Thursday, October 28, 2021

Does it make a difference if Tether lends out new USDt?

I recently tweeted something about the world's largest stablecoin, Tether. It gives me an opportunity to ask a broader question about money in general:

Tether issues USDt, which are U.S. dollar-denominated IOUs redeemable for actual dollars at $1. Unlike a PayPal IOU, a Tether IOU exists on a blockchain.

What the tweet (and linked-to article) is saying, in short, is that Tether has misadvertised itself. Tether says in its terms of service that it only creates new stablecoin tokens, USDt, in acceptance for money. That is, to get $1 worth of USDt from Tether, you need to send it $1 in actual U.S. dollars. But in reality, Tether does not seem to be waiting for deposits to roll in before issuing new USDt. As the FT's Kadhim Shubber reports, it is directly lending new USDt out, much like how a bank puts new dollar IOUs into circulation by lending them out.

I want to use Tether to ask a more general question about the economics of money creation. Granted, Tether is not issuing stablecoins according to its terms of service. But does it really make an economic difference whether Tether lends out new USDt stablecoins or if it only creates them when someone deposits U.S. dollars?

We could also ask the same about PayPal. PayPal only creates new PayPal dollars when someone transfers U.S. dollars to PayPal. But what if PayPal were to start creating new PayPal dollars by lending them into existence? Would the monetary economics of the PayPal change?

I'd argue that it doesn't. But I'd be interested in hearing the other side of the debate. I'll show why the method of issuing Tethers or PayPals doesn't really matter using two quick examples.

Let me quickly outline one assumption I'll be using. The market has a certain demand to hold USDt, and if that demand is exceeded by new issuance of USDt, the excess USDt will be quickly sent back to Tether for redemption at $1. That is, given the existence of a $1 peg, a stablecoin issuer can never exceed the market's demand for a stablecoin.

Say that Tether has $100 in USDt outstanding and also has $100 sitting in a bank account as backing. 

Under the first scenario, one that is consistent with Tether's terms of service, John arrives and deposits $10 with Tether and gets $10 USDt. Now there is $110 USDt outstanding. There is also $110 sitting in Tether's bank account. Next, Tether lends $10 of the $110 in its bank account to Sally at 6% per year. It asks for collateral to protect itself. That is, Tether requires Sally to pledge $15 worth of bitcoin as security.

Now for the second scenario, the one described in Kadhim Shubber's FT article. Tether prints $10 worth of new USDt out of nothing and lends it directly to Sally at 6%. Tether asks Sally to pledge $15 worth of bitcoin as collateral. There are now $110 USDt in circulation. If Tether were to overissue by lending more USDt than the market wants to hold, that amount would quickly reflux back to Tether for redemption at $1. (So if Tether lends Sally $15 USDt but the market only wants $10 USDt, then $5 USDt would quickly be brought back to Tether for redemption. Tether adjusts by reducing its exposure to Sally by $5.)

Under both methods of issuance, we end up at the exact same spot. There are $110 USDt in circulation. Backing that amount, Tether has $100 in cash in a bank account and $10 worth of a 6% loan secured by bitcoin.

So economically speaking, it doesn't matter whether Tether lends out new USDt or if it creates new USDt upon reception of actual dollars. Either way, $10 worth of new USDt will go into circulation. And either way, that issuance will be backed by a 6% loan to Sally collateralized by $15 worth of bitcoin.

The interesting thing is that even though there is no difference between the two scenarios, our language and law distinguishes between Tether 1 and Tether 2. In the first scenario, Tether is considered a fintech, a money services business, or a payments company, and thus subject to a certain set of laws. In the second scenario it is a bank, or a depository, and thus subject to an entirely different set of laws.

But if the two Tethers have the same economic function, why don't we the use the same language and set of laws & regulations for each?


  1. I do not know enough about stablecoins, but for an ordinary bank there is a difference between the two scenarios, as the funding risk is NOT the same.

    Let's take the case of the Tether for USD deposit first.

    This (primary) deposit has stable funding properties, the treasurer at a bank (Tether) can assume that somebody that has just exchanged USD for Tether, is unlikely to withdraw anytime soon (why pay in, if you want your money back immediately), i.e. the behavioural maturity of the deposit is longer than “on demand”. For the treasurer, this means that the deposit can be lent out without impairing the desired liquidity position of the issuer.

