Investing is just as much about learning what to not invest in as what to invest in, and I’ll add GDMN, WisdomTree’s Efficient Gold & Gold Miners ETF to that list. Seems like it would be good for a RP portfolio, but it just isn’t. Check out the backtests, correlations, and my commentary here!
I have been intrigued by GDMN, for quite some time, as it seems to have a lot going for it, at least in theory:
- Heightened exposure to a particular, uncorrelated premium
- Capital efficient
- Low fee
- Investing in commodity producers is favored by some RP investors, namely Evoke Adviser’s RPAR and UPAR funds.
About eight months ago, I started a very small position in it, based on my curiosity. I find that my ability to pay attention to a fund blossoms as soon as I have actual money in it, no matter how little, so I jumped in with a small stake. Since then, my investment has been basically flat, while GLDM, which just tracks the price of gold itself, has been up mid-single digits over the same period.
No big deal, but noticing the divergence led me to wondering why. My expectation was that if gold prices are up, then gold prices PLUS exposure to gold mining equities would be up even more, since all you are doing is adding equities (with positive expected return, right?) that are profiting from the rise in the price of the commodity they produce.
Faced with this small conundrum, I got to tinkering. Initial backtests showed that not only is GDMN not great at enhancing returns from the rise in gold prices, it doesn’t even keep up with them. Furthermore, it may be actively bad for your portfolio, as the gold miner equity side seems to sabotage the gold price side of the fund. Hmmm…
What is GDMN?
First off, from the horse’s mouth:
Highlights:
- A 90/90 fund with long positions in gold miner equities plus cash as collateral for a gold futures strategy.
- Capital efficient, with heightened exposure to gold (well, sort of, as we’ll see) at leverage rates better than an individual investor could get on their own.
- As for the Gold Miner equities portion, it holds market cap weighted positions in 46 worldwide gold producers. The equity position seems pretty similar to GDX, Van Eck’s Gold Miner ETF, which holds 52. I compared the holdings of GDX and the equity portion of GDMN here, and to the naked eye, they are pretty much the same thing.
- Its expense ratio is .45%, compared to .51% for GDX, so if you were scoring at home, you could look at it as a six-basis point discount for adding a+90% position in gold futures on top of the equity portion. Nice.
Ideally, this is a way of adding to one’s exposure to gold, and would seem to be a no brainer for people already investing in GDX. GDMN seems intriguing to Risk Parity folks, too, as many popular RP-based portfolios (The Dalio All Weather, the Golden Butterfly, and then the RPC Stability, Income and Growth portfolios) contain sizable positions in gold.
The capital efficiency allows the RP investor to pursue higher expected returns while still maintaining a diversified portfolio. One way to do this is through GDE, WisdomTree’s Efficient Gold Plus Equity Strategy Fund, which is similar to GDMN, but just has equity exposure to the S&P 500 instead of a basket of Gold Mining companies (in other words, SPY instead of GDX). GDMN would seem to be an enhanced version of GDE when gold prices climb, as the GDX portion would seem to have special benefits from the rise in gold prices.
The Portfolio Backtests
As I mentioned, I started running a few quick backtests of portfolios with GDMN either in place of GLDM,or in place of GDE, and started seeing some troubling numbers. Yet, those backtests weren’t very systematic, and with GDMN’s relative youth (launched in December of 2021), the backtests couldn’t tell you much anyway.
I then decided on a more systematic series of tests using GDX plus GLDM as a proxy. This extends the backtests back to May of 2006. With more systematic and longer backtests, how did our proxy for GDMN do?
1) Golden Butterfly portfolio
This is a modern classic with a healthy gold slice of 20%, big enough that we should be able to get a good look at GDMN’s impact one way or the other. I ran three portfolios, the original with 20% in GLDM, the second with an additional 18% in the S&P 500 and an 18% in GLDM and then 16% subtracted as cash, and then a third with +18% in GDX, 18% in GLDM and -16% in cash.
Golden Butterfly backtest. June 2006 to October 2023
Of the three, the one with GDMN is clearly the worst. Slightly lower CAGR than the Original, but with higher volatility, larger drawdown, and lower risk-adjusted return. The version with GDE was the best of the three for CAGR, due to the higher effective allocation to the S&P 500, but also had the biggest drawdown, for the same reason. In this sense, it may be straying from the original intent behind the GB, and I’ll leave that to the reader to decide. The one with GDMN strays just as much from the original construction as well, and has even worse performance.
