absolute price differential between the final closing price of the expiring contract and the opening price of the new front-month contract, and then adjusting the entire series prior to the current front-month contract by this differential.
Note that the absolute price differential is used and not a ratio.

This 'adjusted' series will have no jumps on roll dates however, there are several issues to be aware of. As the adjustment is made to the entire prior series which has already been adjusted multiple times, the adjustments are cumulative and so the prior price data diverges from the actual price data. In some cases, the prior data very far back can even become negative.
In spite of this, the price charges are still almost identical to the unadjusted data, and so the adjusted data is usually preferable for backtesting and developing trading strategies.