That’s the issue right there. How does Tether save its dollars? We can see it in their transparency report[1]. Whether you believe them or not it’s not just cash in a bank account.

* 0.41% Non-U.S. Treasury Bills

* 55.53% U.S. Treasury Bills

* 0.15% Reverse Repurchase Agreements

* 5.81% Cash & Bank Deposits

* 9.63% Money Market Funds

* 28.47% Commercial Paper and Certificates of Deposit

How much of that is liquid and directly convertible to dollars 1:1 in he next 24 hours? Not 100%.

What happens when they start selling billions in Treasury Bills and Commercial Paper to fund redemptions? The market price of those assets will drop.

What if the value of those assets is already below 1:1 because of recent market events?

What if they’re not being as transparent as they say they are?

> As long as they never spend anything from the reserve, this can't fail no matter how unpopular the currency is.

This can easily fail many different ways.

[1]: https://tether.to/en/transparency/#reports

The statistics you're bringing up are as of March 31. Do note that 6% of reserves are in "Other Investments (including digital tokens)", and Bitcoin (as a proxy for all cryptocurrencies) is down ~30% since then, so that's at least 2% of their assets that have been wiped out by market conditions. Keep in mind that said report also said that, as of March 31, liabilities are 99.8% of assets, so Tether's accounts says it should already be underwater.

(Although, if I'm reading the attestation correctly, all of the assets--including cryptocurrencies--are actually valued at purchase cost and not fair market value, so what the actual present value of those cryptocurrences is now or was 2 months ago is extremely unclear. Transparent is the opposite of how one could describe Tether's financials.)

Also, if 2% of outstanding tether has been lost (forgotten wallet keys etc) then those can never be redeemed and again, tether wins.

Inflation is another factor worth considering here: tethers deposits are deminishing but it's investments are (or should be) shielded.

I think people fail to notice how similar a (non-fraud) tether model is to a traditional bank: you take short term deposits, you make long term loans, and you hope to have enough capital on hand to deal with any runs. Given the liquidity of modern capital markets, it's very rare for the fed to have to bail out small deposit banks. So it's reasonable to assume the same will apply to tether.

A quarter of their investments are commercial paper, which hasn't averaged as high as 2% yield since a brief period in March 2020. Actual cash of course has 0% yield. US Treasuries (sub 1-year), which make up nearly half their assets, also hasn't hit 2% yield any time recently. So no, they aren't recouping their loss on cryptocurrency.

> I think people fail to notice how similar a (non-fraud) tether model is to a traditional bank: you take short term deposits, you make long term loans, and you hope to have enough capital on hand to deal with any runs. Given the liquidity of modern capital markets, it's very rare for the fed to have to bail out small deposit banks. So it's reasonable to assume the same will apply to tether.

One of the reasons why banks rarely have to be bailed out is because there are stringent regulations on bank holdings. For example, a minimum tier 1 capital ratio, the amount of equity that needs to be held to cover unexpected asset shortfalls. This requirement is I believe 10%, and based on the evidence Tether has produced, Tether's tier 1 capital ratio is... 0%. It should also be noted that Tether is perilously close to insolvent, with (claimed) assets about 100-101% of total liabilities; most financial institutions prefer to be at least ~110-115% of total liabilities.

Compare Tether to banks if you want to, just be aware that it just makes Tether's financials look even worse in comparison.

I'm just laying out the maths...

And to be clear, they only need to make 2% total to cover their crypto loses. If the average tether coin exists for 18 months before being redeemed, 1.5% per annum will net them 2.2% over that period and they're golden.

That extra 0.2%, for a $10bn withdrawal is 2million USD in profit right? Not bad split equally between 5 employees, for a month with massive crypto loses and 10bn in net withdrawals...

That’s exactly what’s unethical about it. They’re operating a bank, but have skipped all the regulations and oversight that banks operate with.

I have no issue with Tether operating a fractional reserve deposit system, if they are subject to the same oversight (and insurance) that banks are subject to.

But tether has no risk of contagion to a bank does it?

And markets have never been more stable or deep or liquid.

So the case for regulation here is weak.

Again. I don't actually know if tether is a giant fraud, or how much actual business case there is here for stable coins. I'm just saying, it's sort of easy to make a case at least that they're fine.

