Gold and the central banks: the game theory

Since I (or, rather, “John Law”) first discussed the matter in 2006, the world has made remarkable strides towards understanding the monetary role of gold. I doubt this has anything to do with “John Law,” but you never know—perhaps his post is responsible for a buck or two of the price. All kinds of people read teh Internets.

In case anyone with a lot of dollars is reading today, my position on the gold price is the same as in 2006. If gold will eventually be remonetized, gold is insanely cheap. If gold will never be remonetized, gold is insanely expensive. It’s one or the other. Therefore, if you guess right about this question, you will make huge profits, and if you guess wrong take huge losses.

Your guess is probably better than mine, so I will refrain from making one. I note, however, that in 2006 the remonetization of gold was a decidedly fringe perspective. Now it appears regularly in the headlines. It is still a long way from happening. So there is plenty of time to hop on this bandwagon before it either rolls to glory, or off a cliff.

My specific prediction for the future of the gold-dollar exchange rate is: if there are more sellers than buyers, the gold price will go down. Otherwise, it will go up. You can quote me on this.

Aside from the traditional jewelry markets, which remain quite important as a demand contributor, what sets the gold price is the balance of buyers and sellers in the market for monetary or “investment” gold. If money is moving into gold, gold goes up. If money is moving out of gold, gold goes down. Given the existence of involuntary sellers (such as gold miners), some money always has to be moving in.

(Actually, I’ve never really understood why gold miners sell gold, beyond recouping the cost of mining. Why don’t they just retain all profits, in gold, on their balance sheets? Why resort to old-fashioned dividends? That’s certainly the way we play it here in Silicon Valley. All the more so for a gold miner which voluntarily sells gold, and whose shareholders are thus betting both ways in the gold market. Ideally, of course, the shares of a gold miner would be priced in gold, but this is really asking too much of 21st-century financial innovation. The 22nd might get to it.)

There is some interesting news on the flow front, which is that central banks have become net buyers, rather than sellers, of gold. Obviously, this trend changes the flow and drives the price up. If it reverses, of course, the gold price will go back down. Most industry observers believe that reserve-accumulating central banks will continue diversifying into gold. I am not an industry observer, but I agree.

Now, here is the interesting part. The game-theoretic process which in the past has selected gold and/or silver as monetary goods, which may just as easily operate in the future, and which may even be starting to operate now, is a form of distributed coordination. Previously, I have described this coordination game as a game involving a large number of small players—retail investors, as Wall Street would put it. However, the game theory works just as well for a small number of large players. Such as the central banks.

Moreover, the game theory only works if the players understand it. This is the nature of any Schelling point. Understanding can spread more easily among a small number of large players, than a large number of small players. Therefore, it seems like a fun exercise to restate the theory in these very different terms. After all—anyone can read teh Internets.

The accumulating CBs (China, Russia, India, Japan, the Gulf, etc.) face a simple problem. Some good has to be the international reserve currency, and thus experience price appreciation due to monetary demand. At present, this good is the dollar (and dollar bonds).

The dollar, however, is diluting at a rapid pace, owing to the fiscal incontinence of USG. If you consider the dollar as an sovereign equity instrument, which I do, USG is in what looks a lot like an equity death spiral. USG is continuously issuing large amounts of new equity to finance its operations. This could end well, but history does not suggest that it will.

(Dollar as sovereign equity, Cliffs Notes version: the dollar is not a debt, since it is not a promise of anything. It must therefore be equity. If the dollar is a share, a dollar bond is a restricted share. For a fully-diluted valuation, restricted and contingent liabilities must be included.)

Thus, central bankers (such as those in Beijing) feel their assets are performing badly. This can easily be made a media issue. Which might involuntarily affect their careers. Thus, they seek other currencies in which to park their cash. Preferably, currencies which are not falling like a rock.

However, central bankers have a unique investment problem which you and I do not face. They have so much money that, anywhere they put it, they will move the market. For example, if reserve managers worldwide switch half their dollars into GBP, the following events will occur—coefficients are entirely random:

1. The pound goes up to $8. 2. Britain’s economy collapses. 3. The BNP seize power and print trillions of pounds to pay their skinhead armies. 4. The pound goes back down to 10 cents. 5. The central banker loses the people’s money and is hanged in the street by mobs.

