Engineering Targeted Returns and Risk

[MachineGhost]

Upon re-reading this thread, I find the enhanced AWP matrix difficult to understand compared to the PP.  The multiple assets in different quadrants with quadrant caps without any idea of the relative weights or durations makes me mad!

What I feel like telling Ray Dalio about this is best left said on Deadwood.  What especially irks me are the IL Bonds, the majority-weight Equities in the Falling Inflation quadrant and the lack of cash.

I see a mistake in placing Junk Bonds into Inflation Rising.  That would be only true of a short-term ladder but the same argument could be made for any kind of nominal fixed income, i.e. not an inherent feature of the asset.

Real estate I'm always on the fence on.  Unencumbered property easily fits into Inflation Rising.  But yield (rental real estate, tax liens, REIT's) would be in Rising Growth as yield is economic dependent.  I think like the bonds, real estate needs to be characterized by its form.

Here's the new matrix:

[align=center][/align]

I'm increasingly dubious that Inflation Falling means the same thing as "Tight Money".  Basically, it seems like AWP swaps the economic condition of "Tight Money" for the presumed after-effects of "Tight Money".  Maybe it is a philosophical difference.

[melveyr]

I just really don't like the way he visualizes it. Its really confusing. His chart is just two linear dimensions next to each other. It makes a lot more sense when you think about on a 2D graph like this that shows interactions of inflation/deflation with real growth/contraction...



I think stocks do well when expectations about the black dot are shifting towards the right side of the graph. Gold does best when expectations are shifting towards the upper left, and LTT do best when expectations are shifting towards the lower left.

However, I think that graphic they publish is not how they actually do it. One of the professors at my school said they do something similar to a Barra analysis (essentially a macro factor regression) but with WAY more variables and which they then translate into the dimensions of inflation/deflation and real growth/contraction. They then take balanced risk across those two dimensions...

The PP gets to the heart of it though. There is an interview where Dalio reccommends a portfolio of stocks, long duration bonds, gold, and cash: all held at risk parity. Because LTT, gold, and stocks have all exhibited similar volatility lately I imagine Dalio would see the PP as pretty reasonable for a retail investor.

[MachineGhost]

I think stocks do well when expectations about the black dot are shifting towards the right side of the graph. Gold does best when expectations are shifting towards the upper left, and LTT do best when expectations are shifting towards the lower left.

And cash to the lower right?

[melveyr]

I think stocks do well when expectations about the black dot are shifting towards the right side of the graph. Gold does best when expectations are shifting towards the upper left, and LTT do best when expectations are shifting towards the lower left.

And cash to the lower right?

I think cash could be a third dimension... the special situation where the fed tightens to a level where no asset class is attractive. Aside from diluting the portfolio, it essentially only protects you from unexpected twists in the yield curve. I think any other malaise would be roughly captured by either gold or LTT.

Those twists can be quite painful however... Most notably 1981 and 1994. I think that is the biggest differences between a barbell and a bullet portfolio (for example a PP that was 50% 10 year bonds). MT wrote a poem about it a while ago if I remember correctly...

[MachineGhost]

Those twists can be quite painful however... Most notably 1981 and 1994. I think that is the biggest differences between a barbell and a bullet portfolio (for example a PP that was 50% 10 year bonds). MT wrote a poem about it a while ago if I remember correctly...

From what I see of the black dot, I feel AWP's "Inflation Falling" is a conflation of Real Growth and worse, it double-weights equities and nominal bonds.  It makes me very suspicious that Dalio curve fitted the 1980-2000 period.  So, you may be right that cash is a third dimension or it is really just part of the nominal bond dimension in duration terms.  Either Dalio is right or Browne was right, but both cannot be right.

I did not notice the black dot going into the lower right quadrant on those "tight money" years or at all.  So what exactly would perform best in that quadrant?

So for now, I guess we wind up back with the PP, but with more asset classes to consider:

[align=center][/align]

Anyone disagree with the duration labels for the IL bonds?

