Between Part 1, Part 2, and Part 3, my unexpectedly-ongoing series on gold (NYSEARCA:GLD) has now generated nearly 1,000 comments. The most recent entry got over 20,000 pageviews. I wasn't expecting any of this. Honest.
I also didn't begin this exercise with the intent to become something of an expert on all the ways in which one can push back against the thesis that gold is currently 40% overvalued. But, again unexpectedly, I have. After hearing them all so much, I might say I understand these pro-Gold arguments and talking points better now than the gold bulls who are actually parroting them. I understand where all the holes are and I understand which leaps of faith are worth taking and which aren't when trying to suggest that gold is on the brink of staging yet another massive long-term rally.
My original thesis -- which I was expecting to be controversial -- can be summarized as follows:
- Inflation is the historical trend for gold (or rather, the price of gold is intrinsically linked to the value of the paper currencies in which it is denominated.
- Gold is massively overvalued relative to trend.
- Future returns in gold will be crappy.
The first point allows us to create the following chart. All it does is track gold's actual current price against the price that one would expect gold to trade at if it was simply allowed to increase at the rate of inflation. Or, rather, if gold went up every year by the same amount that the Dollar went down. (As it should over the long run, no?)
In 1997, we started using G20 inflation instead of U.S. inflation. That's a good spot to start too, as gold in the modern era, especially the last decade, has become the province not just of low inflation economies like the U.S., U.K., and Germany, but also higher inflation nations like China and India. The G20 CPI incorporates that.
Keep in mind, all we're doing is establishing a long-term trend line for gold. Everybody's cool with long-term trends, right? We use these all over the place. The only difference here is that this trend line is fundamentally derived. It isn't a linear, mechanical trend line. Nor is it a technical trend or similar to a dynamic moving average. This trend is based on the way that gold, theoretically, should be priced i.e. inverse to the world's paper currencies.
Most of the planet accepts that notion, but if you have trouble swallowing it, let me tell you the following story:
I was at a boutique investment conference a while back with several super high net worth individuals, institutional asset managers, hedge fund managers, and RIAs. One of the guys in attendance had spent his entire career in the gold mining industry. He commented that "an ounce of gold should always be able to buy you a nice suit."
The audience chuckled, reflected, and nodded in agreement. An ounce of gold would indeed have bought you a nice suit in the year 1900 at $20.67, and it'd buy you a nice suit today at $812/oz. (Save your Jos. A Bank jokes for the comments!)
This is what we mean when we say that gold is a store of value that's inversely linked to the purchasing of the Dollar. A nice suit would have cost me somewhere around $20 a hundred years ago. Today I can get one for around $800. Or, an ounce of gold.
If you aren't comfortable with any sort of fundamental trend line, here's what a simple exponential trend line looks like for gold. Obviously, we want to discard all the data before 1974 when the price was artificially suppressed under the gold standard.
It generates the same basic conclusion.
Still think that gold isn't overvalued relative to trend?
What I'm doing is neither witchcraft nor rocket science. I think it shows pretty conclusively that, relative to historical trend, it is overvalued.
You still may disagree, however, in which case please let me know in the comments section why inflation is a bad trend upon which to base a "fair" gold price. Maybe something other than the value of the paper currencies in which it is denominated determines the long-term price for gold?
If you want to say that gold is not overvalued -- and by extension justify a meaningfully higher price for gold -- you have to do one of two things:
You have to suggest that, well, gosh darnit, momentum and psychology or something else is just gonna take it to new highs.
Or you have to say that the trend is broken, will break (careful here), or that it doesn't have any relevance to begin with.
In the case of the first point, you've now got to back it up and submit some evidence for why gold will rally to $2,000, and more importantly, why that price can then be sustained relative to trend.
The latter is even more tricky because you have to decouple gold from the paper currencies in which it's denominated. This goes against pretty much all of modern history and the few things that we can all agree on about gold without debate. E.g. it's an extremely long-term store of value with no other unique intrinsic utility.
So far, none of the 1,000 comments have done this. I wish someone would!
Someone, please, either show me why inflation is a bad trend to use for gold or present me with a logically-sound evidence-based case for why gold prices will go to $2,000 and stay there while the trend catches up.
In the meantime, we can take all that data and fold it into a model that attempts to predict the forward rate of return for gold.
That looks like this:
It's neither a perfect nor precise model. But it's not bad. It does a sensible job ballparking future returns, and more usefully, it excels at pointing out the general windows during which gold is extremely overvalued or undervalued.