If you are anyone who has large amounts of money, central banker or no, your goal is to spend that money in a way that does not move the market. Ideally, you would like to buy at a price set by supply and demand (not including you). You would rather not buy at a price set by supply and demand (including you). This is a tricky task in which many are paid much to succeed.

Market-moving purchases—as we’ll see later in the program—pose a special challenge to accounting. They create what George Soros calls reflexivity. Standard 14th-century Italian double-entry accounting, while perfect if you are running a bodega, is not capable of handling this matter. Because central bankers are not used to thinking about monetary game theory, and have no idea how to integrate this with their 14th-century accounting, they fail to see the optimal strategy.

The problem that breaks Florentine accounting is: if I drive the market up by buying, how should I value what I just bought? Should I mark it to the market price? If so, I am marking it to supply and demand (including me)? Or should I mark it to the price I could sell it for? If so, I am marking it to supply and demand (not including me). The larger the position, the larger the difference between demand (including me) and demand (not including me).

Suppose, for example, that you have 50 billion dollars, and you use this stash to buy the entire 2008 and 2009 peanut crops. You triple the price of peanut contracts. Congratulations! Your position is now valued at $150 billion. You’ve made a 200% profit. You’ve made money just by marking to market. You should be a spammer.

This is called “market manipulation,” or more specifically “cornering the market,” and it happens to be illegal. But even if it was not illegal, it would be unprofitable, because you cannot generally profit with this strategy—as you sell, you are driving the price back down. Your peanut contracts are valued at $150 billion—but can you get $150 billion for them? You can’t buy lunch with peanut contracts.

This is called the burying-the-corpse problem, the corpse being the vast quantity of peanuts that you have bought but don’t intend to eat. The accounting profit is indeed a mirage. Unless of course you can bury the corpse—i.e., get some other fool to take all those peanuts off your hands, at anything like the inflated price you have created.

There is an easy way to avoid this entire weirdness. Spread it around. Diversify. Don’t make market-moving purchases. For the standard large investor, and doubly for the standard central banker, distorting the market with a purchase is considered a rookie mistake.

Thus the old-school CB answer to gold, now just beginning to fade. When asked why a return to the gold standard is impossible, the standard answer is: “there isn’t enough gold.”

What this means is that the stock of monetary gold is relatively small compared to the number of dollars it would have to absorb, were gold to replace the dollar as the international reserve currency. (Ie, not even considering the awful possibility that ordinary citizens decide to redirect their savings into the yellow dog and 100%-backed instruments, obviating the entire concept of a reserve currency.)

Thus, if CBs buy large quantities of gold, they drive the gold price up. Or more precisely, if they exchange large numbers of dollars for gold, they drive the gold-dollar ratio up. Or at least, so theory predicts. And for once, practice seems to match theory—at least, in China:

“Gold is definitely an alternative, but when we buy, the price goes up. We have to do it carefully so as not stimulate the market,” he said.

Indeed. Hu Liaoxian is even trying to jawbone the gold market down:

“We must keep in mind the long-term effects when considering what to use as our reserves,” she said. “We must watch out for bubbles forming on certain assets and be careful in those areas.” […] However, officials in Beijing are aware that China’s $2.3 trillion reserves are now so enormous that the central bank cannot buy much gold without distorting the price, so they have adopted a de facto policy of buying in a calibrated fashion each time prices fall back to their rising trend line—“buying the dips” in trading parlance. Experts say that China is putting a floor under the gold price but does not chase rallies once they are under way.

Either Mr. Cheng and Ms. Hu do not understand the game theory of monetary formation—or they do and they are playing it close to the chests. If they—or their colleagues—ever figure out the game, God help the dollar.

When a CB buys gold, four things happen. One: the CB insures itself against the chance of gold remonetization. Two: the chance of gold remonetization increases. Three: the gold price goes up. Four: the buyer looks good, because the assets he bought went up.

How is this different from buying the pound, or buying peanuts? Because the price increase in the pound, or in peanuts, is unsustainable. What goes up has to come back down. For their own different reasons, the pound and peanuts are incapable of absorbing total global monetary demand, and acting as a stable international currency. Therefore, a sophisticated investor of large money avoids generating phantom profits by distorting the market in pounds or peanuts.