[MachineGhost]

Today I'm poking a hole in commodites being under Rising Growth.  Commodities are a zero-sum game with negative real returns over the "long run" (i.e. the cure for higher prices is higher prices or technological invention or substitution).  Real prices have been continously declining for hundred of years, nonwithstanding short-term inflationary periods.  It doesn't make any sense to me.  Wouldn't you not have exposure to commodity producers/miners that benefit in your equity?  Also, these two charts are not encouraging:

[align=center][/align]
[align=center][/align]

[melveyr]

MG,

I agree that industrial commodities do not appear that useful to a retail investor. This paper offers a compelling reason to avoid commodity indices:
http://www.parametricportfolio.com/wp-c ... -1.CA_.pdf

There is essentially no "market" for the "indices" to track. So they are constructed using arbitrary rules based systems that can easily suffer from front-running. Gold's simplicity really stands out.

Something else to chew on, going back to TIPS, is that they could potentially offer exposure to the entire left side of my quadrant graph (because they are ambivalent about the price level) because in times of slowing real growth, people place a higher value on almost risk free real payments. Gun to my head if I only had two asset classes to choose from, I would do risk parity between long duration TIPS and equities.

Bending the rules a little bit further and acknowledging that HY Corp usually behaves like equities, one could also do risk parity between TIPS and HY Corps. Since the PP does not hold TIPS or HY corporates, this separate risk parity portfolio could be folded in to provide more diversification while still aiming for the economic neutrality.

Or... you could separate the portfolios into Risk On: (VTI, JNK) and Risk Off: (TLT, GLD, TIP) and do risk parity between them. Each portfolio in isolation is roughly diversified across the price level spectrum, so you are allocating risk based off of the real growth axis depending on how much you allocate to either risk on or risk off.

[MachineGhost]

Hmm, TIPS are hard to understand in the AWP context also.  If LT TIPS act just like T-Bonds, then why bother with them in Falling Growth?  And why bother with ST TIPS instead of T-Bills in Inflation Rising?

From what I've read, TIPS have the principal and inflation adjusted by the CPI rate every year, but you won't get paid that compensation by buying a maturity-less bond fund instead of holding individual bonds.  Furthermore, they have the same exact duration sensitivity to rising interest rates as nominal bonds do!  Rising Inflation Does Not Automagically Equal Rising Rates or Vice Versa.  They are not like I-Bonds where there is no interest rate per se, just a fixed real rate at purchase and a variable inflation rate.  Add in the tax complications of imputed income, the current negative real rates which implies you need inflation to be higher than the spread between nominal and TIPS for TIPS to be a better buy or you're screwed, the fact they respond to real rates in the same manner as any real asset (i.e. price goes up while real rates are falling) and it seems to me they are a cure in search of a problem.

In theory I can see how TIPS could cover the entire left side of the [melveyr's] quadrant, but that is not what intermediate durations did in actual practice as late 2008 shows:

[align=center][/align]

I think to get TIPS to work in that dual manner, the duration needs to be split between ST to replace gold and LT to replace nominal.  But, why the heck would anyone do that?  Gold is far superior (is that even arguable?) and who needs the breakeven pressure of inflation adjustments vs rising interest rates over nominals?  Crazy!

Man, I'm sure pooped out on all of this.

[rocketdog]

Today I'm poking a hole in commodites being under Rising Growth.  Commodities are a zero-sum game with negative real returns over the "long run" (i.e. the cure for higher prices is higher prices or technological invention or substitution).  Real prices have been continously declining for hundred of years, nonwithstanding short-term inflationary periods.  It doesn't make any sense to me.

You could say the same about gold, since it's only appreciated 3% annually over the long-term, which barely keeps pace with inflation.  But we're not buying and holding, we're rebalancing when the volatility works to our advantage (buy low, sell high).  Same goes for commodities.  As long as an investment is fluctuating up and down, I'm happy regardless of the so-called "real returns" over the long run. 

[MachineGhost]

I almost hate that I found these ETF's, but if Goldman Sachs is going to be useful for something besides ripping off muppet faces, then it might as well be for this:

http://www.alpsgsetfs.com/momentum-builder.php

Guess what?  Went out of business due to lack of interest.  This was the only momentum fund applying it to asset clasess.  Pity.  You can't sell an idea to the muppets who aren't clamoring for it.