It would have kept you out of gold in 1979-1980 (thank goodness!). And it would have gotten you into gold in 2011 (cha-ching!)
Unfortunately, it also says that 2011 was a level of overvaluation almost on par with 1980. Saying that 2011 was a bubble in gold is almost every bit as controversial as saying that gold is only going to return in the neighborhood of -1%/year over the next decade.
I think those are both statements that have a 90% chance of becoming true in the year 2023. I believe we will slowly recognize that 2011 peak was a moment of irrational extreme. I believe that gold will be a disappointing asset class in nominal terms over the next decade and a potential disaster in real terms.
It's totally possible I'm wrong, though, and that's what I want to talk about today.
We should always be mindful that failure is the best teacher.
A Summary of the Real-life Objections
Overwhelmingly, BY FAR, the most popular objection to this thesis is some flavor of the following:
"The Fed and the global central banks are printing money like crazy and the Dollar and other fiat currencies are eventually going to collapse!!"
I certainly wouldn't advise you to go through the entire comment threads on those pieces the way that I have, but the frequency of other objections aren't even close. Gold bulls may own gold for a hundred different reasons, but when someone like me comes along and tells them that the fair value is still another 40% below today's prices, their knee-jerk response is that I'm wrong because of "The Money Printing."
In the framework of a formal debate, this is a really problematic counterpoint to my thesis.
To start, it's contingent on a future outcome. In order for my thesis to be false, some event needs to materialize in the future. Neither I nor the gold bugs know what will happen in the future. The world's currencies may collapse. Or they may not. Neither of us can say with certitude, despite the confidence that so many of these bullish objectors seem to have.
In a debate, this line of argument scores you zero points. As an investment manager, someone engaged in real-world asset allocation, this line of argument generates zero alpha and could even get you in trouble depending on the degree of wrongness (or make you look smart if you get lucky). It's the same as trading with your gut or based on emotion. There's not a single successful trader in the world who'll tell you that's an effective way to manage real money over the long run.
This line of argument is also problematic in the sense that the evidence and data we have so far suggests otherwise. The world's central banks have been "printing money" for a while now, and in the case of Japan, for a long while. To date, there doesn't seem to be any evidence or even suggestion of hyper-inflation in the world's major economies and currencies. In the case of Japan, the Yen has strengthened a great deal over the last few decades after stimulative policies and government borrowing that'd make our Fed blush.
The second most popular objection is much more sensible. It can be summarized as follows:
"The CPI is bogus and understates the true rate of inflation!!"
Now this is how you score points in a debate setting. One of the first things they teach you in Freshman Philosophy class is that if you want to disprove someone's theory, you attack the soundness of the principles upon which it's based. If you succeed at destroying the foundation for the theory, the entire thing topples unto itself.
In a philosophical sense this is the right way to push back.
The problem with this specific topic, is that attacking the CPI is akin to trying to break the foundation with a wiffle bat.
My response to basically all of these CPI-based comments is:
"Awesome, I agree with you. Send me some better data on global inflation than the G20 CPI that the OECD generates."
That's always met with silence, of course.
The CPI may indeed be bogus. Or it may not. The people that calculate this are really, really smart. Smarter than me and you. But none of that matters because it is, by far, the most widely-accepted and accurate measure of inflation that we have. It's also the one with the longest track record. After its creation in 1919, data was officially backfilled to 1913. It was then estimated all the way back to 1800 by splicing several different historical price indexes together.
CPI isn't the only measure of inflation, of course. There's also the PCE deflator, which is a great metric on its own and it's even preferred over the CPI by a lot of extremely sophisticated monetary economists. The problem for this study is that PCE Deflator is actually lower than the CPI and therefore suggests an even lower fair value for gold than what I've proposed. So that's the wrong way for gold bulls to go about making their case.
There's also the MIT Billion Prices Project. This is really, really cool and I'm surprised more people don't talk about it. With this index, they're attempting to track inflation basically in real-time. And what's more, it only looks at the prices for online retail goods, which is invariably (alongside food & energy costs) the items that most objectors will point out have specifically inflated the most.
As it happens, the Billion Prices Project tracks US CPI rather closely. It's deviated a bit over the last year, but it registered less than the CPI back in 2012 and over the course of its life it has tracked CPI closely enough.
Perhaps CPI is a little understated (and for the record, given the incentives that the government has for it to be low, I do believe could be), but we have some real world evidence suggesting that the degree of understatement is small. Certainly not big enough to support a massively higher inflationary trend line for gold.