2) Capital Efficient Butterfly portfolio
Next, let’s see how GDMN does when compared to other capital efficient portfolios. The Capital Efficient Butterfly is my newest portfolio and takes advantage of the arrival of capital efficient ETFs to create a relatively balanced portfolio with 75% additional leverage. In nominal terms, it is 40% NTSX, 35% RSBT, and 25% GDE, which means that it is effectively 58.5% equities/59% bonds/22.5% gold/35% managed futures.
For the first comparison, I use the GDMN proxy instead of the GDE proxy, which keeps the equity and gold percentages the same. For the other comparison, I’ll use UGL, a 2X levered gold fund, which keeps the total leverage in the portfolio the same, but does drastically increase the impact of gold prices in the portfolio, since now gold is effectively at 45%, and equities are just 36% (I don’t recommend doing this; this is just for illustration!). One note: the Cap. Eff. Butterfly contains managed futures, so the relative youth of the managed futures fund MFTFX constrains the backtest by four years, to May 2010.
Capital Efficient Butterfly Backtest. May 2010 to October 2023
Even when I put GDMN in the mix with its capital efficient peers, it doesn’t come out looking very good. It’s the worst of these three here, as well, and has the lowest CAGR, the highest volatility, and the biggest drawdown. The version with GDE is clearly the best, matching the version with UGL in terms of volatility and drawdown, but with almost 2% better performance.
Isolating Gold Miner Equities
For the last back test, I’ll eschew portfolios and see how a deconstructed version of GDMN compares to its peers. Again, I’ll use proxies for the funds to extend the backtests quite a bit farther back, to June of 2006. I’ll use 90% GLD and 90% GDX for GDMN, 90/90 GLD and VFINX to replicate GDE, and the 180/-80 GLD and cash to recreate UGL. I’ll also put in the unlevered performances of GLD and GDX for comparison (though these won’t show in the link to the backtest below):
GDMN proxy backtest: June 2006 to October 2023
This is like the moment where the mechanic says they’ve found the source of the problem for your car which billows smoke like a locomotive and whose engine sounds like ten money banging on pots and pans. The difference in the portfolios is the allocation to gold mining equities, which just don’t generate returns. They are terrible on their own (in the form of GDX), and nuke the return you would get just on gold alone when combined. We have a pretty long backtest here for a period where gold itself has not done especially well, but it has at least been positive while the performance of the gold miners has been dismal. GDE is obviously the strongest of these due to a great 2010s for the S&P 500, but even gold alone or leveraged doesn’t look that bad.
Correlation to the Rescue?
For our Risk Parity purposes, CAGR is not the be-all, end-all. If an asset class has a return profile that is uncorrelated with other asset classes, especially equities, it can still serve an important function in a portfolio even if by itself it has meager returns. Meanwhile, if an asset has a negative correlation with other asset classes, it can be even better in a portfolio. Uncorrelated and negatively correlated assets will smooth out returns in a portfolio and lead to higher Perpetual Withdrawal Rates.
The catch is that there still need to be positive expected returns, and here, .42% is so low that the correlation really would need to be fantastic. Still, there’s a chance.
Correlation Matrix June 2006 to October 2023
Nope, there is no salvation here. GDX’s correlation with stocks is fairly low, easily in the territory I consider uncorrelated (.5 at the absolute max, but generally in the -.3 to +.3 range), but it’s more correlated to the S&P 500 than gold alone. There appears to be no correlation benefit with gold mining equities to compensate for the reduced returns.
The Verdict
I’ve been keeping track of preferred assets since I started the blog, such as the best fund for commodities (PDBC), the best for board exposure to international equities (DFAX), or the journey to settle on the best US small-cap value fund (first VIOV, then AVUV).
Meanwhile, GDMN:
- Strike 1: harmful performance in portfolios.
- Strike 2: negative performance by itself (or rather, by a similar proxy)
- Strike 3: no real correlation benefit.
Guess what? It’s out.
It looks like GDMN will be the first “rejected” asset: a fund that seems like it might fit well in a Risk Parity portfolio, but upon further investigation, does not. True, I’m no fan of TIPS, and not really one of REITs, but those are asset classes, so they are not really the same. In the case of GDMN, and by extension GDX, they seem to not just fail to help a Risk Parity portfolio, they would seem to make one actively worse.
If you are a Risk Parity-oriented investor who wants gold exposure in your portfolio, sticking with GLDM seems to be the simplest and most straight-forward way to go. If you want capital efficiency in your portfolio, then GDE holds promise, though please do not just jump into the world of capital efficiency without learning more (such as here, here or elsewhere on the internet, such as here). UGL, the 2X leveraged fund provides a different variety of capital efficiency, and may work for some situations, but I don’t know too much about it (yet).
There is still much to be explained with Risk Parity portfolios and the presence of gold in them, but for now, the bottom line: