Key insights
- Analysis of 151k+ dividend ex-date events reveals recovery patterns across security types and yield levels. BDCs show the slowest recovery. High yield stocks take longer to recover. Extreme price deviations at ex-date (high z-score) suggest stronger mean reversion. This data-driven approach offers insights for dividend investors, but its impact on broader US equities is limited.
Been building a dividend intelligence tool and ended up with a database of 151,422 ex-date events across CEFs, ETFs, REITs, BDCs, and dividend stocks going back 17 years. Figured the data was worth sharing since most discussion around ex-date dips is based on gut feel.
Recovery by security type:
Dividend Stocks: 6.7 days average recovery, 71.5% recover within 5 trading days REITs: 7.7 days average recovery, 66.3% recover within 5 days ETFs: 8.1 days average recovery, 62.2% recover within 5 days CEFs: 8.9 days average recovery, 56.2% recover within 5 days BDCs: 12.4 days average recovery, only 45.1% recover within 5 days
Overall median across all 151,422 events: 3 days
The yield effect is real:
Under 3% yield: 6.9 days average recovery 3 to 5%: 7.0 days 5 to 8%: 7.9 days 8 to 12%: 8.3 days Over 12%: 10.1 days
Higher yield means a bigger hole to climb out of. That is consistent across 17 years of data.
The BDC finding surprised me most:
BDCs have the largest average drop at 2.08% AND the slowest recovery at 12.4 days. Only 45% recover within 5 trading days. If you are buying BDC dips expecting a quick bounce the historical data says be patient.
Individual variance is where it gets interesting:
Stocks recover fastest on average but individual variance within each category is massive. Among CEFs with 20 or more cycles in the dataset the fastest recovering funds average under 5 days while the slowest take 3 or more weeks. Both show up as CEFs on any screener. The historical pattern data separates them.
The z-score frame:
A security trading 2.5 or more standard deviations below its 252 day mean at ex-date is a statistically unusual event, not a routine dip. Those setups show stronger mean reversion tendencies than ex-dates occurring near the historical price average.
Happy to answer questions about methodology or specific tickers.