Saturday, November 29, 2014

Gold's rising convenience yield

While I may have taken some jabs at the gold bugs in two recent posts, please don't take that to mean that I have it in for the metal itself. Gold is a fascinating topic with a history that is well worth studying. (See this, for instance). In that vein, what follows is some actual gold analysis.

Something weird is happening in gold markets. The future price of gold (its forward price) has fallen below the current gold price. Now in fairness, this isn't an entirely new phenomenon. Over the last year or two, the price of gold one-month in the future has traded below the current, or spot price, a number of times. However, this observation has grown more marked as both the three-month and six-month rates have also recently fallen below the spot price.

This degree of inversion is rare. Except for a brief flip in 1999 when near-term forward prices fell below zero for a day or two, future gold has almost always traded for more than current gold. See chart below, which illustrates the one-month to twelve-month forward price premium/deficit in annual percent terms:

Here's why this pattern has dominated:

Gold's forward price indicates the level at which a buyer and seller will contract to exchange gold at some point in the future. The seller must be compensated for a number of costs they will incur in holding the gold until the deal's consummation, including: 1) taking out a loan to buy the gold and stumping up interest expenses; and 2) paying to store and insure it in a vault. Together, these are called carrying costs.

The buyer of future gold needs to compensate the seller for these costs. Rather than paying the seller an up-front fee, the buyer builds a premium into the price they pay for future gold in-and-above the current spot price, say $5. The future seller of gold can use this $5 premium to cover their carrying costs, thereby coming out even in the end. So future gold trades above spot gold by the size of its carrying costs.

The current inversion of spot and future gold prices seems to break all these rules. The premium that sellers have traditionally required has not only shrunk to 0 but become a deficit. Put differently, sellers of future gold are no longer demanding a compensatory fee for storing and financing the metal. In fact, they seem willing to provide these expensive services at a negative price!

One explanation for the inversion is that with interest rates being so low, the costs of carrying gold have become negligible. This is only party correct. Minuscule carrying costs would imply a future gold price that is flat relative to the current gold price, when in actuality future prices are below present prices.

That leaves only one explanation for the inversion: there is some sort of hidden non-pecuniary benefit to holding the stuff. In futures-speak, this benefit is typically referred to as a commodity's convenience yield, a term coined by Nicky Kaldor in 1939. An analogy to oil markets may be helpful. Oil prices often invert because merchants see potential for future supply disruptions. Having oil on hand during these disruptions is immensely useful as it  spares our merchant the hassle of negotiating his or her way though an oil supply chain that may be severely crippled while ensuring that customer demand is smoothly met. So the convenience yield can be thought of as a flow of relief, or uncertainty-shielding services, provided to owners of inventories of a commodity. If that relief is sheer enough, than the convenience yield will be larger than the twin costs of financing and storage, resulting in inverted markets. (For an excellent explanation of the convenience yield in oil markets, check out this Steve Randy Waldman post).

That's what appears to be happening in gold. Gold merchants seem to be anticipating choppiness in the future supply and demand of the metal, and see growing benefits in holding inventories of the stuff in order to cope with this choppiness. The convenience yield on these inventories has jumped to a high enough level that it currently outweighs the costs of storing and financing gold, resulting in an inverted gold market.

Gold's convenience yield spikes every every few years due to market disruptions, with the last spike occurring during the 2008 credit crisis, the one prior to that in 2001, and the one before that in 1999 when central banks announced plans to limit gold sales. It just so happens that these earlier disruptions occurred when U.S. interest rates were already high enough that they continued to outweigh the metal's suddenly-augmented convenience yield. Inversions were brief and only on the 1-month horizon. Now that a disruption is occurring when interest costs are near zero, a more sharply inverted market is the result, dragging the 3 and 6-month horizons into negative territory. Going forward, all gold market disruptions could very well create sharp inversions of -1 to -2% in the 1 to 12-month horizons, insofar as we are living in an era of permanently low interest rates.

Is gold becoming money?

A number of gold bugs see the current inversion as something quite momentous. To understand why, you need to know that a gold bug's nirvana is when gold is once again 'money'. When something is money, it is highly liquid. The beauty of owning a highly liquid medium is that it can be mobilized to deal with almost any disruption to one's plans and intentions. Put differently, the convenience yield on stored money is very high. One measure of a paper dollar's convenience yield is the interest rate a government-insured certificate of deposit. Locking away cash for, say, 24 months means that the owner loses all the benefits of its liquidity. With 24-month certificates of deposit currently yielding 0.34% a year, the value of those forgone conveniences is 0.34%.

