Calculating Dividend Yield in the Goyal and Welch Data

Summary

The document clarifies the dividend-yield convention used in the Goyal and Welch stock return predictability data. The dataset’s dividend yield is computed as aggregate dividends paid by all stocks divided by total stock market capitalization measured at the beginning of the year. The cited explanation distinguishes this measure from the dividend-price ratio, which uses market capitalization at the end of the year.

This timing distinction matters when reproducing the paper’s predictor series or comparing dividend-based return predictors: the numerator is the same aggregate dividend amount, while the denominator refers to a different point in time. The document provides a definition rather than a worked calculation, data-cleaning procedure, or empirical evaluation of predictive power. It also does not elaborate on dividend timing conventions or adjustments in the underlying dataset, so replication may require consulting the dataset documentation and paper for details beyond this concise answer.

Key ideas

• The dividend yield uses aggregate dividends paid by all stocks in the numerator.
• Its denominator is total market capitalization at the beginning of the year.
• The dividend-price ratio uses end-of-year market capitalization instead.
• The timing of the denominator matters when reproducing dividend-based predictor data.

Tags

# Dividend Yield Goyal and Welch (2008)


# Dividend Yield Goyal and Welch (2008)












Following the Goyal and Welch (2008) stock return predictability data, does anyone know how they calculate the dividend yield from the dataset that they provide on Amit Goyals website http://www.hec.unil.ch/agoyal/?

## Answer by Eli (score 1, accepted)

https://quant.stackexchange.com/a/38434

They define the dividend yield calculation in the beginning of the paper:

Dividend ratios are the total dividends paid by all stocks (D(t)), divided by the total stock market capitalization, either at the beginning of the year (the dividend yield, P(t − 1) or at the end of the year (the dividend-price ratio, P(t)).

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.