Wednesday, August 26, 2020

18 things about Tether stablecoins

Before I start my list, a bit of introduction.

Tether is a stablecoin. It happens to be the most popular stablecoin in the world.

A stablecoin is a digital IOU that is implemented on a blockchain. In Tether's case, it takes the form of a U.S. dollar-denominated IOU implemented on the Ethereum blockchain. Tether holds U.S. dollars in a traditional bank account. It issues digital blockchain-based Tethers that are convertible into those bank account dollars at a 1:1 rate. This promise is what stabilizes them. And so a user can send some Tethers to another Tether user, say as payment, and neither party need worry about bitcoin-style price disruption.

If you didn't understand any of that, think of Tether as basically PayPal, except on a decentralized database instead of a centralized one.

Thoughts, facts, questions, and interesting tidbits in no particular order:

1. According to Coin Metrics, a financial data provider, Tether is now doing around $3.3 billion per day in transfer value. It just flipped bitcoin's daily volume of $2.94 billion.

2. PayPal did $222 billion in total payment volume last quarter. That's around $2.47 billion per day. So Tether at $2.94 billion is moving more value each day along its network than PayPal is (!).

3. There is a popular theory in the cryptocurrency community that expansions in the supply of Tether are being used to manipulate the price of bitcoin. I don't subscribe to this theory. It strikes me as far-fetched, much like the theories about gold price suppression.

4. U.S. citizens and residents are prohibited from using Tether's version the U.S. dollar (with one small exception).

Tether legal disclaimer [source]

5. Does anyone know who regulates Tether? (Yes, Tether Limited is regulated by the U.S's FinCEN. But who regulates Tether International Limited? Its terms of service says that it is based in the British Virgin Islands. But a search of the BVI's Financial Services Commission doesn't indicate that Tether International Limited has been registered as a money services business.)

6. According to a report by Chainalysis, a blockchain analytics company, Tether is probably being used to evade Chinese capital controls. Chainalysis estimates that $50 million Tethers leave East Asia each day.

7. One reason Tether is popular is because it doesn't collect information about ~99% of those who deal in Tethers. It only does due diligence on the minority who want to cash-out of Tethers (i.e. withdraw Tether dollars in a bank account) or cash-in to them (i.e. deposit dollars into Tethers via a bank account). So if you are content to just accept them and then pass them on, Tether shrugs. This makes Tether stablecoins an incredibly hands-off way to interact with digital U.S. dollars. (By the way, almost all stablecoin issuers adopt the same unknown-wallet-to-uknown-wallet policy as Tether.)

8. In its recent report on "so-called" stablecoins, the so-called Financial Action Task Force hinted that stablecoins must abide by the same diligence requirements as a bank. So I doubt that Tether's "hands-off" policy can last much longer.

9. This hands-offness attracts unsavory users. I've written before about MMM BSC, a global ponzi game that uses PAX stablecoins. (PAX is a smaller stablecoin). According to Coinmetrics, Tether stablecoins on Tron (a blockchain platform) are often used for "dividend schemes." However, I have not run across a single ransomware operator that uses Tether for ransom payments. Bitcoin (and to a lesser extent Monero) still dominate the ransom market.

10. Tether freezes a lot of addresses, according to Eric Wall. More than any other stablecoin.

11. I have never used Tether. (For the sake of experimentation I have tried PAX, a smaller U.S. dollar stablecoin).

12. The collapse of global interest rates since COVID-19's arrival has helped popularize stablecoins. A Tether stablecoin yields 0%. So post-collapse they have become more competitive with dollars in a bank account. Ironically, the collapse in rates also hurts stablecoin issuers. Issuers can't make as much interest on the reserves they hold to back their stablecoins. (More here).

13. Tether was initially launched as RealCoin in July 2014. It rebranded later that year. I speculated about the emergence of Tether-like instruments in 2013. I didn't call them stablecoins. I called them stable-value cryptocoins and stable-value crypto-currency. (Forgive my self-promotion).

I believe it was Vitalik Buterin who coined the term stablecoins?

14. The popularity of Tether hints at the dominance of the U.S. dollar. No one seems to be using euro, yen, or yuan stablecoins.

15. There are currently $13 billion worth of Tethers in circulation. There are currently around $31 billion PayPal dollars in existence (PayPal refers to these balances as "Funds payable and amounts due to customers").

