According to most observers, the speculative bubble-of-the-year award for 2017—had there been one—would have gone to Bitcoin, the value of which soared from under $800 in January of last year to almost $20,000 by mid-December. In fact, though, Ripple (a lesser-known cryptocurrency) racked up an even more astounding achievement, with a 35,000 percent value increase for 2017. Clearly, investors who passed up cryptocurrencies last year missed the boat of all boats.
Still, not everyone is a Bitcoin booster these days. Researchers at Digiconomist and elsewhere have recently attempted to curb our collective crypto-enthusiasm by calculating Bitcoin’s energy footprint, showing that the Bitcoin system currently uses more electricity than the nation of Denmark. The current algorithm for creating, or “mining,” a Bitcoin is based on computing a supermassive number of hash functions. That undoubtedly takes a lot of energy—yet, owing to the secrecy of the process, no one knows exactly how much. Bitcoin trades, which employ blockchain technology (more about that below), also use a lot of energy: Trading a Bitcoin uses at least 77 kilowatt hours (kWh), according to one estimate—enough to power a large American house for a week. At that rate, if all Visa transactions were denominated in Bitcoin, they would use as much electricity as the entire world currently does for all purposes. (Digiconomist estimates that the transaction energy cost is over three times higher, at 250 kWh/transaction.)
Total energy usage for the Bitcoin system (mining plus transactions) is also a slippery figure. Digiconomist estimates 24 terawatt hours (TWh) per year for the energy use in mining the current year’s supply of Bitcoins, as well as for all the transactions of the 16.8 Bitcoins currently in existence. Taking this consumption estimate and dividing by the total number of Bitcoins yields 1,430 kWh per Bitcoin. At the same time, Mark Bevand, an entrepreneur specializing in digital currencies, has used his blog to walk readers through an agonizingly detailed estimate of electricity consumption and comes up with a much smaller figure of 6.78 TWh/year (as of early 2017). Even at that modest estimate, a single Bitcoin represents 404 kWh of current consumption. I’ll use that number for subsequent calculations.
That’s a lot of energy—but compared to what? Traditional currencies also require energy, though estimating how much again requires a little guesswork, as well as basic math. On the high side of energy usage is gold, which still fulfills one of the functions of money—as a store of value—even though almost nobody uses it anymore in ordinary daily commerce. Gold, whose current value is about $1,300 per ounce, costs approximately $1,000 per ounce to mine, according to one estimate. It’s possible to convert that $1,000 to energy in a rough sort of way by using the energy intensity of the global economy, which is presently about 1.75 kWh per 2005 U.S. dollar (the energy intensity of gross domestic product varies a lot by nation and by industry, and overall has tended to fall slowly in recent decades). By this method, the rough current energy footprint of an ounce of gold would be 1,750 kWh. One effort to more directly calculate the energy cost of gold cites 8.5 gigajoules per ounce for two large Australian gold mines, which would translate to about 2,360 kWh/ounce. Since the two figures are within shouting distance, let’s use the lower one. Dollar-for-dollar, at approximate current values, that makes gold roughly four times as energy intensive as Bitcoin (though, as should be clear by now, all of these calculations are only approximate, and making apples-to-apples comparisons is difficult). The high cost of mining gold is one of the practical reasons we don’t use it for all our monetary transactions these days—that, and the fact there’s nowhere near enough of the stuff to account for more than a tiny sliver of global financial transactions.
Paper money understandably has a much lower energy footprint. A $100 bill reportedly costs 12-and-a-half cents to produce, with a rough equivalent energy cost of 220 watt hours, or 0.005 that of a Bitcoin, dollar-for-dollar.
The lowest direct energy cost for money is likely achieved by dollars the Federal Reserve called into existence through its quantitative easing program: Over the past decade, the Fed created about 25 trillion dollars with negligible energy usage by its staff and computers (though it’s hard to access the Fed’s electricity bills in order to confirm this).
Historically, people have tended to want their tokens of value—shells, beads, coins, and so on—to be intrinsically rare or hard to make for fear of inflation and counterfeiting. Paper and ordinary digital currencies have broken with that tendency. These days when a commercial bank makes a loan, it effectively creates the money by way of computerized bookkeeping entries. The energy consumed in the process is in the form of electricity, as well as human metabolic energy used in the activities of negotiating, verifying, managing, and typing.
(Photo: Pierre Teyssot/AFP/Getty Images)
According to Modern Monetary Theory, governments should be able to create as much money as needed, with minimal public debt or energy cost. But MMT acknowledges that, if money creation exceeds biophysical resources, inflation is inevitable. As a way of stabilizing the value of money, some theorists have advocated resource-backed currencies; an energy-backed currency (once favored by Thomas Edison and Henry Ford), for example, could also prevent overuse of resources if it functioned as the key element of a national energy-rationing program.