Now, nobody would think of attacking the idea that inflation -- whatever the rate -- isn't a sensible trend line for gold. Gold doesn't do anything. It doesn't generate earnings and it doesn't pay a dividend. As a result, it's simply a store of value. The free market can take prices all over the place over the short run. But over the long run, as a simple store of value with no earnings or dividend component, gold's price is intrinsically linked to the paper currencies in which it's denominated. As the value of paper currencies drop relative to a basket of goods, the price of gold goes up. Again, nobody on either side of the debate would question this.
The other problem here is that the CPI is, by definition, the way that we measure the decrease in value of the Dollar. Like, that's how we measure it, folks. The CPI measures how much the Dollar falls relative to a basket of goods. This is the definition of the CPI.
Don't like that basket or don't like the math inside the CPI? As I said, bring me a better, more accurate data set.
This is usually the point in the conversation where someone will say, "But ShadowStats says..."
I feel a little dirty giving ShadowStats Guy attention. He sells a service and data and that's fine. But when you get down to it, like most newsletter writers, what he's really selling is support that what you feel deep inside -- that inflation is much higher than being reported -- is OK and that you are OK for feeling that. He gives you a vocabulary and framework for articulating these feelings.
I certainly don't begrudge him this. And I don't want to sound hypocritical either. In one of my businesses, I sell research & model investment portfolios online as well.
The problem is that the ideas he's perpetuating may be socially destructive, and, over the last few decades, has clearly been financially destructive to a lot of investment portfolios. If you've been betting on this idea that the CPI is way too low and someday is going to explode with hyper-inflation, I can't imagine you have many investment dollars left, do you? Like, you're one of those people who are pissed off on the sidelines because you've experienced none of the recovery.
This framework of belief would have had you shorting Treasuries, piling into gold, avoiding or shorting stocks, etc, while you ready yourself for the fiscal-monetary apocalypse. For the most part, those have been bad, bad trades.
I come to use his data!
So let's say I'm wrong. And let's say the ShadowStats Guy has had it right all along. Let's say that this is how we really ought to be measuring inflation, the way we measured it back in 1980! Let's say that inflation really has been averaging around 9% per year since 1999 (which represents a nearly quadrupling of the cost of living since then).
Let's say, that this chart is the way it's really been, the way our government hasn't been telling us:
Now we can take that data and plug it into our inflation-based model for estimating the gold price.
We'll even start it in 1980, too, as that's where his "alternate" model begins.
Are you ready for this?
Do you want to see what the price of gold looks like if ShadowStats Guy has it right?
Is the suspense killing you?
Here it is:
Under the ShadowStats model, the fair price of gold should be closer to $2,600/oz.
Under this alternate measure of inflation, gold is nowhere close to its fair value. It's under-valued by a factor of two. It was even ridiculously cheap in 2011.
With this as the basis for our trend, forward returns for gold would be expected to be in the neighborhood of 17%/year in nominal terms.
Now, that's not exact because I don't have his exact data. I had to back into it, so our already-imprecise model is arguably a little less precise. But his data isn't that difficult to back into, and by 1999, he even admits he's using a constant 7.1% per year that he adds to whatever the BLS publishes. (Which is an approach totally lacking in rigor, but never mind.)
We can use that same data set to look at other real assets. Real estate, for example. Over long windows of time, real estate is another asset class that is driven mostly by inflation. As the Dollars in which houses are denominated decrease in value, the price of the houses increase. In practice, real estate actually grows at a rate slightly higher than inflation over the long run. Real estate has actual utility, and unlike gold, provides a tangible service or tangible yield, resulting in a slight premium to a "completely real" asset like gold.
Here's how real estate has performed relative to our two measures of inflation:
I realize that no sensible individual would use ShadowStats as a truly alternate measure of inflation. But if we were to accept such a measure as we're doing with this revised gold study, we have to accept some uncomfortable conclusions.
We have to be willing to accept that real estate was 40% undervalued during the peak of 2006.
Was real estate actually undervalued in 2006? Was the bubble a mirage? Did we all just get it wrong?
If we're going to take issue with the CPI and claim it's not a real measure of inflation and use a "ShadowStats" model instead to generate a trend and price target for gold, we have to apply it to other real assets in the economy as well.
We don't just get to pick and choose where we apply our alternate method of inflation.
In any case, some individuals will obviously be more comfortable than others saying that housing was tremendously undervalued in 2006. I'm personally not one of those individuals, but again, the purpose of this piece is to illustrate ways in which my inflation-based gold analysis may be wrong.