PayPal Q2 2012 report [link]

16. Tether can't be used in... Singapore? (Also Cuba, North Korea, Iran, Syria, Venezuela, Crimea). Anyone know why?

17. It's common knowledge that the dollar-denominated assets that Tether holds to back Tether stablecoins are questionable. Shenanigans have occurred. (I wouldn't touch them. I wouldn't want family touching them.) But people still hold Tethers and deal in them. Why? Many institutions and trading outfits use Tethers as a bridging mechanism for arbitraging the price of cryptocurrencies across various exchanges. Given that Tether is their preferred medium for this, it speaks to the poor level of due diligence in the industry. It also speaks to the network effects enjoyed by first movers. So-called sophisticated crypto traders are using dodgy Tethers rather than safer but newer stablecoins because much of the liquidity has already attached itself to Tether. 

18. A big part of the cryptocurrency trading universe is denominated in Tethers, not actual U.S. dollars. So if Tether starts to collapse, cryptocurrency prices would actually hyperinflate. (Real cryptocurrency prices wouldn't budge.)

19. Tether has co-opted the unicode currency symbol for the Mongolian tugrik, ₮.

Feel free to add your own factoids in the comments.

Friday, August 21, 2020

The case for banning gold mining

The Kalgoorlie Super Pit Mine in Australia

Does the world need gold mining?

Let's think about what a world without farming look like. If all farming came to a stop, we'd soon use up all of our inventories of wheat, soy, rice, and vegetables. Mass starvation would rapidly ensue. A world without crude oil production wouldn't be much better. We have plenty of the stuff above-ground. But since oil products are destroyed in usage, we'd run out pretty quick. Society would grind to a halt.

But if gold mining were to suddenly stop, nothing bad would happen.

The unique thing about gold is that it doesn't get used up. The main way we consume the yellow metal is by storing it, say in vaults or by wearing it as jewellery. Compared to how we use an industrial metal like copper, this sort of usage is very safe. Copper parts in machinery, for instance, are dissipated by abrasion and wear. But gold just sits there, untouched.

Nor does gold depreciate. Unlike most materials, it is almost indestructible. Copper corrodes, steel rusts, wood rots, and concrete crumbles. But a gold coin from 200BC is still perfectly lustrous.

Nor does the yellow metal suffer from technological obsolescence. Gold keeps doing the same thing it has done for thousands of years.

And obviously we don't eat the stuff.

The upshot of all of this is that most of the gold that has ever been mined continues to exist in the form of bullion or jewellery. The World Gold Council estimates this amount to be around 190,000 tonnes. This above-ground stock of the metal dwarfs the amount of new mine production, which runs around 3,000 tonnes per year.

Were this 3,000 tonne trickle were to come to an end, we'd still have plenty of the yellow metal to meet our needs. The existing stock of gold is incredibly flexible and can repurposed into whatever form we want. Gold fillings can be melted down to mint coins, which can be recycled to produce circuits. Circuitry can be melted down to form bars, which can be melted down to gold fillings.

There is a second reason why we don't need new gold mining.

Our demand for most things is defined in terms of physical units. But our demand for gold is expressed in terms of dollars.

Let me give a better explanation for this difference. To make breakfast for a gathering of friends, say that I plan to get a package of bacon, a dozen eggs, and a litre of milk. When I arrive at the grocery store I discover that the price for these things is higher than I thought, and I can't afford everything on my list. It'll be a disappointing breakfast for my friends.

But our demand for gold is different. Say that I want to buy some gold to hold in a vault. It doesn't make a difference to me that the price is higher than I planned for. If the price is $1000, I'll buy 1 ounce. If it's risen to $2000 I'll buy 0.5 ounces. Either way I end up with the same $1000 worth of gold. I'm perfectly indifferent between these two states of the world.

Because our demand for gold is expressed in dollar amounts, there can never be a shortage of the stuff. If everyone on earth suddenly wakes up wanting to own twice as much gold as before, the price of gold can rise to whatever price is necessary to meet that demand. Not so with pork, or eggs, or milk. If everyone suddenly wants to eat twice as much pork, there's nothing that can satisfy that demand except for a huge ramp up in production.

So to reiterate, society doesn't need new gold mining. The stuff is virtually indestructible, and any increase in demand can be instantly satisfied by a rise in price, not new production.

By the way, I'd have a different opinion on this if we were still on a gold monetary standard. Under a gold standard, all goods & services prices are defined in terms of a fixed amount of gold. A steady stream of new gold from mines would help mute large fluctuations in the demand for gold, thus making the general level of consumer prices more stable. But ever since 1968 the gold link has been severed.

Now let's go onto the next issue. If we don't need new gold, why not just ban gold mining?

Mining imposes many costs on society. To begin with, mining is incredibly invasive. Pascua Lama, a project in Chile that has yet to be developed, originally envisioned relocating entire glaciers to get at the underlying gold formations. The Donlin mine in Alaska involves moving "one mountain to another." Clearing out mine sites destroy forests, wetlands, and displaces wildlife.