With gold, it is different. What goes up can go back down, as it did in the ’80s. (In 1980, it looked rather as if gold was to be remonetized. Then Volcker saved the dollar with 20% interest rates. Of course, at the time America was also a net creditor.) But because we know that gold is a viable monetary system, we know that when gold goes up, it does not have to come down. If it doesn’t come down, that means gold has been (re-)monetized. Peanuts cannot be monetized—they cannot become arbitrarily expensive. Gold can. Therefore, the decision calculi for gold and peanut purchases are fundamentally different.

The gold price has been increasing at roughly 20% a year since 2001. Perhaps coincidentally, the global dollar supply is diluting at rates not too different from this. Betting on the continuation of this trend is not difficult—one the way to bet on it is to buy gold. Which causes the trend to continue. Reflexivity! The dollar itself is a bubble, held up by the monetary demand for dollars—and the dollar does not appear to be an especially stable currency.

Thus there is an entirely different Nash equilibrium out there—one in which all the central banks dump the dollar for gold. This causes the gold price to skyrocket, creating permanent profits for all the reserve-accumulating central banks.

Don’t believe me? Think about it. When remonetization is complete, by definition the CBs will be computing their accounts in gold. Since the price of gold in dollars, under this scenario, is much higher than it is today, it will look like the CB made an enormous profit on the transaction: in exchange for green pieces of paper, now of minimal value, it received good gold. Or it was foolish and held on to the green paper, in which case its bankers are lynched in the street.

This is a self-reinforcing feedback loop. The more gold the CBs buy, the more incentive they have to buy gold. Because if the game ends with gold winning, the game will be scored by how much gold you got for your dollars. This will be a consequence of how soon the CB exchanged its dollars for gold. Devil take the hindmost! A classic panic scenario. A melt-up for gold; a melt-down for the dollar.

In other words, when gold is remonetized, the numerator and denominator on the “gold price” are exchanged. The relevant price is now the “dollar price.” What is a dollar worth? How many milligrams of gold can you trade it for? This piece of paper is a financial security, n’est ce pas? Does this security yield gold, own gold, redeem itself for gold, etc.? No? If you want it to be worth anything, you might want to change that…

Here is the difference between gold and peanuts. No one will ever ask how many peanuts a dollar is worth, because peanuts will never be a monetary good. For one thing, it is too easy to grow them. Gold can be monetized, and peanuts cannot be monetized, because of fundamental physical differences between gold and peanuts.

Every monetary system is a self-supporting market-manipulation scheme of this type. As Willem Buiter points out, money is the bubble that doesn’t pop. In a free market, a currency is stable if and only if the currency is reasonably watertight and does not dilute much. In this panic—the same panic “John Law” anticipated—we see the “dollar bubble” popping, and a new “gold bubble” forming. If someone finds a way to print gold, of course, the gold bubble will pop and some other good will accept its monetary demand—rhodium, perhaps. Or baseball cards.

So, if Cheng Siwei and Hu Xiaolian understood the game theory, they might go ahead and “stimulate the market.” China cannot prevent her purchases from stimulating the market. But she can ensure that when she stimulates the market, she stimulates it first—thus getting the best price. And thus ending up with the most gold.

Collectively, the central bankers of the world might agree that they do not want gold to be remonetized. Individually, it is in their interest to defect from this consensus. As the American Century decays, individual motivations tend to become more prominent. You and I are not in a free market—but the central banks are.

And there is another individual motivation that CBs might have for remonetizing. Suppose a large exporter, such as China, which undervalues its currency and runs a large trade surplus as a result, takes a huge radical step and goes all the way to a 100%-reserve gold currency. The ultimate hard currency. If this succeeds, China is the new England—the financial capital of the world, forever. Everyone else’s money? In a word: pesos. Hard currency is Chinese currency. China’s natural supremacy over the barbarian kingdoms of the West is restored.

There is a practical problem with Chinese remonetization: China has very little gold. Even after the PBoC converts all its dollars to gold—even at the gold price generated by this conversion—it will not have enough gold to back all the yuan in circulation. At least, not at the present yuan-dollar ratio.

When a CB fully remonetizes by converting its balance sheet to hard gold, a fiat-currency note becomes an equity share in the central bank’s gold pool and can be valued as such. What is the share worth, in gold? What are the central bank’s assets, measured in gold? Same question. If the CB has 1000 tons of gold and has issued 1 trillion notes, each note is worth a milligram of gold. Gold accounting—not so hard.

