I recognize that there is a good case for gold at a time when the price level is rising steadily and there are upside risks to inflation and downside risks to the dollar. Let me start there.
I get a little confused trying to decide whether, if we find out that Fort Knox doesn’t have as much gold as we thought, that’s bullish or bearish. I think the consensus is that is bullish because it means there’s one less holder that could potentially sell into the market…but doesn’t that mean there is more actual floating supply? Because that gold is out there somewhere. Anyway, that is all idle speculation anyway until we find out whether the unaudited stockpiles at Fort Knox are intact. (As an aside, the whole monetary history of why there are gold stockpiles at Fort Knox or the New York Federal Reserve is fascinating. I just finished reading Lords of Finance: The Bankers Who Broke the World, which is a little unfair of a title but really interesting.)
But the $2,910 question is whether the case for gold at the current price is good or not. Okay, let’s be real – if people still believed that the current price mattered for future returns then most of the Magnificent 7 stocks wouldn’t be where they are…most of the rest of the equity market, probably too. Certainly Bitcoin. For today, though, let’s pretend that trees don’t grow to the sky and that your future returns depend at least a little bit on the price you buy at relative to the future price you will sell at.[1]
Erb and Harvey wrote a paper in 2013 called “The Golden Dilemma” which had a wonderful chart in it (Exhibit 5) that showed the annualized 10-year real gold return versus the starting real price. The key insight is the translation to real values. Because the gold nominal price does, in fact, rise with the price level over time to reflect the diminishing value of the dollar. But whether it is likely to rise faster than the price level, or slower, or even decline over some period of time is obviously really important to the question of whether gold right now is a good inflation-hedging asset.[2] The chart below shows the Erb and Harvey chart, updated to today and with a more meaningful x axis priced in terms of the starting real price. The vertical line is the current gold price. All other gold prices are converted to their values in terms of today’s price level, and then the return calculated over the subsequent ten years.
The chart highlights that when the gold price is high in a real sense…when it has outrun inflation for a while…it is likely to underperform inflation going forward. When it has lagged inflation for a while it is likely to outperform inflation going forward. There is a fork on the chart of two different periods, the first being the gold prices when we used to care about value more (the lower fork, where the lowest point is January 1980), and the second being the more recent period where value mattered, but less so. The most-recent point for which we can calculate the 10-year real return is January 2015 when the nominal price of gold was $1278 but in terms of today’s dollars it was $1738 because we’ve had 36% inflation since then. Either way, the current price of gold would suggest that while the nominal price may rise over the next decade, that’s likely only to be the case if the price level rises more (that is, negative real returns for gold). That says that while gold is probably a better investment right now that equities – which have deeply negative expected real returns from the current price – it’s not likely to be as good an inflation hedge as, say, 10-year TIPS at 2% real yields.
Now, I have to address one possibility. There are nuts out there who will say “the fork here is evidence that the changes made to CPI caused it to understate inflation by 2% per year – gold, relative to the price level, is suddenly doing 2-5% per year better than it would historically have done.” It is crazy talk, but given the chart it is due a reply. There are two parts of the refutation. The first part is that the bottom fork doesn’t end in 1980; in fact that was the only fork evident in the Erb and Harvey 2013 piece. If CPI was monkeyed with in a serious way as the conspiracy nuts claim, it happened in the early 1980s. Maybe no one figured it out for a quarter century. I think a better argument is that the first gold ETF, GLD (NYSE:GLD), launched in 2004. The abrupt ease in the difficulty of owning and holding gold almost certainly led to an increase in the allocation to gold in global investment portfolios. If that’s the cause, then the happy fork in this card should be the aberration and we should start to converge on the lower fork over time once that change in the acceptance of gold as a portfolio allocation has been fully realized.
The second part of my refutation is to point to another way that it looks like gold might be a bit frothy. People have pointed to the divergence of gold from TIPS (including me, back in 2024), since gold tends to behave like a long-term zero-coupon inflation-linked bond – in other words, it has a lot of real duration, but for some reason not recently. But showing the deviation of TIPS from gold could mean that gold is expensive or that TIPS became cheap. Look instead at the following chart, which relates two zero-coupon real assets: owned homes, and gold. The line is the price of gold divided by the Existing Home Sales Median Sale Price in thousands (so, one ounce of gold will currently buy 0.726% of a median home right now or, conversely, you need about 138oz of gold to buy the median existing home.
The chart shows three distinct extremes prior to the current rise. The peak in 1980 and the trough in the early 2000s are both obviously gold phenomena. The spike in 2010-2012 is a consequence of the housing bust after the GFC, when home prices drastically overcorrected to clear the excess inventory. And then we have the current increase in the ratio, which either means that gold prices are getting too high, or home prices are getting too low.
I think it’s hard to make the argument that home prices are getting too low.
[1] It is really hard to imagine how it cannot be true that today’s price matters for your return experience. The only way it could be irrelevant is if the growth rate of the asset is independent of the price, so that when today’s price goes up the expected future price goes up the same amount. But this leads to the absurdity that there is no price at which you wouldn’t buy an asset. I know this is really, really obvious to most of us but you have to realize that the belief that future prices are untethered to any fundamental value is the only reason the price of BTC – which can be redeemed for exactly zero, forever – is not zero. So lots of people clearly believe that everything has a ‘value’ if someone else will buy it today, even if no one will buy it tomorrow.
[2] People also hold gold for an end-of-the-world/fiat-collapse hedge. This is in my mind a separate case. If the price level goes up 500%, then the question of whether gold returns 490% or 510% is fairly irrelevant to me. The argument I am making is only salient for inflation at non-hyperinflation sorts of levels. Just want you to know I recognize that.