When rock is processed to retrieve gold, the uneconomic reminder, known as the tailings, must be kept in large reservoirs know as tailings ponds. Dangerous chemicals like cyanide are required to "leach" gold from rock. The breach of a tailings pond at the Mt Polley gold mine in British Columbia, Canada led to 10 million cubic metres of water contaminated with arsenic leaking into nearby lakes and rivers.

Mt Polley before and after a tailings pond leak

Earthworks, an environmental group, calculates that 20 tons of toxic waste are produced for every 0.333-ounce gold ring.

Finally, the various steps in producing gold—mining, milling, and smelting—create large amounts of green house gases, as does the associated usage of electricity to power a mine.

So to review, banning gold production won't hurt society—we don't need more of the stuff. To boot, we'd be getting rid of an activity that hurts the environment.

Yes, there are drawbacks to a ban, too.

Mining provides employment. Not only would a ban destroy the livelihood of hundreds of thousands of miners. All the related businesses that depend on the local gold mine, say restaurants or local retailers, would collapse.

To avoid mass dislocation, gold mining bans would have be carried out slowly. Perhaps existing mines could be allowed to operate until they are no longer economic, but gold prospecting, mine extensions, and new mines banned. These policies would have to be accompanied by large budgets for relocation and retraining. 

But even if dislocation could be managed, there remain other problems.

The production of drugs like cocaine is illegal, but this hasn't stopped drugs from being produced. The same applies to gold. No doubt production bans would be fairly effective in countries like Canada and the U.S., but what about Kazakhstan and Guyana?

Unfortunately, a ban on global gold mining could end up replacing relatively safe and clean gold production with dirtier and more dangerous types of mining. A small but significant chunk of global production is carried out by artisanal miners; family-run outfits that do not exist in an official capacity. Because they are small and agile, artisanal producers could more easily evade a ban. Unfortunately these black market producers are not likely to be held to the same environmental and labour standards as large multinational miners.

A ban on gold production could create a more general problem. Historically we have banned products that are dangerous to consumers, say like drugs. But in this case we'd be introducing a new set of criteria for instituting a ban; because some product serves no purpose. Maybe you think that zombie movies are a waste. Does that mean we should stop allowing producers to make new zombie films? This may be a door we don't wish to open.

So as you can see, banning gold production has merits and warts.

But in theory, the idea makes a lot of sense. If something is indestructible, and we only want dollar amounts of it rather than ounces, why the devil are we wasting time and resources producing it? As long as we can successfully shift gold mining communities to other forms of employment, then ending global gold production makes a lot of sense.

Tuesday, August 18, 2020

Bitcoin is an account, not a token

When economists talk about payments, they often make a distinction between token-based and account-based payment systems. In a recent post at the New York Fed's Liberty Street blog, Rodney Garratt & cowriters argue that new payments technologies like bitcoin and central bank digital currency may not fit into these traditional categories. Perhaps it's time for a reorg?

In an account-based system, some sort of database stores account information. For a payment to occur across this database the payer needs to prove that they own a spot in that database, and that this spot has sufficient funds.

With a token-based system there is no database. Instead, objects are used to pay (say banknotes or gold coins). The key feature of a token-based system is that the recipient must verify that the object is valid and not counterfeit.

In short, tokens involve identifying the object. Accounts involve identifying the individual.

Garratt et al argue that a digital currency such as bitcoin is a mix of the two, token and account. Bitcoin is an account-based system because some sort of "proof of identity," specifically a private key, is needed to transact. This puts bitcoin in the same category as a bank account, which also has a process for verifying the identity of users. (Instead of public key cryptography, bank customers must go through a due diligence process, and after that must produce a PIN.)

I think Garratt et al are right about bitcoin being an account-based system. But I don't think that bitcoin also qualifies as a token-based system.

Not a token

A token is an object. And objects can be counterfeited. That's why when Alice pays Bob a $20 note, Bob is responsible for carefully checking it.

Account-based systems don't suffer from counterfeiting problems. It's impossible for a Alice to pay Bob with fake dollars from her Wells Fargo account. Wells Fargo dollars aren't independent objects in the same way that banknotes are. They are entries in a well-secured database. The same goes for bitcoins. There is no way for Alice to make a fake bitcoin entry and send it to Bob.

In addition to counterfeitability, I'd argue that another key feature of a token is that it can be lost by one person, found by another, and then reused for payment. If Bob loses either his $20 banknote or his gold Krugerrand, Alice can find them and spend them.

