It seemed like business as usual yesterday morning, with the SPX marching to yet another all-time high on news that the day of the week ended in "y" but did not start with "s". However, just as the referee decided that Claudio Marchisio was worthy of sending off but not Luis Suarez, Spooz decided to visit the dark side, taking a rather nasty tumble to close below Monday's low print- but not, of course, falling 1% or more, as price action of that magnitude is now banned by administrative diktat. Still, it was a key day reversal from all-time highs....perhaps Italians were the only ones trading?
There are a number of interesting talking points about yesterday's price action. The legend of "Super Tuesday" is once again dying a slow death, much as it did last year, as the outperformance of the second trading day of the week has faded markedly since the beginning of last month. Perhaps "sell in May" only applies to Tuesdays?
However, the key day reversal is the thing that really grabs the eye. After a long, steady, low-volume/low-volatility grind higher, such a reversal certainly rings more than a few alarm bells. It is sort of taken as a given that this is a trend reversal signal, if only temporarily, which would certainly fit with both the skepticism over the current rally and concerns from sources as diverse as the squeeze in energy prices and the eventual end of the taper, etc.
Macro Man decided to put the data to the test. He identified all past key day reversals in his dataset and checked the subsequent returns on a 1, 5, 20, and 60-day horizon. He performed the same analysis on a sub-set of reversals on days when the index made an all-time high. For a control group, he calculated the same returns for all data points in his sample. The results are set out in the table below.
As you can see, while generic key day reversals produce marginally worse than average returns on the day after the reversal, on horizons of a week and longer the returns are substantially better than average. The negative short-term impulse from a key day reversal off of the all-time high would appear to be slightly stronger, though with just 23 observations it could easily just be noise in the data. On longer horizons, the subsequent returns are even stronger than those following generic key day reversals. The moral of the story, evidently, is that markets at their all-time high are bull markets, and it takes more than a little technical reversal to derail a secular bull.
Unfortunately, the Bloomberg dataset only includes daily high/low data since April 1982, a date which coincided with the starts of a rip-snorting 18 year secular bull market. True, the sample also includes the small matter of the GFC, but a look at the price of...err...anything suggests that that is now merely a distant speck in the market's rear-view mirror. It would be interesting to perform the same analysis on a much longer data set, perhaps using the Dow rather than the SPX, which only came into existence in 1957. If anyone has access to such data that they're willing to share, by all means pass it along and Macro Man will do the study.
On the basis of the available evidence, however, the key day reversal looks to be more of a curiosity than a legitimate threat to the market's trend. Of course, you cannot control for other factors which might impact the market...but then again there's almost always a reason to worry about one thing or another. While your author cannot, of course, provide investment advice, he can observe that of all the reasons out there to sell equities, yesterday's key day reversal would appear to merit a spot somewhere close to the bottom of the list.
There are a number of interesting talking points about yesterday's price action. The legend of "Super Tuesday" is once again dying a slow death, much as it did last year, as the outperformance of the second trading day of the week has faded markedly since the beginning of last month. Perhaps "sell in May" only applies to Tuesdays?
However, the key day reversal is the thing that really grabs the eye. After a long, steady, low-volume/low-volatility grind higher, such a reversal certainly rings more than a few alarm bells. It is sort of taken as a given that this is a trend reversal signal, if only temporarily, which would certainly fit with both the skepticism over the current rally and concerns from sources as diverse as the squeeze in energy prices and the eventual end of the taper, etc.
Macro Man decided to put the data to the test. He identified all past key day reversals in his dataset and checked the subsequent returns on a 1, 5, 20, and 60-day horizon. He performed the same analysis on a sub-set of reversals on days when the index made an all-time high. For a control group, he calculated the same returns for all data points in his sample. The results are set out in the table below.
As you can see, while generic key day reversals produce marginally worse than average returns on the day after the reversal, on horizons of a week and longer the returns are substantially better than average. The negative short-term impulse from a key day reversal off of the all-time high would appear to be slightly stronger, though with just 23 observations it could easily just be noise in the data. On longer horizons, the subsequent returns are even stronger than those following generic key day reversals. The moral of the story, evidently, is that markets at their all-time high are bull markets, and it takes more than a little technical reversal to derail a secular bull.
Unfortunately, the Bloomberg dataset only includes daily high/low data since April 1982, a date which coincided with the starts of a rip-snorting 18 year secular bull market. True, the sample also includes the small matter of the GFC, but a look at the price of...err...anything suggests that that is now merely a distant speck in the market's rear-view mirror. It would be interesting to perform the same analysis on a much longer data set, perhaps using the Dow rather than the SPX, which only came into existence in 1957. If anyone has access to such data that they're willing to share, by all means pass it along and Macro Man will do the study.
On the basis of the available evidence, however, the key day reversal looks to be more of a curiosity than a legitimate threat to the market's trend. Of course, you cannot control for other factors which might impact the market...but then again there's almost always a reason to worry about one thing or another. While your author cannot, of course, provide investment advice, he can observe that of all the reasons out there to sell equities, yesterday's key day reversal would appear to merit a spot somewhere close to the bottom of the list.
7 comments
Click here for commentsinteresting analysis, pay attention to sampling error when you move from 268 observations to 23. gut feeling is that if you run a T-test on the means, you may not be able to reject the null hypothesis.
ReplyOn a broader scale, these results remind me of the "bottom fishing strategy" described by Tushar Chande on SPX in late 90s
(buy on a channel breakout sell day)
As long as the major trend is up, even the sell signals are good buys...
Italia is out, now Viva Colombia!
ciao f
Check out the skew/vix spike last Friday - I believe it has a better track record in terms of signalling tops
Replyif you are looking for a longer data set try Yahoo finance
ReplyEnergy stocks finally came off ;-)
Key reversal days do have some power but that doesnt mean they are infallible. Trading rules must be built around them
But nice work MM, I have sold many reversal days before, too bad I hadnt read this a few years ago ;-)
That was a great post. Thanks, MM!
ReplyI don't think the buyers are price sensitive at all - giant asset allocations into equities (like the Norway SWF increase announced recently) dont care about reversal days or bad news. It'll keep going until they stop buying. If you had to allocate a $trillion into global equities, it would take you a while.
Replybecause sovereign funds and all are so great at timing markets aren't they
ReplyShrugs, JBTFD!
Reply