The Money Mustache Community

Learning, Sharing, and Teaching => Investor Alley => Topic started by: fuzzhead1506 on January 19, 2016, 11:33:25 AM

Title: Another Market Timing Method....
Post by: fuzzhead1506 on January 19, 2016, 11:33:25 AM
At the risk of having another thread turn into the discussion around Dual Momentum, here is another method (that brooklynguy mentioned yesterday) that deserves it's own thread, I believe. 

(also optional)

(required prerequisite)

(actual method)

It seems that macroeconomic theory, rather than back-testing or finding past anomalies via "psychological drivers", drives this method seemingly similar to Dual Momentum - similar since it uses Moving Averages, but different since it really has little to do with testing one fund against another.  In fact in the required prerequisite reading the author indicates why moving average methods are preferable to momentum (despite performance being roughly equivalent).

It is truly a long read, especially if you "learn yoself" through both entries.  Based on a few looks over Jesse's graphs, it appears not to have outperformed a Buy-and-Hold portfolio of the S&P 500 index in all periods though... thoughts?

TL;DR: 170 bps over Buy&Hold over the last 100 years - is that enough for you to switch?
Title: Re: Another Market Timing Method....
Post by: CorpRaider on January 19, 2016, 11:49:40 AM
Another interesting post from JL.  Would like to see the so called VTT stats run on a bigger sample of time, perhaps running back through the 20's since I believe he set the CAPE trigger at the long run median of ~16, which has barely been sniffed in the past 25 years (or either use the median CAPE for the actual period tested, or maybe the average).  If the median was basically never touched for most of the sample (he had a post about this phenomenon early on in his series analyzing the CAPE), it seems unsurprising that including that signal did virtually nothing to alter the performance from that of the simple MA. 

I've read a lot of the MA stuff from AlphaArchitect and Meb Faber.  I mess around with it in one account where I have no tax or transaction costs.  Really want to add a valuation overlay, just makes sense to me to use that to skew it long if you view it as a risk management tool with a cost versus B&H, attributable to being out of the market for some periods and a lag as JL notes. It seems that the costs for the MA risk management would be lower when expected forward returns are lower due to high valuations.  Maybe need to use Tobin's Q or something instead of or in conjunction with the CAPE.

Overall I don't like his modification of the MA because it basically presumes that those two economic metrics have additional, superior predictive power for recessions than the market as a whole.  I think he's right that if that is true, it will be (already has been) arbitraged away.  Seems like he's basically saying if you can predict recessions better than the market based on whatever data then you can add that to the market's trend/prediction and outperform.  Yeah if you can predict recessions and their end better than the market you can time it, but good luck with that.
Title: Re: Another Market Timing Method....
Post by: GorgeousSteak on January 20, 2016, 12:12:20 PM
I'm surprised and I guess somewhat disappointed there's only been one response to this.  I thought the posts were really well done and answered alot of questions that were brought up in the momentum post.  They were pretty long and technical, but I didn't think that would discourage this crowd, haha.  The other marketing timing systems presented also seemed pretty smart and well researched to me.  I'm not frantically making changes to my investment plan, but I guess its definitely got me thinking, and I am curious what other smart people think.
Title: Re: Another Market Timing Method....
Post by: brooklynguy on January 20, 2016, 01:22:46 PM
I'm surprised and I guess somewhat disappointed there's only been one response to this...I didn't think that would discourage this crowd

I suspect there are lots of folks in this crowd who read the articles with interest, but we probably need somebody to step forward and start advocating for the proposed market timing method in order to get a lively discussion going a la the DM thread (which sort of lost its momentum (pun intended) when Miles got booted).
Title: Re: Another Market Timing Method....
Post by: GorgeousSteak on January 20, 2016, 05:42:36 PM
Well, I suppose I'm most interested in the criticisms of what he's presented.  I don't see any comments on the blog itself, is there any other place where there is discussion on what he's presented?

One thing that I didn't love was his explanation about how timing the market is just a different way of diversifying your assets, that was more or less equally valid (its like the first thing he talks about in the moving average article).  Perhaps if you could say something like I will randomly pick a month each year where I will hold all bonds, and the other 11 months I will hold all stocks, I could see how this is equivalent to having a 1/12 to 11/12 bond / stock allocation.  To me, theres a huge difference if you're using timing, since its an active strategy that switches your position.

I guess I also thought it got a little weak when he finally presents the growth trend timing strategy.  I guess I thought the reasoning behind including the extra external factors was not convincing and adds to the complexity.  Yes, it clearly backtests well, but not being an economist like him, the reports he's referencing as signals mean alot less to me, so its difficult for me to imagine using something like that that I don't understand very well.  The moving average strategy, although it didn't do as well in his tests, makes perfect sense to my mathematical brain.
Title: Re: Another Market Timing Method....
Post by: frugalstudio on January 21, 2016, 01:28:15 PM
I trade about 20% of our funds using momentum on individual stocks, and it works pretty well. It beats the overall market which is just a very long term momentum index anyway. Best of course for tax deferred accounts.

The article is fascinating, but it seems needlessly complex for a simple idea.

One idea I have explored is using timing indicators to buy on dowswings--really just to more precisely measure when the market is a bargain and likely to bounce with a high probability. It is a bit harder than it seems, but once a non normalized (non-bounded) indicator is setup with the right parameters, you can find great buying opportunities with little risk. Like right now 1/21/16 :-)
This seems more in line with the general MM ideas. You could use highs in the indicator to roll more into bonds, and the lows in the indicator to roll back into stocks. Given that the stock market has ALWAYS recovered to new highs, this actually has worked better than momentum for me. See the attached chart for an example using the S&P. I send a monthly update to my family who joins me in this method and it has been much more profitable than the market in general, and it is simple. It does take guts to buy low when the market is spiking down, but that is what we're trying to do--buy on sale...correct?

"The time to buy is when there's blood in the streets."
Baron Rothschild, an 18th century British nobleman and member of the Rothschild banking family
Title: Re: Another Market Timing Method....
Post by: AdrianC on January 22, 2016, 06:35:59 AM
I intend to read the Philosophical Economics articles but haven't gotten there yet.

Just a thought - certain timing methods have been shown to reduce volatility and drawdown while giving very similar returns to buy and hold. If this allows an investor to forgo the security blanket of bonds the overall portfolio return could be much improved.
Title: Re: Another Market Timing Method....
Post by: CorpRaider on January 22, 2016, 11:42:59 AM
That's what his data pretty much confirms.  All these SMA tests I've seen show lower CAGR but much higher volatility adjusted returns and really not large underperformance from the buy and hold.  I do the 260 SMA in one account with no tax or fee implications and I think I'm going to add a value overlay despite the fact that the one he chose didn't really do anything during the period he back-tested.