I was over on my friend CD’s site and that suckah likes to stir the pot. The topic is interesting enough to warrant a post, but it’s ground I’ve trodden before. Never the less I may have a little better conceptual model than squawking about mean square variance and variable risk. By the way I’m in no way knocking Rick Ferri. I was an investor in the dark ages and in the land of the blind he had one eye. The podcast is here and I recommend!
My conception is a portfolio lives on a plane and has a shape on that plane. It has a center, and the assets fan out from the center in lines like the spokes of a wheel. The shape and area of portfolio on the plane represents portfolio’s total risk, and the spokes individual asset risk and asset correlation.
A square is a more inefficient plane from a risk perspective compared to a circle
The circle represents the efficient frontier. It is the plane where the assets and their correlations come together and form the least risk. The surface area is a measure of portfolio risk, so clearly the circular portfolio holds less risk than the square. Here is the square which holds the assets:
This is not to scale, just a picture of assets as spokes and their relative correlations with each other. The spokes represent individual asset risk and the length of the spoke is adjusted by the assets volatility and it’s percentage in the makeup of the AA. EM is very risky compared to US, 22.73% SD vs 16.67% SD, but if you own only a dab it tends to tone down the risk in the portfolio on a weight adjusted basis. Note bonds form a near perfect right axis with stocks. The correlation between bonds and stock is nearly zero, meaning as stocks gyrate their values wildly due to their risk, bonds don’t budge and are virtually impervious to that gyration. Here is a snapshot of correlations to US stocks
In the diagram the angle between spokes is equal to the correlation. Gold is NOT to scale but gold has a property in a crash I want to talk about later.
The most efficient portfolio is a circle. It is the portfolio of least surface area and therefore least risk. If the radius = 1 the area of a square is 4 and the area of a circle is 3.14, a 27% reduction in area corresponding to a 27% reduction in risk. Again I don’t want to emphasize absolute values but just to give a visual of how to think about portfolio risk. In accumulation the key attribute of a portfolio is return, risk be damned. In accumulation the risk is managed by the W2. You just work more or longer. The calculus changes in retirement once the portfolio is open to withdrawal and SOR and risk becomes paramount, and minimizing risk is a good strategy for portfolio longevity. You want to make the surface area (risk profile) of your portfolio small and circular if you can.
Look at the square and it’s contained assets. Notice how US, Global, and EM all go in pretty much the same direction. Their contribution to surface area is out sized in the vertical dimension. They add a lot to the height not much to the breadth. Bonds add to the breadth. This is why IMHO it’s important to own bonds. They are built in risk management. Bonds further act as a storehouse of value, like a bank, but bonds don’t grow much either. So they add stability but not much return. Re-balancing means you use a little of stocks growth to sock away some value in the bank. You sell a little of the stocks when they are high and stuff that value in the bonds (bank) for later when things aren’t so groovy. It’s a natural thing for a indexer to do, economize and save and invest don’t speculate, reduce the risk, that’s the indexer’s motto. The whole reason to own an index is because of the diversity it provides and diversity to a point improves risk. Diversity is way over done however. Diversity in a given asset class like stocks is asymptotic meaning after a certain point is reached there is barely anymore advantage to be had, and piling higher and deeper PhD just buys you complexity and portfolio drag. A large cap fund becomes diverse with as little as 30 stocks spread across 10 sectors. The DJIA is diverse.
Look at EM US and Global. Are they diverse? NO. They all go in the same direction. They add a little to spreading out the perimeter on the way up. On the way down they all collapse together, and the ones with the greatest risk collapse far more than the ones with smaller risk. So on the way up a dab of diversity, on the way down nightmare. I analyzed EM’s performance in the 2008 crash and if US stocks dipped 50% EM went down 70-75%. If you go down 50% it takes 100% to get even. If you go down 75%, 150% to get even. The thing that gets you even is growth. The relative growth of US stocks is 10.89% the relative growth of EM is 8.66%. How long do you think its going to take for something down 150% to recover at 8.66% growth compared to 100% at 10.89% growth? YEARS is the correct answer yet everybody insists EM needs to be in the portfolio. Quantitatively a stupid idea IMHO.
People criticize factor investing because it may take decades for small cap value to pay off, but blindly put EM in their portfolios which may take centuries to pay off based on some half assed explanation of increased diversity. If the arrows point in the same direction it ain’t diverse. Bonds diversity is called non correlated diversity because it’s correlation is zero. Gold is like that too, non correlated, as is cash (lets say 3 month T-bills = cash). Those are your store houses of wealth, the bank against SOR-Risk if you will. Stocks are the engines of growth. That’s how a portfolio works.
Gold is a special case IMHO. Gold doesn’t return anything it just stores value. It is a commodity not an investment per se’. So why own it?
