Musings, and more musings before the meat
This is a long post which has been divided into two. It has really taken lots of introspection, as it crystallized when being alone in the countryside enjoying the sights of some wonderful lush hills dotted with grazing cattle. This post is a roadmap for the future tasks to be undertaken which I feel will expand the potential of the Dow Theory.
Yes, I am convinced that the Dow Theory has still lots of untapped potential. I alluded to this in past posts, but it bears repeating and expanding in this new post. Hamilton’s insight that the Dow Theory may be successfully applied to any market (not just US stocks) was tentatively proven right here. This is vital, as Hamilton’s insight opens new avenues for the application of the Dow Theory to more markets.
I also wrote that (after thorough back-testing to be done, I might add) the Dow Theory could be the basis for a successful futures hedge fund. This is not an overstatement. We already know that, at least for US stocks, the Dow Theory tracks much better the trend than breakout systems (even if confirmation is demanded), as was explained (for US stocks) here.
As far as I know no single fund is currently doing this (if some reader knows otherwise, please let me know). Hence, it is not outlandish to think that it is likely that the Dow Theory could be a better way to trading commodities than the typical breakout or moving average systems used by most futures funds. Probably, and due to the ample stops resulting from the application of the Dow Theory, we should adopt a shorter time frame (i.e. just requiring 3 or 4 days for an “abridged” secondary reaction to exist in order to trade futures (more on that in a future post). One preliminary test which I carried out with the oil sector (in a very raw state, and not yet polished for posting on this blog) seems to preliminary suggest that the application of the Dow Theory rules with just the exception of only demanding three days for a secondary reaction to exist, results in profits which are larger (percentage-wise) than those obtained by using the usual length for a secondary reaction. I preliminary see some degradation of the profit factor (that is the ratio between gross profit and gross loss) which is logical since as we shorten our time frame, there is more noise and less signal. However, we shouldn’t be purists. I’d rather prefer a system less “efficient” with more trades (and turnover) if I can accomplish more returns and less drawdowns (especially lengthwise, as the patience of most investors, I included, is in short supply). I could be very efficient in catching trends that last two years. However, if the average trade were just of 10%, I could end up making a paltry profit per year. On the other hand, shorter time frame trading, while less efficient (i.e. more losing trades), may result in an average trade of 3% with 5 such trades per annum. Hence, even if less “efficient” I prefer the shorter term system (unless capital constraints become an issue), as I can make percentage-wise more per annum, and more, importantly, each losing trade will tend to be smaller and the higher frequency of trades will get me out of any drawdown sooner. All these are complicated aspects, and before putting money on the line, one must really do his homework. This is what I attempt to do in the coming months or maybe years.
The takeaway is as follows: While I still have to perform many time-consuming tests (as the Dow Theory is not easily encapsulated into an algorithm), it seems that shortening the length of secondary reactions results in a system that may be able to extract decent profits (albeit less efficiently, with more losing trades), generate more trades (good things if they are profitable), cut losses short (as the Dow Theory stop based on the shortened secondary reaction is narrower), and have flattened shorter drawdowns. Such a system would be geared towards futures (commodities). More about the Dow Theory build in stop-loss here.
There are also many issues which I still want to study in the future which are important for investors:
· Deepen the principle of confirmation. What makes a good pair of indices? Should we expand Schannep’s insight of using three stock indices to other markets when possible (i.e. crude oil, heating oil and unleaded gasoline? Or we just make do with two? What level of correlation between the indices is desirable? I already alluded that this is like Goldilocks: Neither too hot, nor too cold. Since the principle of confirmation is vital to add value to our trades, we better be proficient at finding dancing partners when foraging outside the world of US stock indices.
· Provide you with exact figures as to the loss of performance when we cannot entry and exit on the very same day of the signal (at the close). Hint: not so much lost.
· Should we take all trades suggested by the Dow Theory? One tenet of trading is the “risk reward ratio” (RRR). In actual trading, we discriminate. Not all trades are born equal. If the particular setup of the signal shows that there is ample downside (i.e. the primary bear market lows which is our initial stop after we get a bull market signal) while historically the upside has been limited, should we skip the trade? I don’t know now. I am inclined not to dabble with the Dow Theory when applied to US stocks. However, what if I applied the Dow Theory with shortened secondary reactions to a portfolio of commodities? I feel, given the shorter term nature of such a system, that I would be strict in only taking the trades which would offer a good RRR (i.e. likely reward of at least two times the risk taken).
· When dealing with US stocks, should we apply the Dow Theory signals to other indices? What about buying a midcap ETF instead of the SPY when a bull market signal is given? Or a growth ETF? Or an equal weight ETF such as RSP? Could we outperform the official Dow Theory record by doing this? Or are the extraneous ETFs out of sync with the signals given by using the Industrials, Transports and S&P 500?
· And last but not least, could be apply relative momentum to the Dow Theory? The Dow Theory, as with any trend following method is “absolute momentum”. What about of using relative momentum (i.e. buying the best performing index within a given look back period) coupled with the Dow Theory’s absolute momentum? Gary Antonaccy, whom I have lauded in the past, in his seminal book “Dual Momentum Investing” advocated for using both absolute momentum (the asset class in question must be in an uptrend) and relative momentum (the specific asset must be outperforming some peers). From my many posts on this blog, I am convinced of the superiority of the Dow Theory over other trend following methods in order to appraise "absolute momentum". Hence, I feel that if we could couple the more efficient Dow Theory at signaling trends (in any market, by the way) with relative momentum (i.e. buying the strongest US stock sectors only when there is a primary bull market signal) we could further boost the already excellent track record of the Dow Theory. Of course, testing all these thoughts takes time and dedication but I feel the reward is worth, as I’m aiming specifically at:
(a) adding even more outperformance to the Dow Theory (let’s say 2 or 3 additional percentage points p.a.) versus buy and hold;
(b) with some luck (I need to see the tests results) reduce drawdowns (specially lengthwise); and
(c) very importantly, reduce the time of underperformance versus buy and hold. We already know that, while buy and hold is a long term disaster as drawdowns can kill those lacking deep pockets and/or psychological fortitude (more about it here), it can outperform the Dow Theory for long periods of time. I have also explained ad nauseam (i.e. here) that the Dow Theory outperformance is made when big declines occur. Hence, the current hyper bullish environment is not conducive to outperformance for the Dow Theory. We are faring well (as you will see in my next post, no reason to bemoan) but clearly buy and hold is outperforming (at least for now). However, long periods of underperformance (even though one remains profitable) may try the patience of the less committed investors. Hence, I want to explore whether adding relative momentum to the Dow Theory (i.e. by buying the strongest US stock sectors) would result not only in adding some percentage points but also in reducing the time the Dow Theory underperforms buy and hold.
Thus, there are many projects in the pipeline. However, Schannep’s Dow Theory remains the cornerstone upon which I am going to build all these future research. Hence, before we try to push the envelope, it is advisable to make a pause, take a deep breath, and go back to the basics and review how both the S&P 500 (buy and hold) and Schannep’s Dow Theory have fared since 1998. 22 years is quite a good stretch of time to judge. Please mind that Schannep’s Dow Theory has an actual, real money, track record. No pipe dreams. By comparing both investment styles we will be able to reassert Schannep’s Dow Theory superiority.
By the way, soon I will post another chapter of the Capitulation saga. I'm still maturing the already written post.
One Dow Theorist