Made, realised and not spent/withdrawn by them. The required assumption here is that the pool of money behind tether grows / is reinvested. But given they only need to keep 1:1 (assuming even that is true) the investment profits may have been exchanged for hookers and blow for all we know.

Tip of the iceberg https://bitinfocharts.com/bitcoin/wallet/Bitfinex-coldwallet

Also, you're going with the assumption they shall be able to redeem 100%. Crash happens, like we have seen, but everyone cashing out their USDT is not a scenario going to happen. Or if you want to account for this scenario, then you can as well assume that crypto is going to disappear, and that would not happen without a cataclysmic event in the stock market either. Probably we will be back to the stone age at this point, and will have other things to worry about

This seems wildly optimistic. All it would take is for users to adopt some new FOTM stablecoin faster than Tether backers can liquidate their reserves. It needn't be rational, either; it could be catalyzed by, let's say, a *ism scandal involving someone connected to Tether.

Didn't bankers say something similar in 1928?

My long-term treasuries are down well over 10% this YTD, in case anyone wants to know. So if Tether had say, $50-billion in 10-to-30Y treasuries at the start of the year, they only have $45-billion of that now.

There are serious market risks when you buy/sell Treasuries. Yes, they're among the safest instruments on the market, but rising interest rates and inflation are huge issues and absolutely wreck the value of long-term treasuries.

They don't. As stated at the link:

> U.S. treasury bills comprises U.S. treasury bills with a maturity of less than 120 days.

My overall point is that USDT could very well be buying up dollar-backed securities, such as 30-year treasuries, and yet still lose a ton of money if the market moves under them. Unless Tether allows 3rd party audits of their reserves, I don't think its necessarily safe to assume that they actually hold those reserves.

Lastly: a temporary liquidity crisis that drives USDT down far below the peg is something that would be incredibly profitable for the operators. With perfect information, it would be a situation of trading 50 cents for dollars, and could be used to erase a partial deficit overnight.

This is nuts - they dont and it wouldnt make any sense. You cant have a short term cash-equivalent backed with long-duration bonds. It would be a total asset-liability mismatch.

For reference:

T-bonds mature in 20 or 30 years and offer the highest interest payments bi-annually.

T-notes mature anywhere between two and 10 years, with bi-annual interest payments, but lower yields.

T-bills have the shortest maturity terms—from four weeks to one year.

Nice.

Edit: Just providing context for those who may be interested.

It also works because the US dollar has been remarkably stable and most of their holdings are USD. No crypto is so stable, with a bunch of them being about the most volatile assets you can lose money with.

How is that relevant to the claim in question? (Which, if you’ll recall, was whether tether can maintain the peg and redemptions while holding the same assets as MMMFs, which generally do that just fine.)

> It also works because the US dollar has been remarkably stable and most of their holdings are USD.

Okay, now you lost me, and I’m not convinced you have the recent discussion in mind. The original comment was claiming that Tether can’t maintain the peg, because it holds non-dollar assets. I pointed out a trillion dollar industry by that maintains a peg, using those same assets, and you’re saying the non-dollar assets only succeed there because the dollar is stable? Which is somehow an argument about how these assets are good enough for MMMFs to work but not Tether?

Please take a minute to review the thread and see if you’re still supporting the claim I disputed.

Among other things the financial system is a web of trust.

IMO one of the fundamental things that crypto gets wrong is replacing trust with algorithms. I do not think that can be done. Trust is about people.

Time will tell if an algorithm that can automate trust can be found. I do not expect it will

Bitcoin is an example of a system of trust that has worked pretty well so far (although it requires a lot of electricity, but that is the trade off). There are also people on the Bitcoin core team, so there is some trusted element there.

Crypto will likely continue to innovate on algorithms, given the chance.

I think there can be trust in people, plus algorithms, with algorithms taking over more over time. This is already happening even in traditional finance, i.e. giving more control over to algorithms that participate in HFT. People do monitor those, but people monitor crypto, too, and maybe the failures in crypto so far mean too much control has been given.

I would argue that some more things in finance can be automated, without things being so black and white (i.e. no control vs total control given to algorithms). I do think the trust model given to governments and traditional finance gatekeepers can be iterated on, with some regulation involved too. I don't think we've figured everything out yet.