    Now, the case is different, when you lend out money BEFORE you have the funding, as in the deposit creation example. Think about how long somebody, who asks you for a interest paying loan, is going to keep said deposit with you. Not very long as he intends to do something with the deposit (he has to earn the interest on the loan). For all fits and purposes the Treasurer of the bank (Tether) has to assume a behavioral maturity of the deposit is very short-term = “on demand”. If he is already operating at the optimum, i.e. desired reserve ratio, he knows he will have to look for a stable funding source, as the deposit created is flying out of the bank soon…

    For details re-read this article by Tobin.
    d0159.pdf (

    Conclusion: while from an accounting perspective (and in monetary aggregates) a deposit is indeed a deposit, in practice they are not. For that reason banks (Tether?) employ mathematical models to track the payment behaviour of individual deposit accounts in order to better match liquidity of assets and deposits…

    1. "This (primary) deposit has stable funding properties, the treasurer at a bank (Tether) can assume that somebody that has just exchanged USD for Tether, is unlikely to withdraw anytime soon."

      Is that a fair assumption, though? Maybe the person who deposited USD for USDt doesn't want to sit on that USDt. They quickly send it to an exchange to buy a cryptocurrency, and the exchange automatically returns the USDt back to Tether for redemption into US dollars.

      And maybe Sally, the person who borrows USDt, spends them into circulation and they end up in the wallet of saver who lives in Venezuela and has a long-term demand for USDt as an inflation hedge.

      In other words, I'm not convinced that the means of issuing a deposit determines how stable they are for funding purposes.

    2. It is not lending in the usual use of the word. It is more like a swap with a time limit, although it is more complicated than that, with the interest, considering the whole balance sheet.

    3. As I said, I do not know enough about the stablecoin system, in order to give a definite answer. As long as you accept that for banks, where your accounting argument applies all the same, there is indeed a difference (reflected in actual banking practice as well as Basel regulations), I am fine…
      Now back to stablecoins. I read that 70 percent of crypto transactions are done using Tether and that deposits < USD 100k cannot withdraw. This information, if correct, already can serve us to make crude statements about your example:
      1. Somebody who has just deposited Tether in order to gamble in the cryptomarkets, is a very low liquidity risk for Tether, as Tether enjoys a near monopoly position for Crypto transactions, i.e. 70 Percent of the transactions are merely reshuffling of ledger entries. The risk is almost non-existent, if the newly deposited amount is < 100 k.
      Conclusion: The guy in your example does not add substantial liquidity risk to the system as the desire of the guy on the exchange to withdraw (for USD) is unrelated to the fact that a new deposit has been placed into the system (it just could mean his price is marginally better).
      2. This is not necessarily the case for the margin loan. The margin loan is likely to be taken out by somebody who is a large (> 100k) account holder, as he is on average more sophisticated and likely sitting on huge crypto gains. Taking out the margin loan is a nice way for the insiders/whales to take out liquidity without crushing the price while keeping nice optionality (these are non-recourse loans). If my assumptions are correct then crypto margin loans above USD 100 k are a huge liquidity risk for Tether.
      So Sally, cannot spend tether “into circulation” (tether is not a medium of exchange in the real world, is it?) but either continues to gamble on crypto and end up in Venezuela (fine for Tether as these is merely a reshuffling of ledger entries) or she takes out USD immediately, since she wants to cash in on her gains so far…
      Conclusion: I am not convinced that all deposits for Tether are created equal either…

  2. PayPal would never created an account out of thin air for someone without first taking in real dollars. Why would they? Now they owe real dollars to the account holder whenever they want to transfer it to their bank, which they do not have.

    Also, Paypal's accounts are yes IOUS, but they represent real digital dollars, not some stablecoin, like usdt.

    Tether is exactly like the Federal Reserve, they can print without limit. But they are only printing more tokens, not dollars. That is why the government doesn't care.

    Any collateral which Tether may or may not receive for new usdt lent directly does not give value to the tokens that are printed out of thin air. It gives some assurance that the loan will be back with interest- it is value to the loan.

    Ask some bitcoin maxi's what is the difference between what Tether does and what the Federal Reserve does--- I really want to know why they don't like what the Fed does... but they love Tether.