A More Sensible, More Deceptive Option
ShadowStats Guy also publishes a "1990-based" model that seems a little less extreme with the amount by which CPI is supposedly understated.
It doesn't sound like even subscribers are able to download the actual data, but the plug appears reasonably easy to approximate:
For this one he seems to be adding a constant of around 3% per year to the BLS's figure.
So let's just assume that since 1990, the CPI has been understated by around 3% per year, as this ShadowStats "model" suggests. We can plug that back into our gold model and redraw it as follows.
Because I have a sense of humor and appreciation for drama (and because it's possible I've read too much Tufte), I drew that first chart normal scale. This time I'll set it back to log scale so that it doesn't look quite as insane:
What a coincidence!
Honestly, I wasn't expecting this.
Assuming ShadowStats Guy's more recent model is correct, gold is more or less fairly valued today.
Technically, it's suggesting a price of $1,253/oz.
Two Crucial Points
There are two really interesting, important points to take home after this 1990-based study.
The first is that I have no doubt that lots of gold investors might say that, yeah, maybe fair value for gold really is around $1200-1300/oz. This feels really sensible, right? Especially since it seems to also line up with the (all-in sustaining) costs to produce an ounce of gold.
By extension, one might also be inclined to say that, perhaps, ShadowStats Guy has it right after all and the gold market has found fair value today. But this is all an exercise in misdirection.
Don't be misdirected! In order for gold to be fairly valued today, one has to accept that inflation has been underreported by 300 basis points per year for the last 25 years. 3% per year!
Are you willing to believe that?
Because you have to in order to say that gold is "fairly valued" today. (Or, of course, you have to deny the connection gold has with inflation and paper currencies, which as I pointed out, no sensible person is willing to do, nor should they.)
On top of that, we can plug this alternate inflation data -- the data where we say, "OK, gold is fairly valued at $1,253/oz" -- back into our real estate model. When we do that, we also have to accept that real estate was more or less fairly valued in 2006 and today needs to appreciate 65% overnight to reach fair value.
Gold bulls, are you willing to admit this?
The second point -- and this is the bigger one for today's purposes -- is that, ShadowStats Guy's arguably, uh, "more sensible" alternate inflation model still doesn't even come close to justifying $2,000/oz.
To justify those kinds of prices, you have to say that the CPI understates inflation by even more than 3% per year and has been understating it for even longer than the last 25 years.
It's not just me that would be wrong. It's ShadowStats Guy as well. He's too conservative here!
You have to use his other 1980-based model, which shows that inflation has been understated by 7.1% per year since 1999 and a steadily increasing amount between 1% and 7.1% per year from 1980 to 1999.
Do you believe that gold is worth $2,000/oz?
Because in order to, you have to be willing to accept that we've been rocking 10% inflation for the last decade or two.
Are you willing to accept that?
I know your food and energy prices are way up. But has your cost of living nearly quadrupled since 1999? Should wages be four times higher? Was real estate really massively cheap in 2006?
I might be wrong. It certainly wouldn't be the first time. My public newsletter is basically one long, ongoing, well-documented chronicle of my wrongness about things.
So it could certainly happen again here with gold.
To justify a sustainable price of $2,000 or beyond, you have to do one of three things:
- Use some crazy-high alternate measure of inflation that 99% of the world and 100% of the world's legitimate analysts and economic figures scoff at.
- Accept that inflation is not a good long-term trend basis for gold and that gold has nothing to do with the inverse value of the world's paper currencies.
- Have some sort of knowledge about future events. E.g. a total global currency collapse or an overnight doubling of the price index.
So which is it?
I love getting beaten up in the comments section here. It's like an analytical Fight Club. It feels refreshing and it helps me refine the thesis further as more and more people push back and let me have it.
For me to be wrong, for gold not to be fairly valued today around $800/oz and by extension return between -3% and +1% per year over the next decade, one of those three factors has to be the reason.
I don't have enough reliable evidence to suggest that inflation should be higher than what the CPI reports.
I don't have any reason to believe that gold isn't intimately linked to the value of the world's paper currencies.
I don't have any knowledge about future events, nor can I forecast them with confidence.
I think this is an acceptable and rigorous framework for understanding the fair price of gold relative to its long-term trend.
But like I said, maybe I'm wrong.
Disclosure: I am long GLD. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.
Additional disclosure: For additional disclosure, see here.