Asset Allocation

[MachineGhost]

I found it interesting to read the comments from McKelvey about the average investor passively investing in low-cost index funds.  He all but said their problem is they lack any knowledge of asset allocation.  I have never been a Boglehead, but doesn't Bogle preach that in his education?  It would seem to be a critical oversight, if not.

Anyway, both Hulbert Financial Digest (it tracks and rates newsletter advisories) and academia have both stated that about 90% of the long-term peformance differences among newsletters or fund managers are due to the asset allocation and not any market timing or stock selection.  While that isn't quite literally true in the way it is promoted by self-serving Wall Street (upon a really critical closer look), it is still true enough to serve as useful meme and investment framework.  My impression is that if you happen to not have any exposure to whatever asset class is the fad du jour of the moment, you will lose out big time.  No amount of market timing or stock selection in the other asset classes can make up for the steep performance hit.

There are three ways to engage in asset allocation.  There is strategic, tactical and dynamic.  We're all familiar with strategic as that is what the PP uses.  It is initially set fixed and grows unbounded, until a rebalancing event.  There are a lot of ways to weight strategic, but naive diversification by either capital or risk is robust.  Tactical is the realm of sector rotation and market timing.  Overlayed upon strategic, it exploits the business cycle or market trends by changing [intra]asset "tilt" weights or moving in and out of cash, such as Mebane Faber's QTA model.  Dynamic is shifting in and out of entire asset classes alltogether, which is what DecisionMoose does.  Tactical and dynamic commonly use momentum as the criteria for making portfolio changes, but that is because it is the easiest to implement vs econometric modeling or even forecasting.

Now personally in my PP v1 I use tactical allocation, specifically complex market timing.  The next iteration of the PP v2 that I'm currently working on as a new portfolio will use the other kind of tactical, sector rotation.  Dynamic is still down in the lab because I always find the risk to be too high, but I have my eye on macroeconomic regime switching.  I like my maximum drawdown to be no larger than 10%-15% for any kind of portfolio.  Even the strategic PP is too risky for my blood.

[BearBones]

Good description. What technique is it when you invest based on media hype, piling in at market peaks such as tech stocks in 2000, international equities in 2007, and gold in 2012? Many/most professionals are not exempt from this dangerous heard mentality.

[melveyr]

In some sense the wide 15/35 rebalancing bands introduce a tactical component because momentum of asset class performance plays a role in a vanilla PP's performance.

[Pointedstick]

Good description. What technique is it when you invest based on media hype, piling in at market peaks such as tech stocks in 2000, international equities in 2007, and gold in 2012? Many/most professionals are not exempt from this dangerous heard mentality.

The "depressingly normal" technique? 

Great post, MachineGhost. I've recently been tinkering in my mind with a tactical VP that has three cash buckets, each of which shifts in and out of gold, bonds, or stocks according to macroeconomic indicators such as real interest rates and the marker's Shiller P/E. But I've been having trouble settling on the actual buy/sell points. Buy gold when real rates are above 4%? Why not 3 or 5?

[MediumTex]

When a person is in a 100% rational state of mind he can do all sorts of interesting and creative things.  When, however, he gets frightened or greedy his behavior seems to become much more predictable and I've always thought that this is what the Wall Street firms depended on happening over time with most investment decisions--i.e., sooner or later the most rational investor will succumb to greed or fear and that is when the market will take his money away from him.

Like a gambler in a casino, I think that most investors who tinker with their portfolios long enough will find that the house takes most of their winnings, if not their principal as well.  This is not true of all investors, but it's probably true of many who imagine themselves to be skilled at "beating the market."

"Adam Smith" explores this idea in subtle and entertaining ways in his excellent books The Money Game and Supermoney.

[MachineGhost]

Good description. What technique is it when you invest based on media hype, piling in at market peaks such as tech stocks in 2000, international equities in 2007, and gold in 2012? Many/most professionals are not exempt from this dangerous heard mentality.

That one sure is hard to ignore!  I suspect its because we're all hard-wired with envy and our self-worth and status is not solely determined in an isolation bubble.

https://en.wikipedia.org/wiki/Herding_i ... et_bubbles

It goes without saying that if you can't handle tracking error, you can't handle tactical or dynamic asset allocation.  There goes 95% of Wall Street.  It's particularly painful right now with the market shooting off into bubbleland lunacy again and having minimal exposure.  I've actually been thinking recently I should be a masochist and invest in the strategic PP just to have the actual experience, both good and bad, and to serve as a control to tactical and dynamic.  Its hard not to wince at missing V reversal bottoms in a multi-asset hedged portfolio like the PP.  It could be as simple as just decreasing all assets in proportion to increasing cash to get the maximum drawdown under control without compromising the hedge relationships and V style bottoms.

[MachineGhost]

Great post, MachineGhost. I've recently been tinkering in my mind with a tactical VP that has three cash buckets, each of which shifts in and out of gold, bonds, or stocks according to macroeconomic indicators such as real interest rates and the marker's Shiller P/E. But I've been having trouble settling on the actual buy/sell points. Buy gold when real rates are above 4%? Why not 3 or 5?

