Do calendar effects found in stock markets also appear in broad stock index futures? In their November 2011 paper entitled “Calendar Anomalies in Stock Index Futures”, Oscar Carchano and Angel Pardo investigate 188 possible cyclical anomalies in S&P 500, DAX and Nikkei index futures contracts (derived from day-of-the-week, month-of-the-year, weekday-of-the-month, week-of-the-month, semi-month, turn-of-the-month, end-of-year, holidays, semi-month-of-the-year, week-of-the-month-of-the-year, Friday the 13th, Halloween effect and quarterly futures expiration). They note that small trading frictions and ease of shorting promote exploitability of anomalies in futures markets. They assume round trip trading frictions of 0.05% for assessing net profitability. Applying tests not dependent on type of return distribution to stock index futures prices from December 1991 through April 2008, they find that:
- Only three of 188 effects exhibit statistical significance in both of two equal sample subperiods: (1) long the turn-of-the-month for S&P 500 Index futures; (2) short the fourth-week-of-September for S&P 500 Index futures; and, (3) short the fourth-week-of-July for Nikkei futures.
- The turn-of-the-month “+1” effect (buying at the close on the last trading day of the month and selling at the close the next trading day) applied to S&P 500 Index futures is the only calendar effect that persistently exhibits net profitability over the entire sample period. Specifically:
- The net cumulative profit for this strategy over the sample period is 27.5%.
- Profit concentrates in next-day rather than overnight returns.
In summary, evidence indicates that a compressed turn-of-the-month effect is the sole likely calendar effect for broad stock index futures.
Cautions regarding findings include:
- As noted by the authors, the sample is very short (16 observations, and only eight for subperiods) for annual effects.
- The half-sample persistence test may not fully account for data snooping bias (lucky findings) associated with testing 188 potential anomalies on three price series.
- Some traders may not be able to achieve the assumed level of trading frictions.