They have 39B in US Treasury Bills, how much do you think these will drop if they sell? The truth is next to nothing. and a 39B moat for sell offs seems very reasonable

Buy Tethers with what? If the money is in fact saved in regulated banks or other such instruments, there's nothing liquid left to defend the peg on the exchange.

Instead, the standard "backed stablecoin" approach is to make money with a small spread on redemptions/creation, while allowing others to do that hard work. Tether takes (https://tether.to/es/fees) a 0.1% spread on redemptions, so if you hand them $1mUSDT you'll get back $999k USD.

> But they should be solid as long as they have made more money on those investments than they've spent in salaries and yachts and what-not, and Tether was already huge back when BTC was below $2k.

One possible "bank run" scenario is that their backing is in fact stable, but it is illiquid. Suppose Tether invested part of its reserve in long-term loans to another company (like Binance). If Tether ever faces a crisis of confidence, it would face large-scale redemption requests, but it may be unable to call in its loans to fund those requests. That would leave Tether unable to redeem its currency, and a public suspension of redemptions would drive a further exodus.

In theory, every single USDT in circulation could be redeemed at a moment's notice. Unwinding some $73 billion in investments would be a Herculean feat even if everything is fully legitimate.

The easy way to protect against this is to not contractually promise instant redemption. I think this is what tether actually does, but I could not find a source. Regular savings account banks do typically do this, for example the bank has the right to ask for 7 days to honor a withdrawal.

If you are tether and your asserts are in bonds that mature in under N days, then you could just promise redemption within N days to eliminate bank run risk. You would still of course have counter party risk that when the bond matures it is not paid back.

Source?

I'm familiar with savings accounts described as "instant access" or "easy access" where you can get your money out whenever you feel like it.

Unless the bank actually markets an account as a "notice account", can they really ask for 7 days notice?

> Advance Notice of Withdrawal: Under federal law, we must reserve the right to require you to give us at least 7 days written notice before you take money out of your 360 Savings. (This hardly ever happens but legally we have to say it!)

Interesting, and this appears to be federally-mandated. Wow.

Are there other jurisdictions where this kind of rule exists?

https://www.moneysavingexpert.com/savings/savings-accounts-b...

I'm sure you didn't mean it that way, but that sounds rather like an excuse for a bank to hold on to someone's money. Withdrawals don't have to be in cash, banks can issue cheques or money orders (always assuming they still exist in your jurisdiction!) If not the bank can simply pay out your balance by electronic transfer to an account you specify.

[1] https://www.investopedia.com/terms/n/notice-of-withdrawal.as...

I don't think that's reasonable for Tether, at least not with a bulletproof N. I can't quickly find any specific redemption guarantee, but their March reserves report notes that their US Treasuries (the largest single claimed category) can have maturities up to 120 days. The commercial paper category claims an average duration of 44 days.

I acknowledge the complexities of investing that collateral in more or less liquid instruments, but in principle the notion appears sound. I'd be happy to give you an interest free IOU in any amount you like if you provide me cash money for that nominal value in exchange.

That is, IMHO, a questionable assumption.

Now, "fraud" in this case too strong a term. The intent to defraud may not be explicit in Tether's internal conversations and practices.

All business ventures by definition involve risk, and corporate structures and venues and liability shields and so forth exist precisely to create a space for risky ventures to be undertaken without failure leading to death or personal poverty for the principals. In those cases, the customers, investors, and other participants are also partaking in the risk, based on information that is deliberately- sometimes responsibly, sometimes incidentally, and sometimes maliciously- incomplete.

At the end of the day, when the full accounting is known, a post facto judgement of fraud vs speculation may be rendered.

But the assumption that Tether is behaving as one might want one's bank to be behaving, in terms of customer/user risk exposure, should be strongly, strongly questioned.

The result was the same “slow up fast down” sawtooth we see in the real market, and investments that had real alpha were crowded out by investments that simply pushed risk into the tails. It took like an hour to assemble that agent-based simulation and it scared me out of finance. Most financial engineering seems to just be Martingale betting schemes that guarantee small to modest returns 99% of the time.

Also imagine that they took the actual money that was given to them and, instead of keeping them in a bank account, said they were going to invest them. (Or did not say it, and did it anyways). Imagine they invested it in amazingly profitable sectors like the housing market in China which is definitely not crashing right now.

Tether isn't backed.

0. https://www.cnbc.com/2021/02/23/tether-bitfinex-reach-settle...

I think this is the only way a "stablecoin" can function as designed... but it is not possible to construct an entity that has any scalable incentive to provide the backing that would create such a coin. For that entity, it is nothing but downside.

Therefore, stable coins are a fiction on par with perpetual motion machines. In the short term there's all sorts of perpetual motion and over unity machines... in the long term, not so much.

Backing a currency isn't about taking deposits from people. You can't back with deposits because your liabilities = assets then. You have to put your own stuff up, but then all you are doing is running the risk that people will redeem it away from you. Backing a currency is basically just giving your stuff away with extra steps.

And if people redeem from you, just add a clause where if the liquidity pool is gone you have to wait X days before getting it redeemed.

It’s like panicking because you’re because your bank only has $30k cash onsite, or your money market fund is only 3% cash.

Disclaimer: I’m still slightly short Tether, just think this a bad reason to be. Search my history for the details.

https://ag.ny.gov/press-release/2021/attorney-general-james-...

That doesn't necessarily mean that Tether's assets are less than its debts. I don't know the details. I just know that the NY Attorney General finds their activities to be shady.

In the case of Tether, the company represented that each of its stablecoins were backed one-to-one by U.S. dollars in reserve. However, an investigation by the Office of the Attorney General (OAG) found that iFinex — the operator of Bitfinex — and Tether made false statements about the backing of the “tether” stablecoin, and about the movement of hundreds of millions of dollars between the two companies to cover up the truth about massive losses by Bitfinex.

also

Tether’s claims that its virtual currency was fully backed by U.S. dollars at all times was a lie.

The settlement also bars Tether from trading with New Yorkers, and the must submit regular reports about their compliance with this prohibition.

I suspect the AG got what they wanted, they just managed to do so without taking time in the court system. I would assume that Tether realized that it wasn't worth trying to defend themselves and so agreed to settle. My use of "successfully prosecuted" might not have been the right way to phrase it.

Reading between the lines - there was a period in the past where Tether was not fully backed. If Tether was still, as of today, not fully backed, would NYAG conclude this case?

I suspect the AG thought it had a bigger case involving ongoing fraud but settled as it could not find evidence of any ongoing false claims.

While biased in Tether's favor, worth a read: https://tether.to/en/stress-tests-resiliency-and-bank-runs/

"Regulators and economists have long questioned whether Tether has enough assets in its reserves to justify its stablecoin’s purported peg to the dollar.

“USDT is, quite simply, fully backed by collateral,” Tether said in a statement Monday.

“It has maintained its peg because every USDT is redeemable for dollars via Tether, and as such any time the price goes below $1 investors can earn a profit by buying USDT for a discount and redeeming it with Tether.”

The company previously claimed tether was backed one-to-one by dollars in a bank account, but subsequently revealed it was using other assets including commercial paper — short-term corporate debt — and even digital tokens as collateral after a settlement with the New York attorney general.

Last week, Tether said it reduced the amount of commercial paper it owns and increased its holdings of U.S. Treasury bills. For the first time, the British Virgin Islands-based firm said it also holds some foreign government debt. Tether declined to comment further on the source of its funds, but said it is pursuing a more thorough audit of its reserves."

But such a bank, when you show up with your $80bn, in order to protect their ability to reliably offer those kinds of terms, will want to ask you a few questions about where you came by that cash.

And if you can only say ‘I have no idea’, and when they then ask ‘how much of it belongs to sanctioned individuals or is criminal proceeds?’ Your best guess is ‘not none of it’, then the legitimate banks are going to walk away from that conversation.

So then, yes: where are you going to keep that $80bn?

This is not someone you have a direct relationship with. And now those stablecoins ‘belong’ to that new person.

Then that person uses the stablecoins to pay another person for something else, and eventually a completely different person can now come along and go through your KYC process and cash out the coins.

All above board and legitimate.

Except in the middle there’s a part where you might actually be acting as a bank for an international drug cartel or a sanctioned Russian oligarch. You just can’t be sure.

