The problem of fake gold bars
You don’t need to be a conspiracy theorist to find this worrying: a 1kg gold bar, certified as 99.98% pure by XRF (X-ray fluorescence) tests, turns out to have been drilled out and largely replaced with tungsten. This bar was discovered only because it was 2 grams lighter than it ought to have been: the forgers failed to add quite enough gold to the outside of the bar to make up for the weight lost when they replaced gold with tungsten. But if they’d gotten the weight right, it would probably still be circulating today.
Of course, there are means of testing gold bars beyond just weight and XRF. If you can weigh the bar accurately, then you can test for purity by essentially dropping it in a bucket of water and seeing how much the water level rises: a gold-covered tungsten bar will displace more water than a pure gold bar. Alternatively, for $3,000 or so you can buy a micro ohm meter, which is easily sensitive enough to tell the difference in conductivity between a pure gold bar and one which is largely tungsten.
But as far as I know, these tests are not particularly commonplace among gold dealers, and in any case only a minuscule fraction of the gold bars in the world are physically traded, changing hands from one owner to the next — the obvious point at which tests like these would be conducted. If you own gold — if you’re a central bank, say, or a physical-gold ETF, or even if you’re just a Ron Paul or Glenn Beck type with your own personal stash — then there’s no realistic chance at all that you’re going to go bar by bar through your own holdings, testing each one.
In the case of gold, then, what JK Galbraith famously called “the bezzle” — the amount of wealth that people think they have, which in fact they don’t have — could be truly enormous. If there are 1.3 million salted 400 oz bars in existence, and each one is 75% tungsten, then that makes 390 million ounces of gold which in truth isn’t there. At $1,660 per ounce, that’s over $600 billion which people think they own but don’t. To put that number in context, it’s roughly half the total quantity of subprime mortgages which had been issued at the height of the housing bubble.
On the other hand, it’s also possible that the number of salted 400 oz bars is zero, that the 1 kg bar recently discovered was an amateurish aberration, and that there’s nothing to worry about at all. But the economic incentives here are so enormous, for people looking to make a fortune in the fake-gold-bar industry, that I very much doubt no one has tried. And statistically speaking, if a bunch of people have tried, then some subset of those people will have succeeded.
So what to make of the fact that no salted 400 oz bars have yet been found? On the one hand, it can be viewed as very worrying — as being an indication that the gold markets are very bad at uncovering such things. On the other hand, you’d think they’d be good enough at it that they would have uncovered at least a small percentage of the salted bars outstanding — and if they’ve uncovered no such bars at all, then that can be taken as an indication there aren’t any salted bars outstanding.
In any case, there’s clearly now serious tail risk for anybody in the physical-gold market. And like most tail risks, measuring and/or insuring against it is extremely difficult. Any store of value has problems, be it fiat currency or sovereign debt or bitcoins. This latest discovery just goes to show that the problems with gold aren’t just the obvious ones surrounding things like the risk that the price of gold might plunge. There are non-obvious ones, too, which have the potential to be even bigger.
Update: Thanks to smart comments from BronSuchecki and Tim Worstall. I forgot that gold is actually used, once in a while, to make things like jewels — which means that bars are melted down pretty frequently. If there was a significant number of salted bars out there, we’d know about it, since you’d notice when one of them got melted down. Which isn’t to say that there are no salted bars at all, but it certainly does seem to imply that there’s nowhere near 1.3 million of them.