One last post to wrap up this discussion about short-term (daily) stock market mean-reversion becoming stronger (read part 1, part 2, and part 3). This has been a pretty geeky conversation so far, and I’m afraid that I haven’t answered the most important question…so what?
Well…if you’re a trend-follower or buy & pray type, none of it matters. More importantly, if you’re unable to minimize your transaction costs, none of it matters.
But if your trading resembles anything like ours (swing trading near-daily with low transaction cost vehicles**), all of this definitely matters.
Below are two examples of short-term strategies that have been positively impacted by the acceleration in short-term mean reversion.
Adaptive Daily Follow-Through
This is the same short-term strategy I talked about in The Simple Made Powerful with Adaptation and that I track daily on the free State of the Market report.
Strategy performance frictionless (red) versus S&P 500 (blue) from 1955:
[logarithmically-scaled]
Note the strength of the strategy’s performance today (right end of graph), not seen since the 60’s and 70’s. Enough said.
For all intents and purposes, these results could be duplicated in today’s market (net an annual fund expense ratio) using leveraged mutual funds…see note at end of post.
Scotty Strategy
Long-time readers know that I never release backtested performance for my proprietary programs because it fundamentally goes against my belief that investors should base their investment decisions on actual results (read why), but I’m bending my rule a bit to prove a point.
This is a backtest of my newest strategy, Scotty, from 2000 including all transaction costs (but excluding our 2.5% annual account management fee):
[logarithmically-scaled]
Note the uptick in performance in the last year. Scotty isn’t purely a short-term strategy, but the short-term component of the strategy is really taking off, directly as a result of the increase in daily mean-reversion. Enough said.
The point of these two examples is to say that, at this moment in time (subject to change with a portfolio-crushing lack of notice) short-term mean-reversion is the stock market play du jour. Not respecting this shift in the markets and following the CNBC’esque view of the world (the market rallied today, the bottom is here!) is quite possibly the easiest way to underperform even the sad saps on Wall Street.
[Edit: click for a summary of all posts in this series]
Happy Trading,
ms
** Side note – because I’m asked after every post where I talk about low/no transaction costs vehicles, I’ll answer in advance – I only trade leveraged mutual funds from Rydex, ProFunds, and Direxion (not to be confused with leveraged ETFs), that are designed to be actively-traded. These do not carry per-transaction costs and exhibit high r-square to their underlying indices (meaning, unlike ETFs, they track their indices very tightly). The downside of these of course is that they do not allow for intraday orders (stop, limits, etc.)
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Evolving Markets, Follow-Through, Stock Market Mechanics | 7 Comments
Michael,
I have enjoyed the series on the increasing strength of reversion to the mean, and thanks for the help with that spreadsheet a few weeks ago. Hope you don’t mind, I reference your work with a link on my blog. Have a great week.
Michael,
I don’t understand how you implement a strategy like daily follow-through using leveraged mutual funds. These funds require you to send trade orders significantly in advance of the close – but the strategy seemingly requires you to trade at the close based on the price at that exact time. What am I missing?
Thanks
RE to Josh: Good question. These funds do not require you to send orders “significantly in advance of the close”. The exact fund cutoff time depends on the fund company and specific fund, but is usually 5-10 minutes before the market close. In some uncommon instances, this would lead to a trader taking the wrong position (because the market shifts directions in the last 5-10 minutes), but those instances should be few and far between. Hope that helps.
michael
Hi Michael,
How does revert to mean come into question for say the Japanese stock market, it is down 80% from the last peak, about 18 year back.
http://finance.yahoo.com/echarts?s=%5EN225#chart1:symbol=^n225;range=my;charttype=line;crosshair=on;ohlcvalues=0;logscale=on;source=undefined
Amit
Amit – couple of thoughts. First, I rarely do any work with non-US markets and all of the observations in this series were specifically about the S&P 500. Second, short-term mean-reversion and the long-term trend are two entirely different things. I often talk about the idea that (at this moment in time) the market tends to trend over long time horizons (ex. 50/200-day crossovers), but is very much contrarian over short time horizons (ex. adaptive next-day follow-through).
The point of that second thought is to say that even if Japan exhibits strong short-term mean-reversion (I have no idea if it does), that does not have any impact on the broader long-term trend.
Just my $0.02.
michael
Fair enough.
It is also likely that short term trends might continue, particularly with an adaptive strategy even if some of the Economic factors are up in the air.
Amit