- The "Tuesday Effect" has been powerful in 2014 -- until last week.
- There are good reasons for skepticism.
- The lesson can be applied to other alleged seasonal patterns.
How should traders and investors take advantage of seasonal patterns?
There are so many opportunities for dividing the calendar! This year, the best has been to go long on Tuesday. Here is the 2014 summary from the Bespoke Team:
You could have done even better if you had sold short on the other days of the week.
You might well be skeptical about this easy system. When I offered it for comment to the Scutify community they suggested that there were not enough cases and that it did not make sense. Market veteran Robert Marcin suggested that this pattern would quickly be learned by the market. Monday buying would anticipate the Tuesday effect, and it would quickly disappear.
In fact, prior years have had different days as the best performers.
So far, so good. There have been eighteen Tuesdays to establish and test a pattern.
- There was no a priori hypothesis
- There is no "out of sample" data that could be used for a test
- The results are "data mining" covering a specific time period.
Those who are skeptical are fully justified. The pattern broke down last week.
Let us take the conclusions and find a new setting. How about the Presidential Election cycle. This is getting a lot of buzz as "experts" opine about why the second year of the second term should be bad. How many cases are there? If you wanted to get 18 (the same number for this year's Tuesdays), you would need to go back 144 years. Would that really be relevant?
We have better data on the "Tuesday effect" than we have on the oft-touted Presidential Election Cycle.
Most of those discussing seasonality violate all three of the points listed above.
Challenge for the Seasonality Crowd
Why hasn't the market "learned" this behavior and adjusted?
Did any of those parroting "sell in May" advise you to "buy in October" last year?