Friday, December 6, 2019

Dow Theory Special Issue: Hamilton’s insight that the Dow Theory may be applied to many markets seems to be right

 Applying the Dow Theory to oil markets and towards building a Dow Theory portfolio

Hamilton was Charles Dow understudy, and writer of the book The Stock Market Barometer

As I have written here, Hamilton was of the opinion that the Dow Theory may be applied to other markets. Based on that insight, when I started this blog, I wanted to apply the Dow Theory to precious metals, interest rates, and occasionally, even to Chinese stocks.

As I wrote in my previous post, from eyeballing the charts I know that my applying the Dow Theory to the pairs SLV/GLD and SIL/GDX has outperformed buy and hold since August 2012 while dramatically reducing drawdowns (when I started the blog). In order to draw final conclusions, I have to wait, though, until the current swing for those markets makes their final lows. As readers of this blog know, in order to properly compare the Dow Theory and buy and hold performance, we need to wait until the final lows of the primary bear market are made (the further decline following the primary bear market signal). However, we don’t know the final lows until the next primary bull market signal is given. Hence, I still have to wait some months to give final figures.
Therefore, while I wait the final figures of outperformance for precious metals whose trades have been publicly made known on this blog in real time, I decided to apply the Dow Theory to the energy sector. My premise was as follows. Oil stocks should more or less move in tandem with their underlying commodity. If oil is weak, oil stocks should, more or less, be weak too. Hence, I took USO (crude oil) and XLE (oil stocks) ETFs. While maybe not the best pair to apply the Dow Theory, I did it as per the following rules:

·        At least 15 trading days on both indices in order to declare a secondary reaction.

·        At least 1/3 minimum retracement, save when it has been a very long secondary reaction (i.e. 50 days, etc.) and retracement for both indices nears 1/3.

·        At least two days pullback/rally after the secondary reaction to setup the ETFs for a primary bull or bear market signal.

·        Both the secondary reaction and the subsequent pullback/rally must exceed the minimum movement (volatility adjusted). Hence, negligible movements, even if the satisfy the time requirement, do not qualify.

·        Both for buy and hold and the Dow Theory I divided the hypothetical trading capital by two. One half for USO and the other half for XLE.

·        I calculated the results and drawdowns in two different ways: One way, by only going long and being in “cash” (out of the market) when there was a sell signal. The other way, by going short when a sell signal was given, and covering and going long when a buy signal was given.

As you can see, I applied a “classical” Dow Theory flavor. I didn’t apply Schannep’s, as I just had two indices.

First off, I’ll analyze the results by just taking long trades and being out of the market when a sell signal is given. In another post, I will discuss results (truly good, by the way) when shorted when a sell signal was given.

The test spans from 08/05/2009 (first long taken) to 08/11/2018 (last long closed). Since 08/11/2018, the Dow Theory did not signal any new long position.

Here you have the highlights:

·        Total long trades taken: six. Since the test spans ca. 10 years, we have a “typical” classical Dow Theory duration for the average trade, which is more than one year.

·        Maximum drawdown buy and hold: -64.33%

·        Maximum drawdown Dow Theory (closed trades): -17.95%

·       Final capital buy and hold (starting at 100): 90.93. Hence buy and hold lost for an allocation of 50% to each ETF ca. 9.1%

·       Final capital Dow Theory (starting at 100): 105.28, which translates to an outperformance versus buy and hold of ca. 1.5% p.a. (annualized). Please mind that the negative “gravitational” force against long positions has been very strong during the period tested, as there was a maximum decline of -64.33%. Confronted with such a bearish force, the Dow Theory managed to remain positive by going long. This is a huge accomplishment.

·     Hence, the Dow Theory reduced the maximum drawdown by 72.08% when comparing the drawdown of buy and hold to that of the Dow Theory. This dramatic drawdown reduction confirms what I have repeatedly asserted: The main source of the Dow Theory outperformance is loss avoidance. Hence, when markets (as US stock indices in the last 10 years) lack significant drawdowns, the Dow Theory will surely underperform. However, “underperform” does not entail “losing”, just winning less.

·      Therefore, Hamilton seems to be right. While I want to perform more tests to extraneous markets, it seems that the Dow Theory may be successfully applied to other markets.

