Short-term update: discussion in Bloomberg and Barron's.
One should not have their portfolio linger for the next ~4% single day up-move in the market, similar to what we witnessed Wednesday August 26, shortly after the black swan smash. Many think that the sizable market swings seen in the past month in the U.S. (and the past summer in Asia) involve equal magnitude swings in both directions. This flawed mastery was reinforced by the immediate ~4% one-day jump in August - slightly less than the single day up-move of 4% we last saw in 2011. However despite the general up-trend of markets over lengthy periods, as we'll see below, extreme up-days (e.g., not 0.5% or even 1.5% but >2% per here) are generally far more rare than equal magnitude swings to the downside. Be careful recklessly dawdling, as you constantly hope to ultimately see many such clustered days to recover lost ground.
What we observe over the past 9 years (2015 data annualized) are five things. Note that all data above use continuous return calculations to focus on return symmetry, per the fifth item below. The first thing is that there have been 100 trading days with a >2% move. That's great! And it accounts for only ~4% of the trading days. On the flip side, there have been 129 trading days with a <-2% move (colors in the chart are slighted softened for easier identification). And this accounts for nearly 6% of the trading days (so extreme up-days come around only ¾ as often as extreme down-days). Incidentally, as we've achingly seen so far this year these extreme up and extreme down days do not always have to be interlaced with -or adjacent to- one another. In other words, when you see such an extreme day, don't overly expound what it might mean for seeing another one immediately again (in the same direction).
The second thing is we can decompose these values in terms of scoring each annually:
One should not have their portfolio linger for the next ~4% single day up-move in the market, similar to what we witnessed Wednesday August 26, shortly after the black swan smash. Many think that the sizable market swings seen in the past month in the U.S. (and the past summer in Asia) involve equal magnitude swings in both directions. This flawed mastery was reinforced by the immediate ~4% one-day jump in August - slightly less than the single day up-move of 4% we last saw in 2011. However despite the general up-trend of markets over lengthy periods, as we'll see below, extreme up-days (e.g., not 0.5% or even 1.5% but >2% per here) are generally far more rare than equal magnitude swings to the downside. Be careful recklessly dawdling, as you constantly hope to ultimately see many such clustered days to recover lost ground.
What we observe over the past 9 years (2015 data annualized) are five things. Note that all data above use continuous return calculations to focus on return symmetry, per the fifth item below. The first thing is that there have been 100 trading days with a >2% move. That's great! And it accounts for only ~4% of the trading days. On the flip side, there have been 129 trading days with a <-2% move (colors in the chart are slighted softened for easier identification). And this accounts for nearly 6% of the trading days (so extreme up-days come around only ¾ as often as extreme down-days). Incidentally, as we've achingly seen so far this year these extreme up and extreme down days do not always have to be interlaced with -or adjacent to- one another. In other words, when you see such an extreme day, don't overly expound what it might mean for seeing another one immediately again (in the same direction).
The second thing is we can decompose these values in terms of scoring each annually:
- 6 years where number of extreme down-days > number of extreme up-days
- 2 years where number of extreme down-days = number of extreme up-days
- 1 year where number of extreme down-days < number of extreme up-days
The third thing is we can note that this 9-year period covers a diverse range of economic events: from the global financial crisis (GFC), to my years in TARP, to globally-synchronized central banks' monetary easing, to record economic output, through this year's summer market calamity thusfar. Think about that, even with the gentle ride upward that we've been accustomed to, through many years of near zero interest rates and central bank asset purchases, we've only had 7 extreme up-days in the 3 years covering 2012, 2013, and 2014 (<1% of the days!) This has been a fanciful period well documented (here, here, here, here), and includes prolonged periods of relentlessly premature warnings from Nobel-laureate Professor Shiller who helped provide data for a previous article.
The fourth thing is that so far in 2015, we see volatility on the rise versus the prior couple years and are assuredly in the long tail of risk in both directions. But as is unquestionable from the chart above, we are still looking better (in terms of frequency and severity) versus the entire GFC-era years. For more on contrasting risk measures with that period, see "market crash statistics" and article.
The fifth thing is for one to make up for a large extreme down-day, one needs an even larger magnitude extreme up-day to make up for it. This is why we showed the values in the chart using continuous return rates. For example, if you had $100,000 in the S&P 500 before China's "Black Monday" (the multitrillion dollar wealth evisceration on August 24 impacting everyone from individual speculators, to index-linked ETF holders, to many famous hedge funds), then after that day's -3.9% (non-continuous) move your portfolio would be worth $96,059. A +3.9% (non-continuous) rebound from this crash would then only equal $99,845. You instead need an even-greater +4.1% (non-continuous) rebound to properly return to the $100,000. The data in the chart above makes for easier comparisons by converting all returns to a continuous format, where the black swan event was a -4.0% move, requiring a similar +4.0% one-day rebound (which we have not seen since 2011). This also suggests that the one-day extreme rebounds are slightly more rare versus what is historically shown.
The summary here is that if you are standing around for a streak of acute extreme up-day moves (a.k.a. "rips", "relief rallies", or "short-covering"), then you are apparently going to be waiting a long time. Don't play crazy games. If it comes at all (and certainly it may: again anyone who acts as if they surely know is just reading tarot cards), then you will -by the nature of the many more extreme down-days- have lost much more capital in the interim.
On an aside: our articles have recently enjoyed far-reaching appreciation, from both outstanding friends and newcomers. Here's a sample. "Market crash statistics" was shared by Zero Hedge. "Transfiguring 985 million into 1 billion" was shared many hundreds of times including by a producer at NPR, by securities and regulatory professionals, CEOs of both Futures Company and KDnuggets, Bloomberg, solidly #2 on redditt, and translated for different news globally.
"The value-at-risk fiasco" was also shared hundreds of times, including by numerous industry titans, and professors, Bloomberg, Top News in Zero Hedge, and Quantocracy. "Year of the slaughtered black sheep" was shared by Abnormal Returns. "Waiting to be right" was shared by Bloomberg.
A valuable company for which I am on external council was recently acquired by investment force BlackRock, and later this year will have a nice analytical finance article, published by the CFA. Lastly, the world's largest statistical software company has enjoyed this blog and offered to generously assist in a future article. The adventure rocks on...
A valuable company for which I am on external council was recently acquired by investment force BlackRock, and later this year will have a nice analytical finance article, published by the CFA. Lastly, the world's largest statistical software company has enjoyed this blog and offered to generously assist in a future article. The adventure rocks on...
No comments:
Post a Comment