We tested the claim "A payout ratio above 100% predicts a cut"

What we found

Payout over 100%: 20.2% cut their dividend within 12 months, across 2,131 observations. Payout 60% or less: 8.2%, across 9,862. The gap is 12.0 percentage points, with a 90% range of 9.1 to 13.7.

How the test worked

• Payout over 100%. Every company in the universe, where payout ratio (dividends as a share of earnings) > 1 on each date.
• Payout 60% or less. Every company in the universe, where payout ratio (dividends as a share of earnings) <= 0.6 on each date.
• Universe: US-listed companies with a market value of at least $2bn, observed every six months from 2015 to 2024.
• A cut means the dividends paid over the next 12 months fell more than 2% below the 12 months before, with one-off special dividends excluded.
• The test was registered as number 3 before it ran, with the effect it had to show set in advance.

Data quality

• Observations with no cut label (forward window past the data, or no trailing dividend): 23.
• Rows where two independent yield measures disagreed by more than 2x: 1,127.
• Rows above the 25% yield ceiling: 56.
• Dividend payers whose yield could not be cross-checked: 8.
• Payout over 100%: 121 observations had no share price a year later (taken over or delisted). A takeover ends the dividend, so these can count as cuts.
• Payout 60% or less: 321 observations had no share price a year later (taken over or delisted). A takeover ends the dividend, so these can count as cuts.

Limits

• Market value above $2bn approximates, but is not, FTSE index membership.
• Companies that delisted have partial price histories, so the survivor correction is incomplete.
• This tests what happened to dividends, not what an investor would have earned.

Data

Download the per-period data (CSV). Dataset 2b6b505eef56, gate rules version 1.