As he was cycling yesterday, it occurred to Macro Man that he had left out a fairly obvious component in his analysis of key day reversals and the S&P 500- namely, how they work to the upside. Wednesday's rally in the face of an abject GDP revision suggested that his caution about Tuesday's reversal was warranted....but what would the data say about bullish reversals?
Obviously, some tweaks to the study were required, as he's fairly certain that he won't find many key day reversals off of the all-time low in his sample. In lieu of the "all time high" filter in his bearish reversal analysis, Macro Man substituted a 6-month low to capture a reversal off of a defined trend. Otherwise, the definitions were simply (ahem) reversed from yesterday's study (i.e., a reversal was defined as prices making a lower low than the previous day's, only to close above the previous day's high.)
On the face of it, bullish key day reversals appear to work a bit better than bearish ones. As the table below suggests, in the 225 observations since 1982 stocks fare marginally worse than average over the week following a bull key day reversal, but comfortably better than average on 20- and 60-day horizons. On the rare occasion when you get a key day reversal off a 6 month low, stocks crush it. Perhaps the Fed should order key day reversals the next time they wish to implement unconventional monetary policy!
Of course, with only five observations, the results of the pattern off of 6 month lows is by definition fluky. In fact, there are so few observations that we can examine them one-by-one:
June 1984: Frankly, this is long enough ago that even Macro Man cannot recall exactly what was going on...after all, he had just finished the seventh grade. Although the reversal pattern was good for a short-term kick, the SPX subsequently made a fresh low a month later before subsequently exploding higher a few days thereafter. You'd have been hard pressed to feel comfortable holding a long position then, and you'd be hard-pressed to explain that rally on the key day reversal now.
July 2002: Stocks suffered badly in the summer of 2002 as the corporate accounting scandals that took down Enron, Tyco, Adelphia and Arthur Andersen claimed their biggest scalp in Worldcom. The latter declared bankruptcy on July 19, sending stocks spiralling lower until they put in an impressive reversal on the 24th. The reversal worked well for a month or so...until it didn't. Still, you'd have to say that it nailed a tradeable bounce.
October 2002: Corporate credit got crushed in the late summer/early autumn of 2002, taking stocks with it after the optimism fostered by the previous reversal. Sharp-eyed readers may have noted a subsequent October reversal on the previous chart, which worked just as well as the previous one. Indeed, even better, as the gains held up for longer...though stocks did subsequently swoon late in the winter in the run up to the invasion of Iraq in March 2003. (Sobering thought: junior analysts during the Iraq invasion are now MD's in their 30's.)
January 2008: Ah, Jerome! Markets dumped in January 2008 for no obvious reason, forcing both the Fed and (put your headphones on) some punters to freak out. The Fed delivered an intra-meeting cut on January 22nd, and when it emerged immediately after that rogue trader Jerome Kerviel was behind the selling, markets regained their footing rather quickly. Regained it, that is, until they had a look at Bear Stearns' liquidity position, at which point we were at fresh lows less than two months later.
November 2008: Ummm, nope. (Twentysomething holders of limit long positions in assorted low-quality risky assets are invited to consult the scale on the right hand side of the chart.)
So there you have it. Bullish key day reversals off a six month low are normally good for at least a short-term bounce....unless they aren't. The track record in nailing significant trend reversals is dubious at best, however. This is a prime example of why Macro Man uses technicals as a piece of the puzzle, but almost never as the whole puzzle unto themselves.
Obviously, some tweaks to the study were required, as he's fairly certain that he won't find many key day reversals off of the all-time low in his sample. In lieu of the "all time high" filter in his bearish reversal analysis, Macro Man substituted a 6-month low to capture a reversal off of a defined trend. Otherwise, the definitions were simply (ahem) reversed from yesterday's study (i.e., a reversal was defined as prices making a lower low than the previous day's, only to close above the previous day's high.)
On the face of it, bullish key day reversals appear to work a bit better than bearish ones. As the table below suggests, in the 225 observations since 1982 stocks fare marginally worse than average over the week following a bull key day reversal, but comfortably better than average on 20- and 60-day horizons. On the rare occasion when you get a key day reversal off a 6 month low, stocks crush it. Perhaps the Fed should order key day reversals the next time they wish to implement unconventional monetary policy!