So when a gold bug's dream becomes reality and gold overtakes the dollar, yen, pound etc. as the world's most-liquid exchange medium, that is the equivalent of saying that gold is providing investors with the market-leading monetary convenience yield. And a permanently high convenience yield would result in a permanently inverted gold market (or at least a much flatter one).

Is the current inversion an indication that gold is becoming money? I don't think so. If the augmented convenience yield on gold was in fact rising due to gold's liquidity having surpassed that of fiat money, we'd expect this to be reflected not only in near-term forward prices but along the entire horizon of forward prices. Not only should the 3-month forward prices be inverted, but so should the 3-year forward price. Is this the case? Not really. If you've seen Crocodile Dundee, I'd suggest you go and check out this hilarious post by Bron Suchecki illustrating the extent of gold's inversion. If you haven't seen the movie (you should), check out the chart below.

The first data point is the spot price. Gold forward prices are inverted after that, but only over a narrow range of five or six-months. By mid-2015, forward prices return to their regular pattern of trading at a premium to current prices.

So no, gold is not becoming money. Rather, we are running into some short-term jitters, and merchants think that holding the stuff provides a few more ancillary benefits than before.

Could these short-term supply & demand problems crescendo into longer-term problems, resulting in inversion beyond 2015? I don't think so. Unlike oil and most other commodities, the supply of mined gold is never used up. Ounces that were brought out of the ground by the Romans are still in existence. This means that supply disruptions should never pose a significant problem in the gold market since gold necklaces and fillings can be rapidly melted down into bars and brought to market. While we care if Saudi stops all oil production or if the U.S. corn harvest is terrible,  if South Africa ceases to produce gold—meh.

This means that the convenience yield on inventories of gold will almost always be less than the convenience yield on stocks of oil, since the sorts of disruptions in the gold market will always be shorter and less extreme than in oil markets. Oil supply shocks can be so sharp and enduring that oil's convenience yield remains elevated for long periods of time. The result is an inverted oil market over the entire time horizon. Such inversions are fairly common events in oil markets (once again, see Bron's post).

Gold shocks can never be enduring, so the types of price inversions we'd expect will be fleeting and only appear in the near-term time horizon. Like the one we are seeing now. In sum, we've seen this all before, and no, it's not the end of the world.


  1. Another good post. The Bank of Canada promises that the future price of the financial asset it issues will always be falling at (approximately) 2% per year relative to the spot price.

    But could we imagine a world where the future price of gold is always less than the spot price?

    1. Thanks Nick. I think a world in which gold dust or coins had entirely displaced paper media as highly-liquid exchange media might demonstrate a future gold price below the spot price, or at least a very flat curve. People would be willing to hold these bits of gold despite a negative/flat return since their monetary conveniences would more-than outweigh the capital loss.

      Do Bank of Canada liabilities fall because of the bank's promise? Wouldn't they have fallen anyways since they are highly liquid and don't need to tempt holders with an expected capital gain?

    2. JP Koning, when you say, "Do Bank of Canada liabilities fall because of the bank's promise? Wouldn't they have fallen anyways since they are highly liquid and don't need to tempt holders with an expected capital gain?"

      -It reminds me of your great, "sign wars" post. It all makes me wonder whether, when bank reserves are in short supply, they can be issued at a premium because they have a scarcity value for use for bank settlements. Once QE has created a glut, then new bank reserves will only fetch a "fair value" that represents what value they have as a store of value. That fair value does not incorporate the premium price that creates inflation.

      I really don't know, is my understanding mangled on this?

    3. I think that makes sense. The premium at which reserves can be issued is their liquidity premium. It arises in part because a central bank like the Fed keeps reserves in short supply. Central banks used to massage this liquidity premium, raising it a little bit with open market sales and lowering it with open market purchases, so that the price level would fall either faster or slower. When QE brought a glut of reserves, their marginal liquidity value fell to 0 at which point there no longer existed a liquidity premium to shrink in order to create more inflation.

