Gyroscopic Investing

http://seekingalpha.com/article/830031- ... -portfolio

I looked but didn't see a post about this article's approach to the PP, and it intrigued me. If there has already been discussion on it, please show me.

Assuming the author's facts are accurate, it seems to indicate that in a simple 4 ETF Permanent Portfolio, it is significantly advantageous to purchase the top two performing ETFs when you add to your portfolio every 6 months.

According to the table in the article (looking at Janary of 2005 through August 2012),

Buy and Hold resulted in a 107.7% total return, with a CAGR of 10%, and volatility of 9.3%, while the max drawdown was -16.71%
Buy and Hold with an annual rebalance resulted in a 102% total return, with a CAGR of 9.6%, and volatility of 7.7%, while the max drawdown was -14.03%
Using the tactic of purchasing the top 2 funds every 6 months resulted in in a 213.9% total return, with a CAGR of 16.10%, and volatility of 12.4%, while the max drawdown was -14.2%

Currently, my investing is set up automatically to contribute to all the funds equally, and it ends up occuring about every 3 months.

Am I reading this right? What am I not seeing? I understand the volatility is a bit higher for this method, but the max drawdown is less and the CAGR and Total Returns are significantly higher. For me, it seems like a no-brainer. Which, I am certain, is the reason I am hesitant to actually implement it.

It really comes down to how much you want to buy into momentum. There has been a lot of academic research that concludes that there is a momentum effect in the financial markets.

When you look at the 15/35 bands you can see that it benefits most from short-term momentum (you hold more of assets that have done well recently) and medium term reversion to the mean (after a long run you distribute your profits into assets that have under-performed). So the 15/35 already has a momentum bias, it is just arbitrarily determined by when you bought into the PP.

I am not sure I would be comfortable explicitly chasing momentum. Yes the 15/35 bands have some momentum effect, but I mainly use them because they involve fewer transactions. I will probably continue to buy the laggards as I accumulate.

Slotine wrote: These investments are bad as they just add more risk and volatility for no good reason.  They're back-test fitted.

Yeah choosing which MA to follow strikes me as really arbitrary, without much theoretical logic behind it. People just pick which one worked in the past.

It's an interesting article, but there are many holes within.  In addition to fitting an arbitrary MA to fit the intended results, this "study" consists of a very small sample of 2005-2012 data. 

If anyone were inclined to follow this system, I suspect your best bet would be to dedicate a very small portion of the variable portfolio to buying or writing the appropriate bull or bear spread using options against the portion of the portfolio that you would otherwise sell/buy based on the indicator.  At least that way you aren't disposing of the actual assets and kicking yourself when you discover the moving average was simply noise 30-60 days later...

On the opposite end of the spectrum, there was a discussion not long ago about whether it is better to buy the lagging assets vs buying equally into all 4 groups. I think somebody did some back testing and proved that the latter approach was better (I think maybe it was Sophie, if memory serves me right). I don't recall that anyone tested this theory but I did suggest myself that maybe taking advantage of existing momentum might be a better idea. Someone pointed out that this was incorrect PP thinking because it amounted to "market timing", to which I had to wonder why buying lagging assets wouldn't be the same.

I haven't done it yet but as long as you continue to re-balance when you hit the bands I don't see anything wrong with taking this approach and I might very well do it when I make my next purchases (which I do about every six months).

MangoMan wrote: Virtually every blended asset fund, be it stocks/bonds or PRPFX, basically buys the lagging asset. They aim to keep the portfolio percentages relatively constant over time, so when money flow in/out exceeds the cash on hand, they buy/sell assets to keep as close to target as possible. This necessitates buying the lagging asset by design.

If virtually everyone does it then that feeds my contrarian instincts even more.

Seems to me the whole idea of leaving the PP alone until you hit the 35/15 rebalancing bands speaks to the concept of allowing positive momentum to have its way so I don't see why it would be a bad idea to go for the best performing assets as opposed to the lagging ones. If you buy lagging assets you don't know whether you are buying on the way up or the way down, do you? Likewise, if you buy the best performing assets you don't know either. So what is the difference? I have no mathematical formula to go with but  on gut instinct I'm not surprised by a  chart that shows buying the best performing assets does better over time, simply because I think betting on upward momentum will generally do better over time than betting against it. The only drawback to this that I can see is the psychological factor of getting caught up in chasing after the best returns and forgetting the 35/15 rebalancing bands.