...but isn't the whole point of Tether to stay as far away from the traditional banking system as they can?

This of course includes avoiding - as much as possible - all the KYC/AML legislation that traditional banks are obliged to follow?

The only way I can think of a stablecoin being totally independent of banks is to store their reserves as bank notes, and when someone wants to cash out USDT to USD they have to go pick up the bank notes. This glosses over the obvious challenges of storing billions of dollars in bank notes and distributing those notes during a cash out. The bank notes are also worth less than face value because you have to physically store and secure them.

If one's business is "providing a clean link to the US financial system", why would one choose to avoid financial services regulation?[0]

If one is clean, or at least trying to look clean, why would one choose an accountant based in the Cayman Islands?[1] This bit in particular is beyond parody.

[0] https://www.coindesk.com/policy/2021/01/26/what-tether-means... [1] https://www.coindesk.com/markets/2022/01/26/tethers-new-acco...

Or they can invest in money market funds, or similar short term investments. Or commercial paper (unsecured short term debt). From what I read, they invested a lot in commercial paper to Chinese companies. It's all a game to eek out a slightly higher yield.

They face the same problem as any bank. Except banks have a regulatory framework that tells them exactly how much they can invest and in what, and every decade or so the bank gets bailed out.

[1] Especially in a near-zero interest rate environment.

The fact that tether has never been audited makes the former more likely than not.

This is exactly the case and exactly what most people believe has happened. It's a minor miracle that usdt hasn't evaporated already, it's a matter of time.

So what you're describing does exist, but isn't what USDT is. IMO the infatuation with USDT has always confused me a bit; why would anyone use USDT over USDC in the first place?

I don't think even banks do monthly external audits.

USDT does quarterly external reports: https://tether.to/en/transparency/#reports

(Don't get me wrong, I assume everyone in this space is some combination of fraud, ponzi, and money laundering)

My model is the same as your model, you're just making assumptions I'm not.

Here’s the same example, stated more explicitly:

1. I buy 10 USDT from Tether in exchange for $10 worth of Bitcoin. Tether now has 1:1 reserves of Bitcoin backing USDT.

2. The price of Bitcoin decreases by 50%. Tether no longer has enough reserves to cover all the USDT in circulation.

3. I want to sell my 10 USDT back to Tether and get my $10 worth of Bitcoin. But Tether only has $5 worth of Bitcoin in its reserves! USDT will lose its peg.

Tether acts more like a central bank, they allow people to trade USD for USDT for exactly $1. Trading USD directly can be difficult because there are extensive KYC rules and not everyone has access to a USD bank account.

On some exchanges, there is still a demand to trade in USD and it commands a premium due to the difficulty in trading it. This generates demand for USDT and slightly increase the price, suppose to $1.02 on this exchange.

Arbitrage traders can now buy USDT for $1.00 and sell it on this other exchange for $1.02. The tether corporation simply holds the original $1 (of actual USD). They should have $1 USD for each USDT in circulation, and that's a lot of 'float' that they can invest in safe things like short-term US treasuries to make money off simply holding the cash. No need to speculate on Bitcoin or other risky investments, they are making plenty off just holding cash-equivalents.

That's the issue — they don't have $1 USD for each USDT in circulation.

Tether has seen over 10B in withdrawals over the last couple weeks. The total circulating supply is 73B so that was something like a 12% withdrawal over a very short time window.

Consider that banks are only required to maintain a 5% leverage ratio. 12% is a pretty extreme test, it's enough to cause most banks to fail. It also does not make sense for Tether to take on additional risk. They can make a nice yield by investing the 70B+ in very safe highly liquid short-term US treasuries.

They've admitted to it in court. Multiple times. Patrick Mackenzie (patio11 on here) has a couple good articles about the fraud. [1][2]

> 12% is a pretty extreme test, it's enough to cause most banks to fail.

A bank can cover its customers' withdrawals by borrowing money from the Fed. If a run happens on Tether, who's going to rescue them?

> It also does not make sense for Tether to take on additional risk.

It might not make sense. That doesn't mean they didn't do it.

[1] https://www.kalzumeus.com/2022/05/20/tether-required-recapit...