But Bitcoin is not an energy-backed currency; it’s just an energy-intensive currency. Having a Bitcoin doesn’t entitle anyone to exchange it directly for 404 kWh of energy; in fact, at current ballpark retail values ($0.10 per kWh for electricity, $10,000 per Bitcoin), a single Bitcoin could buy 100,000 kWh of electricity. The point is that there are less energy-intensive currencies, and a more widespread use of Bitcoin therefore implies more energy demand for society as a whole, which uses an enormous amount of currency to do its business.
This is a problem because delivering energy requires infrastructure and fuel. If the world is to avert catastrophic climate change, we will have to reduce consumption of fuel (coal, oil, and natural gas) and build vast new infrastructure (solar panels, wind turbines, and energy storage) very quickly. Our only realistic pathways for doing this entail reducing overall energy usage as much as we can without wrecking the economy. Therefore, the 21st century happens to be a terrible time to introduce a new fast-growing currency that could dramatically increase society’s energy needs.
In fact, the situation may be worse than portrayed so far. The technological basis for Bitcoin transfers (which occur after mining) is the blockchain, a digital ledger in which transactions are recorded chronologically and publicly. There is widespread and growing interest on the part of banks, credit card companies, online retailers, and even governments in using the blockchain to facilitate global commerce and electronic data storage. It’s also being used to guarantee ethical supply chains. There are even utopian notions that blockchain technology, combined with a universal basic income, could do away with the monopoly powers of banks and governments in money creation, thus decentralizing the economy and empowering ordinary people in ways no mere political revolution has ever done. But since every block in the blockchain is a ledger (and therefore a file) that exists in many copies, and the chain is always lengthening, the computer resources required for the calculation, transmission, and storage of this information are always increasing, as are the energy requirements.
A dramatic increase in adoption of the blockchain (and of the cryptocurrencies that use it) would therefore seem to depend on somehow improving technology so that each computer calculation uses dramatically less energy. Computational energy efficiency is commonly measured in floating point operations per second, per watt—or FLOPS/W. This efficiency has risen exponentially in recent decades, but so has overall computation, leading to steady increases in total energy demand to power computers of increasing number and variety. There is certainly the prospect for further dramatic FLOPS/W improvement, but the pace of society’s explosive growth in the use of computers for every imaginable purpose could easily overwhelm such improvements, even without added energy demand from the widespread adoption of cryptocurrencies based on the blockchain.
So what does all of this really mean, and where is it taking us? The current fascination with Bitcoin and other cryptocurrencies (Ripple, Litecoin, Ethereum, Peercoin, ShadowCash, etc.) forces us to think about money itself—what it is and what it does. We may increasingly find that the distinct functions of money—as a store of value, a measure of value, and a facilitator of trade—can best be served by an ecosystem of fundamentally different currencies. The current cryptocurrency mania also leads us to think again about how money relates to real value—which ultimately comes from energy and physical resources like food, forests, fisheries, soil, water, and mineral deposits. Having more currency options will be of little use to us if our real sources of value are depleting, eroding, and disappearing. An ideal currency should give us feedback with regard to the status of these real sources of value and thereby help us conserve them rather than spending them into oblivion.
Over the short term, the demand for Bitcoin may continue to grow. Financial institutions and private investors may turn to it because it is an uncorrelated asset: Even though its value is highly volatile, it is not negatively or positively correlated with stocks, bonds, or gold—which increases its utility as a portfolio addition. Further, those who worry about currency controls in the event of a financial crisis may increasingly see Bitcoin and other cryptocurrencies as providing a form of security that gold and other stores of value can’t match. Overall, digital assets seem poised to transform many aspects of our financial system. Likewise, there appears to be little headwind against the adoption of the blockchain by industry and government.
Sooner or later, the energy appetite of most blockchain-based systems is likely to be reined in by replacing the proof-of-work algorithm (which governs Bitcoin mining and exchange, and requires an expensive computer calculation) with a proof-of-stake alternative (in which miners are chosen by the size of their stake in the system)—a move that would trade a little transaction security for far greater energy efficiency. Moreover, cryptocurrencies that use less energy than Bitcoin will continue to gain market share. As for Bitcoin itself, unless its energy efficiency improves dramatically, demand will eventually be self-limiting. Of one thing we can be confident: If the energy requirements for Bitcoin were to grow to equal the entire energy consumption of the world, the currency’s real value would fall to approximately zero. That’s because there wouldn’t be anything left in the economy to exchange it for.