Here you have the breakdown of the trades taken and other interesting data:
Trades taken by applying the Dow Theory to oil ETF (USO) and oil stocks ETF (XLE)

Starting and final capital for buy and hold and the Dow Theory
Maximum drawdown comparison between buy and hold and the Dow Theory

This exercise has been useful, as I have always thought that US stocks have behaved too nicely for the last 120 years. By “nicely” I mean that there have been few drawdowns and their depth was quite mitigated. By “nicely” I also mean that even in those rare occurrences of big drawdowns (i.e. 1932, 1974, 2000-2002, 2008-2009, and to some extent October 1987) the Dow Theory was “nice” enough to signal a bear market before the really big declines occurred. As I wrote here, there is no guarantee that the Dow Theory (or for that matter any trend following method) can get us out in time when markets become unruly, as Chinese stocks behaved in August 2015 (see here and here, if you want to see scary market action even when applying the Dow Theory). Hence, it is a good exercise to apply the Dow Theory to many markets, and especially markets that have experienced deep declines before a primary bear market signal was signalled. What if 2008 had not been preceded by a relatively soft decline at first (secondary reaction) followed by a rally (setup for primary bear market completed), and would have declined overnight by more than 30% with no prior warning? When one applies the Dow Theory to precious metals, oil markets (as with this post), wheat/corn, etc., one can experience first-hand much more unruly markets. What if one day US stocks behaved like that? Are Dow Theorists and trend followers in general prepared to weather such a storm? Personally, I don’t like to fake reality and say, “well this will probably never happen to US stocks, after all the Fed will take care of an orderly decline” or say something like this “the US is different, such drastic declines are unheard of”. In markets everything is possible. It is your capital (or your client’s) what is at stake and I just don’t want to hope for the best. I must mentally rehearse worse scenarios ahead than those previously seen. Maybe I am paranoid, but the best offence is good defence. Drawdown reduction is key to long term investment survival.

 I just see two ways to play defence against markets that get “unruly” and experience drastic declines before a primary bear market signal is given.

The first line of defense is to diversify across time frames. This idea is not new but it is worth repeating: We can trade the very same markets with alternative definitions of secondary reactions. Hence, we can split our trading capital between two different ways of applying the Dow Theory. For instance, one demanding strictly 3 weeks confirmed and 1/3 retracement and performing volatility adjustments so that there is a minimum movement; and other one, just requiring 10 or 8 days for a secondary reaction to exist and giving away the 1/3 requirement (as does Schannep) and being less strict with the minimum volatility adjusted movement. By splitting the capital between two alternative ways of applying the Dow Theory (one more reactive, with smaller losing trades, but with a lower profit factor, and the other one less reactive, with better long term profit factor but more prone to occasional big losing trades) one is more likely to avoid a monster losing trading when markets become crazy.

The second line of defence is to diversify your capital across many markets. As an example, one could devote 1/5 to US stock indices, 1/5 to precious metals, 1/5 to energy, 1/5 to US interest rates and 1/5 to agriculturals (i.e. pairs as wheat/corn, soybean/soy meal, etc.). Thus, if you are caught wrong footed, your losses will be, nevertheless, smaller. A 30% loss (or drawdown resulting for some consecutive losing trades, or just by one monster losing trade) when one has just allocated 20% of his capital to the unruly pair of markets is just a 6% loss for the whole portfolio, which is bearable (and quite likely some other market will be winning, hence further mitigating the 6% loss). The key is diversification, as all futures hedge fund manager know. No futures hedge fund manager in his right senses would ever attempt to apply his trend following method to just one market, no matter how good past performance has been in the past. However, we, Dow Theorists, seem to be blinded by the generous Dow Theory record when applied to US stock indices. While I have not yet applied the Dow Theory to, i.e. French or Spanish stocks indices in the last 15 years, one thing I know is that, their charts look quite pathetic, nothing short of the booming pattern seen in US stock indices. What if one day the US becomes like those old Europe laggards? What if one day the US manages to screw its own economy up by imitating the failed policies of some old Europe countries? Would it then be so easy to make 7% p.a. by just being buy and hold? And how would the Dow Theory perform under such conditions?

Well, enough for today, I feel this has been a good “stress test” as we have seen that even when confronted with a profoundly bearish environment, the Dow Theory managed to even make some modest gains. Off course, there is no guarantee to make positive results even under strong secular bear markets. However, it seems clear that the Dow Theory will most probably do better than buy and hold or even many trend following systems based either on moving averages or breakout systems.

In the second post of this Hamilton saga, we will see what would have happened if we had shorted. As an appetizer: Very positive results. While I am not friend of shorting US stocks, it may make sense in some markets. More about this in a future post.

The Dow Theorist

No comments:

Post a Comment