Estimating Daily S&P 500 Dividends from Price and Total Returns

Summary

The document asks how to obtain daily or weekly S&P 500 dividend data for studying short-term changes, including during COVID-19, when dividend records are commonly available at lower frequencies. One proposed approach is to compare simple price returns with gross total returns, whose difference can indicate the dividend contribution over a period. The document does not provide a derivation, data source, or implementation details for turning that comparison into a daily series.

A second answer sketches an alternative estimate: infer a geometric growth rate from quarterly dividends, convert that rate to a daily rate, and project a stream forward from a starting dividend. This is a modeling assumption rather than observed daily payment data. The response itself cautions that dividends are not paid every day, so such a constructed series would represent an assumed accrual or estimate, not the actual timing of distributions. The discussion does not compare the approaches or validate either against historical observations.

Key ideas

• The difference between price returns and gross total returns can help estimate dividend contribution.
• A geometric growth rate inferred from quarterly dividends can be converted into a daily assumed rate.
• A projected daily dividend stream is an estimate and does not represent actual daily payments.
• The document gives no validation, data source, or detailed method for constructing the daily series.

Tags

# How to find/calculate daily S&P500 dividends?


# How to find/calculate daily S&P500 dividends?












I need a time series with daily dividend data on the S&P500. I understand most dividend data is on a quarterly basis but I'm looking to model short-term effects of COVID-19 on dividends so I need a daily/weekly frequency. Does anyone have any ideas?

Is is possible to use the S&P500 price and total return indexes to calculate daily dividend? And if so, how do you do this? I have tried something like this but then I get stuck with a static 252-trading day moving average.

Note; interpolating quarterly data to a daily frequency is sadly not an option.

## Answer by user28909 (score 1)

https://quant.stackexchange.com/a/54368

I would use the difference between simple return (price change) and total return gross of dividend withholding tax.

## Answer by JazKaz (score 0)

https://quant.stackexchange.com/a/54338

Are you trying to estimate what would happen if there were daily dividends?

If so you could try:

- Find the geometric $g$ rate of all quarterly $D$

- Convert the quarterly $g$ rate into daily,

- Find starting day $D_{0}$ and from there calculate $D(1+g)$... till you’ve found all wanted daily dividend streams

This would be how I’d attempt to find it. This is weird tho... dividends aren’t always paid, especially not daily.

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.