In other words, to fully remonetize to gold, China (or any country, even the US with Fort Knox), will have to devalue its currency. The entire event obviously devalues all fiat currencies against gold, because it sends the price from $1000 an ounce to something more like $20,000. But since the US has a huge hoard of gold, the dollar needs to devalue less against gold than other currencies, such as the yuan. Thus, in any full remonetization, China must devalue the yuan against the dollar.

Now: class, what is the economic effect of a devaluation? What are central bankers around the world struggling desperately to accomplish? Same question! The answer is: “stimulus.” Devaluation is inflationary; inflation is a stimulant. Age-old economic truths. China’s low gold reserve is not a problem, but an opportunity. It is not a bug—but a feature.

Of course, if you resort routinely to devaluation or any other stimulus, you become a stimulus junkie. You get chronic inflation—fiscal needle-tracks. But this cannot happen as the result of remonetizing to gold. Because—duh! At the end of the game, your economy is on the gold standard. You are a healthy financial and industrial power, not a strung-out inflation junkie. England in the 1870s, not Argentina in the 1970s.

And ideally a you’ve gone all the way to a 100%-reserve hard-money standard, which means no more booms and busts. So, when I propose remonetization, what I’m saying to central bankers is the equivalent of saying to a heroin addict: after a hit of this junk, you’ll feel so high for so long, you’ll never want to shoot up again! If you can find a junkie who wouldn’t take this offer, your country has some very unusual junkies.

In previous centuries, the phenomenon we know as a “recession” was sometimes described as a “shortage of money.” I too am experiencing a shortage of money—can anyone give me some money? Doesn’t everyone have a shortage of money? But the term, which seems silly, is actually quite reasonable—a shortage of money is actually an excess of debt. An economy experiences a shortage of money when it has run up unsustainable debts. It is collectively bankrupt. Its actors tend to find themselves individually bankrupt as well. Hence, recession.

For instance, the classical gold standard of the Bank of England era failed because, after World War I, the gold cover (ratio of CB-guaranteed gold debts, to CB gold) was just too high. This, too, was a shortage of money—a shortage of gold. The correct answer would have been to share the pain equally all around, by devaluing the pound and other currencies against gold—i.e., making them reflect their actual gold value. But this was too politically painful; so instead, the entire system was allowed to collapse. Some consider this a good outcome.

But the global economy does not have a shortage of gold. It has a shortage of dollars. It is not oppressed by unpayable gold debts. It is oppressed by unpayable dollar debts. (If lending were restricted to self-interested market actors who need a dollar to lend a dollar, i.e., not the Fed, you’d really see that shortage.) When you revalue gold, carried on the balance sheet at $42 an ounce, to the post-remonetization price of tens of thousands an ounce, what are you doing? Creating one heck of a lot of new dollars.

Therefore, if Cheng Siwei and Hu Xiaolian follow my advice, join the bubble instead of fighting it, and unilaterally remonetize with devaluation, China should experience the following results:

First: dramatic stimulus to the Chinese economy. Second: dramatic inflow of gold into China. Third: dramatic, but temporary, inflation, in the context of boom conditions and general prosperity. Fourth: eventual stabilization, with China a wealthy First World country which is the world’s financial and industrial leader. Fifth: large gold-plated statues of Cheng Siwei and Hu Xiaolian are erected all over China.

Of course, remonetization would also produce enormous profits for gold speculators. I.e., anyone else who jumps on the bandwagon. Currency transitions are inherently turbulent processes. Central banks, as they should, hate turbulence.

On the other hand, if there is no option but to make the transition, it may be better to rip the band-aid off fast than slowly. It is certainly better to understand the situation—and once you understand it, the outcome seems difficult to resist. As in many cases, the CBs may just need to move faster and more decisively than the speculators.

The reason I still expect gold remonetization to happen—in the long term, not tomorrow!—is that there’s simply no other viable alternative. Everyone knows that the global economy needs a new currency. Most can see that the dollar cannot be saved or replaced by any other sovereign currency or basket thereof—because no central bank could tolerate the economic effects of the upward revaluation that would result if its currency replaced the dollar in CB portfolios. When the impossible is eliminated, the improbable becomes a certainty.