    1. "PayPal would never created an account out of thin air for someone without first taking in real dollars. Why would they?"

      I'm asking you to think hypothetically.

    2. It seems this already occurs operationally with Paypal credit. Pay pal Credit has been running for 5 years.
      Person A borrows on PP credit which is run by a credit card company / Bank . Person A spends the credit with Person B who has a Paypal account. New Paypal balance has been created and Person B can now withdraw dollars.

    3. IIRC that sort of lending is being done with a partner bank. PayPal isn't actually lending out new PayPal dollars. The bank is.

    4. Yes. I did identify that in my comment. I would have thought the characteristic of the transaction that would interest you from the context of "moneyness" is that new Paypal balances are being created in return for debt contracts.

    5. idk any bitcoin maxis that disapprove of what the FED does... We basically just acknowledge that they will keep doing it because the system depends on it. FED injecting liquidity in the system is not only vital for Bitcoin, its what makes it a better purchase than a US treasury over a 4-10 yr period.

  3. I'm not sure I follow your second scenario example. If the demand for USDT is less than what has been issued then redemptions are made in USD, not BTC. So, for example, Tether starts with 100 USDT issued and 100 USD backed, then issues 100 USDT and receives 150 BTC in collateral, someone redeems 100 USDT for 100 USD. Tether now has no USD collateral and is left with only BTC that cannot be easily converted to USD. Tether must break its peg and likely collapses.

    1. Sure, but the same thing would happen if Tether tried that under my first scenario.

      Using your numbers, Tether issues 100 extra USDTs for 100 USD, then lends the 100 USD it receives to Sally (taking 150 BTC as collateral). The demand for USDT suddenly falls by 100, which means that someone takes 100 USDT into redemption. This exhausts all of Tether's USD assets, leaving it with only BTC.

      So it doesn't matter how Tether issues USDT. Either way we get to the same endpoint.

    2. Sure, we do get to the same endpoint....eventually...
      Since the world is never in equilibrium and dynamics do matter (who redeems when and where and how) there is a difference between scenario 1 and 2, as outlined in my hypothetical examples above..

      Just because something is a demand deposit, doesn't mean I cannot safely lend against it, if it meets certain criteria. One demand deposit is drawn down every month (say the wage fund of a company) the other stays there forever as precautionary savings...
      Eventually both demand deposits are going to be withdrawn, but in the real world there is a huge difference...(dynamic vs equilibrium economics)

  4. Just a few quick points, if I may.

    1. PayPal does not create “PayPal dollars”. In fact, it creates no money. Therefore, it cannot lend out anything other than the money it receives (or acquires). As opposed to PayPal, a normal bank can. Hence the significant differences between the two regulatory environments.

    2. Point 1 above should explain the difference in your two examples. But just to make sure: under ex.1 Tether lends out 10 USD, currency it received and it doesn’t control; under ex.2 it lends out 10 USDT, currency it controls, i.e. which it could create out of thin air and hitherto without any limit.

    3. On to your main question: if the 1$ peg would be effective, you’d be right: there would be no economic difference. But what makes you think it’s the case? Based on what I’ve seen so far, in all likelihood a sizeable chunk of USDT is not even collateralized. Have you seen Tether’s full balance sheet? Cause I haven’t. Short of full transparency, the only event which will put this to the test will be a run on USDT.

  5. I am not sure what "economically speaking" means but the risk to Tether holders is obviously different in these two scenarios. If a large and quick drop in Bitcoin value occurred, Tether would remain liquidity run resilient in the first scenario and not resilient in the second scenario. This resilience difference would have an impact on the likelihood of such run - it is rational to try to be the first out of the door with your USD in case if a 90% reserve bank and not rational if the bank is 100% reserved. This difference would have an impact on the willingness of the public to hold Tether even prior to a Bitcoin drop. So the two situations are obviously different from risk point view and will result in different current demand for Tether.

  6. In the US Tether is considered a commodity or futures trading platform because on October 15, 2021 the Commodities Futures Trading Commission fined Tether Holdings $41 million for making untrue or misleading statements about the asset backing for tether tokens. For a 26 month period beginning June 1, 2016 Tether held sufficient reserves only 28% of the time and never completed an independent audit. Tether is not a fintech or a bank but a commodities/futures (?) scam. See the following link: