Work On Your Golf Game

In this blog I often look at individual companies and do bottoms up analysis. I tell myself I'm looking for within sector alpha, and maybe I even find some from time to time.  But mostly I do it because I'm a degenerate speculator (gambler) and enjoy it.  This post will be a bit different.

Today we are going to talk about asset classes and re-balancing. The nice thing about asset classes and re-balancing is that implementing such strategies doesn't take a ton of time, so you can spend your time doing other things you might enjoy better like working on your golf game, fishing, playing catch with your kids, or binge watching the wire on HBO.


I've included code so you can recreate backtests I use for illustration. But it's worth noting these are for illustration only. The time period they are run on is just what I had easy access to data on, you'll get different results in different time periods and while I don't know what the results would be going forward I can assure you they would also be different. There is no accounting for slippage, taxes, broker commissions. I selected these with full hindsight bias.  

rebalance royalty backtest
rebalance royalty backtest. GitHub Gist: instantly share code, notes, and snippets.

All of these backtests are for education and illustration and aren't mean to be actual allocation strategies.

Profit Pools

Within industries different types of companies have different ability to capture profit due to their place in the industry. Some will have consistently great returns on capital and some will have consistently terrible returns on capital.

In gold mining sector the royalty companies have consistently great returns on capital. Meanwhile junior miners and explorers tend to have consistently terrible returns on capital. We could make a basket of companies that represented the royalty companies and went long that basket, and we could make a basket of junior miners and exploreers and short that basket. Then we wouldn't have to follow any individual companies. After all some royalty companies will do badly and some junior miners will do well, but it's hard to know which ones. But it's relatively easier to know that the fist basket will do better than the second basket.


Imagine you have a 60/40 stock/bond portfolio where your goal is to have 60% stocks and 40% bonds. The prices on each change and every so often you sell the one that is more than its target allocation and buy the one that is less than its target allocation.  What is really happening here? In my view there are two main competing things in tension.

Regression to the Mean

Things that have done well tend to do less well going forward and things that have done less well tend to do better. Stuff returns to average. So if stocks went up to 70% of your 60/40 portfolio it's probably a good time to sell high (stocks) and buy low (bonds). This is a way of harvesting volatility.

Ray Dalio lays out a theoretical argument for this in a youtube video showing a theoretical portfolio of assets with 10% returns and a 10% standard deviation and various levels of correlation, then showing the probability of lowing money in a given year.

Dalio's point here actually is subtly different in that losses and gains from uncorrelated return streams offset, lowering volatility in the short term. He then argues elsewhere that the lowered volatility allows leveraging returns, but we'll get into that later.

Trimming Your Flowers and Watering Your Weeds

Some people automatically sell the “winners”—stocks that go up—and hold on to their “losers”—stocks that go down—which is about as sensible as pulling out the flowers and watering the weeds. Others automatically sell their losers and hold on to their winners, which doesn’t work out much better. Both strategies fail because they’re tied to the current movement of the stock price as an indicator of the company’s fundamental value.
-Peter Lynch
In fact, when we own portions of outstanding businesses with outstanding managements, our favorite holding period is forever. We are just the opposite of those who hurry to sell and book profits when companies perform well but who tenaciously hang on to businesses that disappoint. Peter Lynch aptly likens such behavior to cutting the flowers and watering the weeds.
-Warren Buffet

Going to our 60/40 portolio over a long period stocks have a higher expected returns than bonds. Adding lower returning assets and then rebalancing to continually buy more of them has a negative effect.


Asset Return Max Drawdown Sharpe Ratio
SPY 9.28% -48.85% 0.55

We'll use the S&P500 $SPY exchange traded fund as a standin for stocks. We kind of expect stocks to return 8% a year with wild swings, so this timeframe of our backtest is actually pretty typical. The drawdowns can be stomach churning, but 8% or 9% a year compounded (before inflation) can slowly build wealth and increase spending power.

Physical Gold

Asset Return Max Drawdown Sharpe Ratio
GLD 1.35% -13.11% 0.10

We are using the $GLD exchange traded fund as a standin for physical gold. Physical gold doesn't produce anything, so over a long enough time period we might expect the return to be similar to inflation, basically holding its value, with wild swings above and below inflation along the way.

I'm not a gold bug, I'm not going to suggest puttting all your money in Gold, because you aren't likely to increase your spending power over time. But let's see what happens when you combine Gold with stocks.

Asset Return Max Drawdown Sharpe Ratio
75% SPY, 25% GLD 8.66% -39.19% 0.60

Adding some gold reduced returns a bit, but also reduced our drawdown and increased our Sharpe Ratio. So maybe we could sleep better at night. But it's hard not to think we are giving up some returns for that good night sleep.

Royalty Companies

My whole blog is mostly about how royalty companies are a better investment. You can argue that todays valuations are higher than this backtest period, or that it's a more competitive space so royalty companies are paying more for royalties now. I'd argue that royalties actualy benefit from higher inflation and this backtest period was mostly a low inflation environment.  Regardless, here's the numbers for the period:

Asset Return Max Drawdown Sharpe Ratio
33.33% FNV, 33.33% WPM, 33.33% RGLD 14.40% -60.02% 0.59

With the benefit of hindsight you would just stick it all in FNV of course. But that would have been a bumpy ride and now you are in the bottom's up stock picking game instead of the sector game.

Asset Return Max Drawdown Sharpe Ratio
FNV 15.44% -57.74% 0.69

Junior Gold Miners

Asset Return Max Drawdown Sharpe Ratio
GDXJ -4.23% -87.07% 0.03

Yes, I'm sure there are better time periods for junior miners than this backtest period, but I stand by the statement that it is a horrible asset class.  

Luckily for us it is a very liquid low cost short, or if you are in a short restricted account it has a very liquid options market where you can buy puts. This kind of tail hedging has a steady strem of small losses with brief periods of large payoffs precicesly when the market is down. It's beyond the scope of this article, but worth mentioning.

Long/Short Royalty/Juniors

If you have two assets with good positive expected returns and low correlation you want to be long each and rebalance. If you have strongly correlated assets and one has signifcantly higher epected returns than the other you generally want to be long the higher returning one and can hedge by being short the lower returning one.

When you short a stock you get the cash for selling the stock up front and take on the liability of buying back the stock later. This makes it pretty easy to increase your long portfolio to more than 100% of your portfolio value. Hedge funds might then tout their net market exposure as the percentage long minus the percentage short. This is usually correct with the difference between gross and net being your factor exposure.  

However, sometimes your longs with go negative and your shorts will go positive and your gross exposure will be realized. It's also worth noting that your shorts have potentially infinite loss potential. Shorts are playing with fire and should be done in small doses.

Here's some various combinations of agressiveness on long/short. These range from 75% net long to 100% net long with a maximum of 50% on the short side.

Asset Return Max Drawdown Sharpe Ratio
25% each FNV, WPM, RGLD, -25% GDXJ 9.65% -30.89% 0.69
33.33% each FNV, WPM, RGLD, -25% GDXJ 12.22% -42.33% 0.65
33.33% each FNV, WPM, RGLD, -50% GDXJ 12.52% -34.64% 0.76
50% each FNV, WPM, RGLD, -50% GDXJ 17.32% -54.41% 0.69

It's probably worth mentioning that Bernie Madoff's returns that were too good to be true were 11.68% with a sharpe ratio of 0.656. So these returns are notable.

Long/Short S&P500,Royalty/Juniors

Asset Return Max Drawdown Sharpe Ratio
SPY 9.28% -48.85% 0.55
100% SPY, -25% GDXJ 9.17% -26.69% 0.56
33.33% FNV, 33.33% WPM, 33.33% RGLD 14.40% -60.02% 0.59
25% each FNV, WPM, RGLD, SPY 13.61% -53.69% 0.63
25% each FNV, WPM, RGLD, SPY, -25% GDXJ 12.26% -32.25%% 0.78
50% SPY, 16.66% each FNV, WPM, RGLD 13.15% -47.30% 0.70
50% SPY, 16.66% each FNV, WPM, RGLD, -25% GDXJ 12.06% -20.65% 0.97
62.5% SPY, 20.83% each FNV, WPM, RGLD, -25% GDXJ 15.23% -27.1% 0.96
62.5% SPY, 20.83% each FNV, WPM, RGLD, -50% GDXJ 14.45% -22.05% 0.86
37.5% each SPY, FNV, WPM, RGLD, -50% GDXJ 18.12% -41.5% 0.86

There are several things to note here.

  • If you look at the first two lines I'd choose the second every time. Sharpe ratio is probably not as good of a barometer as a ratio of annual returns to max drawdown is. Though that isn't perfect either as Treasuries win that with low returns.
  • Short Junior Gold Miners is maybe even a better hedge for the the S&P500 than for gold royalty companies. My gut says that it is less about leveraged exposure to the price of gold and more about exposure to risk on / risk off assets with different long term expected returns.
  • If you started with $10,000 and made 9.28% per year (backtesting 100% SPY) for 20 years you'd end up with $58,994. If you make 13.15% (backtesting 50% SPY, 50% royalties) you'd end up with $118,328, more than twice as much money. No leverage, no shorting, slighly less max drawdown percentage.
  • While none of these are more than 100% net market long the portfolios with total long/short more than 100% can bite you more than the backtesting indicates if the longs go down while the shorts go up simultaneously. Some of the return lines have some eye popping numbers, if there was certaintly on those last 4 lines I'd sign up. But the future doesn't look like the past and 125% long 25% short can be a scary place to be when the markets decide to punish you.


Oil companies stock prices are largely correlated with the price of crude oil. It's tempting to make this a factor, however the low but positive returns and big drawdowns make it pretty tough as a passive weighting to rebalance. Though admittidly this is backtest period includes a negative oil price shock that is not typical.

Asset Return Max Drawdown Sharpe Ratio
SPY 9.28% -48.85% 0.55
XLE 4.35% -59.04% 0.24
SPY 80%, XLE 20% 8.92% -49.71% 0.56
SPY 80%, XLE -20% 6.43% -33.17% 0.47
SPY 100%, XLE -20% 7.81% -41.87% 0.49
SPY 100%, XLE 20% 9.17% -49.48% 0.55

It is possible to do better using rig counts, commodity price relative to production cost curves, etc. But that kind of defeats our work on our golf game philosophy as taking a view on a commodity price is a lot of work.

Follow up work is to find where are the profit pools and poor capital returns in the energy sector. Juniors versus majors? Upstream vs midstream? Offshore vs onshore? Oil royalties? Oil tankers? Drill Rig Operators?


You can find a couple baskets of long only (S&P 500 large cap and precious metals royalty companies in our example) and with periodic rebalancing and without leverage do quite well. It doesn't take a ton of time, though you do have to weather some ugly drawdowns. You could smooth the returns with some basket shorts, or juice them a bit with some leverage, both of which might increase your risk.

You don't have to find a lot of baskets. In Ray Dalio's theoretical example there isn't much benefit beyond 15, and most of the benefit is from the first 4.

So my challenge to the reader is, what are some of your baskets?