I think Mebane Faber produced some charts of varying interest rate levels and which assets performed best.  That's probably what you need.

[MachineGhost]

Here is a very simple implemention of dynamic asset allocation as applied to the PP:

[align=center][/align]

When the line goes below zero, switch to BIL (T-Bills).  Above zero, switch back the respective ETF (SPY/EDV/GLD).

[MachineGhost]

How is the "zero" calculated?

EMA smoothing of the difference between the two EMA smoothed ETF's.

[Reub]

MG, is this what you do in actual practice and if so how has it worked?

[MachineGhost]

MG, is this what you do in actual practice and if so how has it worked?

I have implemented tactical and so far I think it is too early to determine how well it will do because we need another peak-to-trough trend to compare any advantage (I was just in strategic cash, stocks (less) and bonds (more) during 2007-2008 after having proactively dumped PRPFX).  For now, tactical is underperforming strategic (which is OK so far since I'm more concerned about downside risk, than upside gains).  The annual return charts I posted last year show the underperformance.

One way to glean the portfolio advantages of tactical allocation is to read this paper: https://papers.ssrn.com/sol3/papers.cfm ... id=2148660

Managed futures use a trend-following tactical approach.  And equity hedge funds are typically stock picking strategies, typically relative momentum or other exotica.  The paper will show the out-of-sample performance of managed futures as being far superior to equity hedge funds in terms of reducing left fat tail risk and shifting the distribution of returns to the right.

[MachineGhost]

I have decided to drop any kind of tactical asset allocation for the PP and adopt the strategic.  My reasoning upon further study and backtesting is that even though the tactical PP can perform somewhat better on a risk-adjusted basis than the strategic PP, I feel human mistakes, transanction costs and taxes whittle away the edge.  The net difference in the end doesn't seem worthy of the stress and discipline involved.  I will be folding dynamic asset allocation for the PP into my trading VP where I am much, much happier spending my time.

That's not to say I can stop gritting my teeth about the maximum drawdown risk of the strategic PP, but I have to make a choice: its either that or its the above.

[AgAuMoney]

Good description. What technique is it when you invest based on media hype, piling in at market peaks such as tech stocks in 2000, international equities in 2007, and gold in 2012? Many/most professionals are not exempt from this dangerous heard mentality.

That herd mentality is precisely dynamic allocation.

The timing is a problem, but timing is NEVER perfectly correct except by chance.  It is also never perfectly incorrect.  The only difference between your perception of "the herd" and perfectly correct is the timing of the sector rotation.

The timing, or when to shift, is determined by looking at trends.

Obviously you will never shift perfectly correct because the trend has not yet been established.

You might shift perfectly incorrect, but the odds are strongly against you.

So the difference is how soon you can correctly assess a trend.  Too early and you'll hop on the wrong trend.  Too late and you'll miss most of it, just like the herd you mention.

[AgAuMoney]

I do not do dynamic, like you I judge the risk/cost of being wrong to be too high.

I do a slight tactical shift on top of the standard PP.  But only within (at least so far) the bounds of the PP.  I also do a slight tactical shift on my VP, but even less than the PP.


Gold example, in early 2006 started putting all new investment into gold (and silver).  By the end of 2007 was rebalancing on slightly tighter bands out of stocks and longer term bonds into cash.  In summer 2008 moved about 10% out of stock and LTT into cash by borrowing 50k from my 401(k).  In late 2008 moved that cash into gold.  In early to mid 2010 liquidated that gold to pay off the 401(k) and pick up stocks.

Stock example, in late 2007 thru late summer 2008 I was lightening up on my stock portion.  I didn't have my PP in a separate account at that time so it was harder to figure the stock percentage, but overall my portfolio went from about 70% stock at the market peak in summer 2007 to about 30-35% stock by September 2008.  In Oct 2008 I started to rotate money into stocks.  In January 2009 more into stocks at a rate of about 2x-3x my entire normal monthly investment amount.  In Apr 2009 I accelerated the rate to more than my monthly gross income going into stocks each month.  By Jan 2010 I had moved more than 1 year gross income from cash into stocks.

LTT example, last fall I dropped my long-term treasuries down to about 20% instead of (at that time) almost 30%.  By the time I acted they had been dropping for a couple of months, had a blip upwards, and started dropping again.  So far so good.  I will rebalance more into LTTs when they get down to 15%, but unless I'm convinced the trend has reversed I'll not bring them up to 25%.


At no time has any PP asset class been less than 15% or more than 35%.  (Except outside my PP where I do hold more stocks and gold (and silver)).

[MachineGhost]

How is the "zero" calculated?

EMA smoothing of the difference between the two EMA smoothed ETF's.

I realize I didn't answer this in plain English.  The zero line represents cash.  If cash vs cash was plotted, it looks like this:

[align=center][/align]

BTW, I have compared the ETFs to total return indexes and SPY, GLD and TLT specifically have the least tracking error out of all the alternatives, especially for total return 30-year constant maturity Treasuries.