Copper, Abundance, and Scarcity
Julian Simon vs. Thomas Malthus
In August 2022, my former co-blogger Arnold Kling wrote a critical review of Superabundance. He argues that the changes in the prices of exhaustible resources reflect new information that speculators receive. Therefore, he argues, changes could just as easily be in one direction as in another. He bases his reasoning on a famous 1931 article by economist Harold Hotelling. But when I taught that article to the students in my energy economics class, I pointed out that virtually all the assumptions in the Hotelling model are at odds with reality. Laying out how that matters would take me too far afield. …To Arnold Kling, I would propose the same kind of bet that Simon proposed. Choose ten exhaustible resources, price them now, and wait eight years.
I decline the bet, but let us discuss a single resource: copper.
The Malthusian argument is that copper prices should be headed up, because copper is an essential component in electricity transmission. The goal to “electrify everything” is going to make copper expensive.
The Simonian argument (due to Julian Simon) is that human ingenuity is the ultimate resource. We may be able to find more copper somewhere beneath the earth. We may find a cheap substitute for copper. We may find a way to get more electricity transmission out of a given amount of copper.
The Hotelling-Kling argument is that regardless of whether Malthus or Simon is correct, the futures price of copper already reflects this. There is no reason for me to bet against the futures market, since I have no superior information about the outlook for supply and demand for copper.
There is not a futures market that goes eight years out. For data on copper futures, I used a Google search to arrive at Market Watch. As of January 12, this showed the price of a pound of copper for delivery in various months of 2024, including:
January $3.7415
February $3.7410
March $3.7355
April $3.7500
…
September $3.8150
The September price is 1.96 percent above the January price. What does that tell us? Is that a Malthusian increase or a Simonian decrease?
What if the bet were about the copper price in September? And what if the price turns out to be as predicted by the futures market? Who wins the bet?
That depends on what you use for comparison. Compared to zero, the price will rise by 1.96 percent, and Malthus wins. Compared to the likely path of wages, Simon probably will win. Compared to the expected path of the Consumer Price index, it may be too close to call.
The Hotelling point is that speculators can store copper. If you own a copper mine, you can leave it in the ground. If you are a buyer of copper, you can keep it in a storage facility.
As a speculator, you face a “now or later” decision about when to unload your copper. If you unload it now, you can take your $3.7415 and invest it, earning interest. If you keep it and sell it in September, you will get $3.8150 minus whatever it costs you to store it.1
The January price is known as the spot price. The September price is known as the futures price. They are linked via arbitrage to the interest rate and storage cost.
The spot-futures link means that all of the information about the copper market is contained in the futures price. The market decides what copper will be worth in September. It then works backward, using the interest rate and storage cost, to find out what copper should sell for today.
If we had a futures market for copper that went out eight years, then the price for delivery eight years from now would tell us what the price should be today. The spot-futures differential would reflect the interest rate and storage cost, regardless of whether the future is Malthusian or Simonian.
I take this theory of speculation seriously. It argues that the Malthusian/Simonian argument is already embedded in today’s price of copper. The outcome of the bet depends on these things: how interest rates and storage costs evolve; and whether today’s speculators are over-estimating or under-estimating long-run demand and supply conditions for copper.
I have no comparative advantage in assessing these prospects. And winning or losing a bet would be a matter of sheer luck.
It seems to me that as long as a commodity is highly storable, the Hotelling model should apply. I cannot think of what other criteria the commodity has to satisfy.
Note that in February and March, the futures price is below the January price. It seems as though storage cost is negative. If you want to find explanations for this, look up the term “backwardation.” This phenomenon interested Keynes, among others.
Currently there is a vast reserve of about 700 million metric tons of copper in storage.
Well, in what you might call "active storage", that is, in the form of every finished product ever made of copper. All those wires, electrical components, pipes, pots, coins (er, the 20 microns near the surface anyway), and so forth.
That's over 30 times current annual production and at least theoretically able to be mobilized by scrapping the lowest-valued of these uses and recycled into whatever today's highest-valued use happens to be. Use of copper scrap accounts for something like 1/5 to 1/4 of the global refined copper supply, so it's not negligible.
Non-consumed commodities which are recyclable at market-competitive costs add just another little wrinkle to the Hotelling story, but that wrinkle doesn't detract from that model's fundamental correctness.
Copper is unlikely to be in short supply for several decades at least, possibly never
From the copper alliance (which appears to be most of the copper mining/smelting industry)
"Global copper reserves are estimated at 870 million tonnes (United States Geological Survey [USGS], 2020), and annual copper demand is 28 million tonnes. Current copper resources are estimated to exceed 5,000 million tonnes (USGS, 2014 & 2017)."
See https://copperalliance.org/sustainable-copper/about-copper/cu-demand-long-term-availability/
Reserves are usually ore bodies that people expect to be exploitable at roughly current prices in the relatively near future (a decade plus). In the case of copper, in 2020 the world had about 30 years of reserves and a couple of centuries of resources at current annual usage levels. Even if new copper usage doubles we're looking at 15 years of reserves and a century of resources. We are not going to run short of copper any time soon, and possibly never.
There are two reasons why never is plausible
One thing to recall is that the developed world switched from lead to copper water/gas pipes about a century ago and is now switching away from copper to plastic PE/PEX. That means there are thousands (millions probably) of tons of copper pipes that can be recycled into wiring. That's another decade plus of easily exploitable supply.
Secondly the goal is to get electrical current and signals from A to B. Copper is just the currently best/cheapest means of doing so. Development of room temp superconductors, aluminum? alloys or any number of other things I don't know about may end up producing a better (cheaper) way to transfer electricity. We've got a good century to figure this out.