But account-based systems don't have this feature. Wells Fargo dollars are database entries. It'd be impossible for Alice to lose a Wells Fargo database entry or Bob to find one and reuse it. And the same goes for bitcoins. Alice can't lose 0.2234 bitcoins, nor can Bob find Alice's 0.2234 bitcoins.

Gift cards: account or token?

What about gift cards, say like those iTunes cards that they sell at pharmacies?

Because gift cards come in physical form they may seem like tokens. But the card is just one element in an account-based system. A $20 iTunes balance exists in a database run by the gift card system operator, Apple. Whereas a $20 banknote can be used without the authorization of the central bank, an iTunes card has to be granted permission before it can initiate purchase. If the database entry to which an iTunes card is linked is empty, then a payment will be denied.

Like other account-based systems, counterfeiting isn't a problem with gift cards. Database entries can't be produced with an inkjet printer. (Sure, an iTunes card can be lost and reused by the finder. And so it seems like it should be classified as a token. But that's only because the "key word" or "password" is literally baked onto the card. Losing an iTunes card is like losing one's private bitcoin key.)

Bitcoins and gift cards can be turned into tokens

There is an interesting situation in which gift cards or bitcoins can be converted from account-based systems into token-based systems.

Instead of using her $20 banknote to pay Bob, Alice may choose to pay Bob by passing him a $20 iTunes card. Or maybe she gives him a physical Opendime hardware wallet that holds $20 in bitcoin. Bob could in turn pass these devices on to Charlie, to whom he owes $20, and Charlie on to Dave, etc.

In both cases, accounts are being repurposed as tokens. What is happening is the physical object, or key, that provides access to underlying bitcoin or gift card database entries is being moved from one person to another. But the actual database entries to which the Opendime and iTunes card are linked are not being updated.

Bob will have to treat both of these proffered instruments like cash. He'd be worried about counterfeits, and would want to verify that neither the gift card nor the Opendime device are fake.

Let's vizualize all of this into a table:

The table suggests that bitcoin exists in a totally different category than a banknote (unless those bitcoins are embodied in the form of an Opendime device). You'll also notice that the table differentiates between open accounts and closed ones. Let's get into that next...

Bitcoin and Wells Fargo are different types of accounts-based systems

If bitcoin qualifies as an account-based system, it is certainly different from Wells Fargo's system. I'd argue that Wells Fargo operates a "closed account" system while Bitcoin functions as an "open account" system. The key feature of an open system is that anyone can get an account. As Jerry Brito would put it, they are permissionless. A passport or driver's license isn't required to gain access, so even fraudsters can climb aboard.

Gift cards, non-reloadable prepaid cards (like Vanilla cards), and e-gold (a 1990-2000s era payments system that didn't require people to use real names) are also open account systems. Anyone can self-initiate account opening and start to make payments. No one gets kicked off. (We could further differentiate between centralized and decentralized open account systems, with gift cards being the former and bitcoin the latter.)

What qualifies something as a closed account system is that not just anyone can get access. Want to use a Wells Fargo account for making payments? You'll have to make it through Wells Fargo's application process, which involves giving up plenty of personal information. And only when a Wells Fargo employee has opened an account can payments be made.

Turning bitcoins, gold, and cash into closed account systems

Incidentally, it's also possible to convert tokens and open accounts into closed accounts. The London Bullion Market Association (LBMA) manages a walled garden of gold bars. Each bar is carefully vetted prior to entry into the system, and once in the system all movements are tracked by monitoring serial numbers. LMBA gold has ceased to be a token and has become a closed account. (I recently wrote about the LBMA system for Coindesk.)

Central bans could do the same with cash. They could set up an online bank note registry, and all cash users would be required to sign-up and record the serial numbers of note each time they receive a new one.

As for bitcoin, many people already keep their bitcoin in at regulated services like Coinbase that require customer ID. One day it may only be possible to send bitcoins from one regulated service to another, in which case a big chunk of the Bitcoin network will have become a closed account system, not an open one.

Why does all this matter?

Specialists need to be able to have conversations about their subject matter. Categorization is one of the ways to make these conversations flow without chaos. Are the categories we use for conversing on the topic of the economics of payments still sufficient? The world has changed. We've got new entrants like bitcoin and central bank digital currency. Garratt et al suggest that aging categories can "slow down progress in understanding intrinsic differences between the growing set of digital payment options and technologies."

But this case I think the old taxonomy are still useful, albeit with some tweaks.

P.S. Lawyers, regulators, users, and developers will all have different taxonomies for payments. Inter-taxonomic conversations are difficult, since a single term can mean something totally different. So stay within your taxonomy to avoid confusion.