I split the difference between the "buy evenly" and your "buy the best-performing" by buying assets in their current proportions.  That's the only way that the portfolio is not distorted by my contributions.

Xan wrote: I split the difference between the "buy evenly" and your "buy the best-performing" by buying assets in their current proportions.  That's the only way that the portfolio is not distorted by my contributions.

This by far makes the most sense to me.  I'm not looking forward to implementing this method though.  My funds are split across 3 accounts (rollover IRA, Roth IRA, trad. 401k).  Starting in January my company is offering a Roth 401k so I'll have funds spread across 4 accounts to juggle.  Ugh...

RuralEngineer wrote: This by far makes the most sense to me.  I'm not looking forward to implementing this method though.  My funds are split across 3 accounts (rollover IRA, Roth IRA, trad. 401k).  Starting in January my company is offering a Roth 401k so I'll have funds spread across 4 accounts to juggle.  Ugh...

I'm at 9 accounts between me and my wife.

Juggling isn't that hard but this is easy for me to say since I already invested a lot of time in figuring it how to do it. You will too eventually.

Tax considerations - much more complicated but I don't think anyone can predict what is going to happen to taxes in the near or long term (though I would bet on them going up).

notsheigetz wrote: I'm at 9 accounts between me and my wife.

Juggling isn't that hard but this is easy for me to say since I already invested a lot of time in figuring it how to do it. You will too eventually.

Tax considerations - much more complicated but I don't think anyone can predict what is going to happen to taxes in the near or long term (though I would bet on them going up).

"Never challenge worse."  9 accounts sounds horrible.  I'll take your word that you get used to it.

Luckily I'm still young enough that I don't have the money to justify having any non-tax advantaged accounts.  Also luckily my wife is a teacher and makes no money, preventing me from having to worry about her side of the finances.  On second thought, I'd rather be rich with a stay-at-home wife and crap loads of money in taxable accounts.

I put the chances of overall taxes going down somewhere between an extinction level asteroid impact and the super-volcano under Yellowstone erupting in my lifetime.  What I worry about is some genius moving from a primarily income/cap gains/dividend tax structure to a VAT.  That would screw the pooch of all retirees with a majority of their funds in post tax accounts.  Luckily a VAT would be pretty regressive so hopefully the liberals would put a stop to that.  I think the best bet by far is to sock as much of my retirement into post-tax accounts as possible.  Particularly since I'll have access to a Roth 401k and the $17,500 contribution limit.

Out of curiosity I tried some so-called tactical measures, i.e. using different EMAs (monthly). I used a 25% allocation to VTI,TLT,GLD and SHY, re-balancing it every 12 months using Adjusted Closing data from Yahoo from 1 Feb 05 to 1 Aug 12.
I go into Cash if the Price is below the corresponding EMA for a particular asset. The results are show below. For Sharpe, I used 1% as the risk free return.

The general conclusion is :
1. You get the best annualized return with no EMA tactic, although you will get a larger draw down. The Sharpe value is also the highest
2. The MAX DD does not steadily decline with increasing EMA parameters. In fact there is a peculiar drop for EMA(10)
3. There is an almost steady increase in the Standard Deviation as EMA parameter increases.

All in all it seems, as has oft been repeated, the basic PP with re-balancing appears to be the best if you can withstand a 10-15% draw-down. If you cannot it is still better to use the basic PP but increase the amount of Cash to reduce the DD %age and td deviation.

EMA Ret MaxDD Stdev Sharpe
1 9.6% -12.3% 7.2% 1.18
2 6.5% -9.0% 5.8% 0.95
4 6.5% -6.6% 5.8% 0.94
6 7.3% -6.0% 5.8% 1.08
8 7.5% -6.4% 5.9% 1.12
10 6.9% -10.0% 6.0% 0.97
12 6.9% -9.4% 6.0% 0.99
14 7.3% -9.4% 6.2% 1.03
16 7.6% -8.5% 6.2% 1.07
18 7.7% -8.5% 6.4% 1.05
20 7.7% -8.1% 6.3% 1.06
22 7.6% -8.1% 6.2% 1.06
24 7.5% -8.1% 6.3% 1.03