Of course, with only five observations, the results of the pattern off of 6 month lows is by definition fluky. In fact, there are so few observations that we can examine them one-by-one:
June 1984: Frankly, this is long enough ago that even Macro Man cannot recall exactly what was going on...after all, he had just finished the seventh grade. Although the reversal pattern was good for a short-term kick, the SPX subsequently made a fresh low a month later before subsequently exploding higher a few days thereafter. You'd have been hard pressed to feel comfortable holding a long position then, and you'd be hard-pressed to explain that rally on the key day reversal now.
July 2002: Stocks suffered badly in the summer of 2002 as the corporate accounting scandals that took down Enron, Tyco, Adelphia and Arthur Andersen claimed their biggest scalp in Worldcom. The latter declared bankruptcy on July 19, sending stocks spiralling lower until they put in an impressive reversal on the 24th. The reversal worked well for a month or so...until it didn't. Still, you'd have to say that it nailed a tradeable bounce.
October 2002: Corporate credit got crushed in the late summer/early autumn of 2002, taking stocks with it after the optimism fostered by the previous reversal. Sharp-eyed readers may have noted a subsequent October reversal on the previous chart, which worked just as well as the previous one. Indeed, even better, as the gains held up for longer...though stocks did subsequently swoon late in the winter in the run up to the invasion of Iraq in March 2003. (Sobering thought: junior analysts during the Iraq invasion are now MD's in their 30's.)
January 2008: Ah, Jerome! Markets dumped in January 2008 for no obvious reason, forcing both the Fed and (put your headphones on) some punters to freak out. The Fed delivered an intra-meeting cut on January 22nd, and when it emerged immediately after that rogue trader Jerome Kerviel was behind the selling, markets regained their footing rather quickly. Regained it, that is, until they had a look at Bear Stearns' liquidity position, at which point we were at fresh lows less than two months later.
November 2008: Ummm, nope. (Twentysomething holders of limit long positions in assorted low-quality risky assets are invited to consult the scale on the right hand side of the chart.)
So there you have it. Bullish key day reversals off a six month low are normally good for at least a short-term bounce....unless they aren't. The track record in nailing significant trend reversals is dubious at best, however. This is a prime example of why Macro Man uses technicals as a piece of the puzzle, but almost never as the whole puzzle unto themselves.
9 comments
Click here for commentsC Says
Reply"uses technicals as a piece of the puzzle, but almost never as the whole puzzle unto themselves"
Indeed. Moreover big moves typically typically come off elongated sideways periods of action so personally I look at months ,or weeks ,but hardly ever days action. On that note the USD/YEN is in the zone to deliver some meaningful move after going sideways all year.
Which leaves us asking why bullish and bearish should work differently. I have a sneaking feeling it's related to the same psychology as carry trades where up moves are grinds and down moves are sharp, with the stop loss bias being to the sell stop in a market that is naturally long. In these cases of these saw tooth markets key day reversals are tied in with the psychology of falling knives rather than countering a grind.
ReplySo I suspect, though would be interested to know empirically, if the bias is proportional to the extent of carry differentials. In stocks its the dividend yield and in FX it will be the interest rate diff. The bias always being referenced to the higher yielder being the base re high or low as of course with all these things we can just 1/x the price ratio to make the observation invalid (EUR/GBP or GBP/EUR).
Pol
That July 02 post made me all nostalgic for when companies who ran out of money used to default and punters engaged in a thing called "risk aversion"
ReplyFor equities, the difference between bullish and bearish is that there is an equity risk premium which acts as a tailwind for higher returns. Generally speaking, the risk premium rises as stocks fall (and weaker hands no longer want to buy / stop out of longs). At a significant low, you could easily see risk premium adding a higher than normal tailwind higher. On the bear side, you are betting that things get meaningfully worse or that risk aversion rises... Both of which are significantly harder to capture than an abnormally high risk premium.
ReplyStocks red... oh the humanity!
ReplyBullard mentions rates aren't priced in, spooz immediately fall 30pts lol. Equities and property are now juggling for pole position as the winner of the "biggest ponzi scheme" contest.
ReplyMM, nice mention in Grants!
ReplySince macro man is going all TA on us, check out this site. Pretty 1990's style but has some interesting info on a plethora of TA patterns, though on individual stocks mostly
http://thepatternsite.com/OutsideDays.html
Add lower quality US credit (junk, PIK and leveraged loans) to the list of Ponzi assets. That market at the moment has a whole lot of noob participants who have never participated in a downturn and hence haven't experienced the joys of vanishing liquidity in thin markets.... "NO BID". Got popcorn? It should be fun. No shorting of equities until spreads gap wider.
ReplyKeep updating the more news and updates on stock market, as this news and updates are really very beneficial to all of us.
ReplyStock Tips