Behavior of GLD in 2008 crash
Notice how GLD soared while stocks crashed and recall correlations. This is a strong negative correlation in the face of a crash. Soaring gold gives you something to “sell high” when everything else is going to hell and all the stock arrows are pointing into the ground. You want some hamburgers? Sell some GLD when the S&P is in the toilet. It’s a hedge against SORR. What about GLD today?
It’s time to buy a little GLD (buy low) to get ready to hedge. This is still a strong negative correlation but more approaching zero correlation. Gold gets quiet when stocks go up and becomes quite volatile when they go down. So you can trade some GLD volatility for that stock volatility and reduce the over all volatility. That’s why I own some GLD. Not for return but as a volatility hedge.
So what about re-balancing? If you put some money in the bond (bank) on the way up, you have some money to spend when the crash comes. If you own GLD or maybe a cash equivalent (more later) you have something to buy hamburgers with, so take some money out of the bond bank and buy stocks LOW. The rule is buy low sell high, or sell high buy low. People say “you can’t time the market” but they’d be wrong. If you look at growth after a recession it tends to explode to the upside and buying low is exactly the “right time” to buy. The feature of this kind of risk management is it’s automatic, no human intervention required except to do the mechanics of re-balancing year in and year out. Look at the 2009 chart above the right time to buy was Jan 2009 and if you re-balanced in Jan 2009 you hit a home run. That’s called market timing, it’s mechanical and that part is critical. It takes the dumb assed human who’s constantly trying to maximize profit out of the loop. Dumb assed humans can’t time the market but a mechanical system can capture at least some extra growth.
What about cash equivalents? I consider a tangent portfolio a cash equivalent. What is a tangent portfolio?
The tangent portfolio is the roundest portfolio. It is a portfolio of stocks and bonds that has the most return for the least risk and the least risk is the roundest area. If you look at this portfolio it’s expected return is 6% but it’s risk only 3%. It’s mostly bonds with a dab of stock so if bonds are horizontal it’s just enough stock to turn the bond line into a circle from a risk perspective. If you own 12% stock and the market drops in half you now own 6% stock barely a budge on the net portfolio value. It drops 6%. If you’ve been re-balancing on the way up you’ve stored extra value in the bonds anyway, so a few years in you could loose 6% and still be money ahead. If you wan’t to store some equivalent to cash for a rainy day this is a pretty good way to do it. Lets say you store 2 years of WR (say 200K) in a tangent for 5 years and the crash comes.
lets say you can tighten your belt to 85K/yr.
You have about 3.5 years of money to live on without touching the main portfolio. This is a strong hedge against SORR since you leave the main portfolio alone while using the tangent to buy hamburgers. I did an analysis and burning this fuse portfolio needs only happen once to change the trajectory of SORR in a bad crash to something sustainable. You don’t need to keep refilling this bucket if you have it full to start. It just changes the probability of success to the good in a very Bayesian way. The cost? 2 more years of work. A little GLD and a little Tangent = a lot of horse power when it comes to combating SORR. If you never use it good deal. Just die richer or blow it on a Bentley at 85. Again GLD (non correlated) and a tangent (relatively non correlated since its mostly bonds) reduce the risk.
The last portfolio risk rounder is the Roth. Getting some dough in a Roth is insurance because your retirement is NOT going to happen as you think it is and you can be woefully underfunded.
A mere 200K in a Roth 20 years later is 650K and that buys a lot of end of life care. Remember if you’re married you have to fund 2 end of life plans. You don’t get to suck up all the dough and leave your wife hanging. In your death she will have higher taxes and a reduction in SS income and if you don’t plan for that it will bite her in her ass, no thanks to you Mr bogglehead tycoon. 650K is a nice back up portfolio and virtually eliminates SORR from your life and her life. Notice that 69% of the backup portfolio is accrued interest aka free money when it sits in a Roth.
You further reduce risk by proper tax planning but that’s another post. CD’s post was on complexity and kind of addressed the complexity in a single portfolio. Portfolios can be analyzed simply on the efficient frontier plane to reduce complexity.
Here is a 5 fund portfolio of 3 kinds of stock funds, bonds and GLD
It expects a 7.5% return at 12% risk
Here is a portfolio with the same return but half the risk
You tell me which one you want to own in retirement. This is not complex. The efficient frontier calculator is effectively a square rounder. You feed it some assets and it will tell you the assets and allocations of the best circle. You may feed it 10 assets in may only choose 3 to give you the best circle. The tangent is the best circle. It will then create a line of “better circles” for various risk/reward pairs, and you just pick off what you like. It will also analyze your particular asset allocations you fed it and give you the overall risk and reward of that portfolio.
What is complex, is planning for the unknown SORR and the the hedges described above do that with minimum muss and fuss, but you don’t get to retire at 29. The complexity also comes from trying to win against an arcane tax code designed to separate you from your wealth. You only get to spend what they don’t take,