    4. Yes, thanks, I seem to have got my head around it now!

      What I'm still struggling with is whether the central bank paying interest on reserves (after having created a QE glut) can genuinely substitute for the "liquidity premium" in terms of generating inflation. My gut feeling is that perhaps you can't substitute for a scarcity induced liquidity premium with interest on reserves. In the interest on reserves situation, the reserves will be issued at a "steady state fair value". That situation comes about because only banks get the interest on reserves and they don't need to pass it on to deposit holders and the banks don't set general prices. In the scarcity (pre-QE) situation, banks did need to compete for reserves so they did pass on the liquidity premium value to depositors. But I'm happy to be corrected on this.

    5. The liquidity premium arises from the convenience yield on reserves, which is just a market return. The interest rate paid on reserves is also a return. As they are both returns, they are substitutes: one can do the job of keeping the inflation rate on target as well as the other. The one difference is that at some point the supply of reserves will be so high that the convenience yield can no longer be reduced. The IOR, however, can in theory go towards some infinitely negative amount.

  2. I have no idea how significant gold (and gold derivative) trading really is for banks. Like you say, gold bugs seem to spout a lot of hot air and that makes the whole subject very murky.

    How much basis is there in the legend that Gordon Brown sold off UK gold stocks because they had been leased to banks that sold them and then couldn't buy them back from the market -ie it was really just a bank bailout? If banks have once again got in such a pickle then the issue would be a risk of bank collapse rather than the (inconsequential) risk that people won't have new gold rings etc.

    I suppose people who currently hold the gold would be persuaded to sell it into the market if the price were high enough but perhaps banks won't be able to wait that long or afford to pay enough to cover their positions. The current gold lease rates may reflect a perception that there is a risk of such a fiasco.

    1. I guess I'm just wondering whether gold currently has a convenience yield not because it has become sort after as a general medium of exchange but rather because there is (perceived looming) very specific need for it to cover short trading positions.

    2. Stone, as I pointed out in the post, I don't think any of gold's growing convenience yield is related to any supposed augmentation of its qualities as a general medium of exchange. Your 'looming need' theory makes much more sense to me.

  3. Interesting post. Thanks.

    I tend to eschew the term 'Gold Bug', but there seems to be great sport- Bron, Kid Dynamite - in being really derisive toward people so termed. The thing I struggle with most - and I've asked KD point blank about it - is that no one has been able to articulate a path through our current debt crisis that doesn't lead through a currency crisis on the way to a debt-clearing deflationary bust. The currency crisis will be the result of a policy choice, but the debt crisis is the result of past policy choices.

    Gold seems, in the past, to have done well during periods of currency debasement, and fared well during the 30's as well.

    I will be happy to be wrong about this - I don't want my family to live through the crisis that is coming - but I have yet to get a reasonable explanation that explains how an open money spigot ends well. That's, I believe, the thing that the folks you deem "Gold Bugs" struggle with.

    When I asked KD about it, after some of back and forth, he finally admitted that things are going to end very, very badly. So perhaps the disconnect is merely in timing - with many seeing Gold as a store of value during crises, while others are more interested in shorter term trading.

    One thought about the disconnect between spot and paper (CME) prices. The preponderance of CME trading is not about physical changing hands, but is about trading profits - making it an easy target for manipulation by interested parties (Fed, etc.) . And lest you say the Fed would never manipulate gold, consider that central banks worldwide are buying equities, the CME published pricing for central banks trading the ES, and think back to Paul Volcker's comment about the late 70's, something,like "The big mistake we made was in not controlling the price of Gold".

    If this makes me a gold bug, so be it, but to me it seems basic prudence.

    1. Very good post , one way of clarifying the question is to read

    2. "....a path through our current debt crisis."

      Maybe a debt crisis is on the horizon, I have no idea. But I don't see any evidence of a 'current debt crisis'. Yields on government debt are still very low. Why can't we just muddle on?

      As for manipulation, one central bank is buying equities, the Bank of Japan, but they aren't trying to set a price. Not like the London Gold Pool back in the 1960s, which you may be referring to. But that was to enforce its $35 promise. I don't think central banks really care about the price of gold anymore.

      Thanks for the link, interesting.

  4. A futures contract that is in backwardation incentivizes holders of future production or holders in storage to sell at the spot. Basically it is getting supply on the spot market due to the market's perception of a current supply shortage. This typically applies for oil or other commodities. The fact that it is happening in gold tells me that there is a great amount of current physical demand.