[2] https://www.kalzumeus.com/2019/10/28/tether-and-bitfinex/ (2019)

>The Consolidated Reserves Report alleges that Tether’s reserves included, as of March 2022, $4,959,634,446 of “Other Investments (including digital tokens).” A 3.27% decrease in the value of these investments wipes out all Tether equity and causes their tokens to be undercollateralized.

First this is wrong, equity includes Enterprise Value which the article seems to ignore. The ability to control $70B of float is worth quite a lot enterprise value. They should be able to take out debt against that EV (or even sell shares/equity) if needed to re-collateralize.

They can also handle another 73-5=68B worth of redemptions before they need to touch those "Other Investments".

There's still some risk if the price is rapidly moving and they can only get $9.99 for the BTC you sent them, but that could be mitigated by ordering the transactions to keep the peg: you send $10 of BTC, they sell it (and "only" get $9.99), they give you 9.99 USDT and say, "tough, what you thought was $10/BTC was really $9.99/BTC"

Yes, why indeed? That's a great question for Tether! If they were holding enough actual cash reserves to cover all the USDT in circulation, there wouldn't be an issue.

Suppose that's all bitcoin, which is valued at "cost less impairment" (i.e. the lowest price bitcoin reaches since the purchase), and bitcoin drops by 10%. That would wipe out about $490mn of equity, putting Tether underwater. The same could happen with a similar drawdown on $3.7bn in "corporate bonds, funds and precious metals," which are marked to market.

They are the only stablecoin I would ever consider buying, because they by now have enough resources to fake it long term.

I've read that the main reasons that people invest in a stablecoin (from least to most) is to: a) buy/sell other crypto easily, b) store illegal gains without drawing attention to yourself c) move your money out of the country.

Reason C is especially pertinent to China, and the environment there (from what I read on U.S. media, which is HIGHLY unreliable) is not pushing them to run the bank.

IMO, the biggest risk for Tether is that the AG suddenly announces that they won't be able to trade in the States, as that will cause a run, and that is definitely possible.

https://makerdao.com/whitepaper/Dai-Whitepaper-Dec17-en.pdf

It does not perfectly track USD but stays within a couple percentage points. Arbitrage traders automatically stabilize the coin thanks to the way the smart contracts are arranged.

The USD:USDT exchange rate should always be 1:1. If 1 USDT is, due to market pressure, suddenly worth $0.98 USD, Tether's responsibility involves some analysis of either waiting for the market to resolve itself (maybe its a transient thing, and its an arbitrage opportunity, so The People may step in), or act as the source of market liquidity; they buy USDT up, with their dollar reserves. Now they, as they always do, have a reserve of USDT they can re-sell for $1 USD without minting new currency, which is convenient (though ultimately a liability on their balance sheets).

But there's a second situation: the exchange rate hits 1 USDT == 1.02 USD. This would happen if demand for Tether were very high; people want safety and stability, so they demand USDT. But: they aren't interested in "leaving crypto"; they don't want to convert USD to USDT, they want to convert crypto assets to USDT. The open market doesn't have enough liquidity to support the 1:1 exchange rate, so the price of USDT starts rising.

This absolutely happens in exchanges which price crypto assets in USDT, which is many. One exchange says 1 BTC = X USD, another says 1 BTC = X*1.01 USDT, and there's now an inter-exchange arbitrage opportunity based on the promise that 1 USD should be 1 USDT. That arbitrage, in some market conditions, can act as an upward force on the price of USDT.

It is ALSO Tether's responsibility to cool off the market. What do they do in this situation? There's only one solution: they need to increase the liquidity pool of USDT. In other words, there are sources of demand for USDT outside the scope of the purist viewpoint of "give me one USD and you get one USDT" rooted in their responsibility not just to act as a reserve bank for USD/T, but also a market maker. Markets misbehave; Tether promises stability.

What's the source of these USDT? They probably have some USDT in their reserves. But beyond that, they need to be minted! Their promise is that every minted token is backed with USD. Assume they're keeping their promise, then follow the line of questioning into where the USD comes from. Return on USD investments for sure. Bank loans? External investor capital? CEO working nights at McDonalds? Regardless, it gets minted and then sold (probably at $1 USD); they make the money back.

And, of course, there's the situation if that assumption is wrong. They have no external source of capital; they just mint tokens to meet demand, hoping that they get sold at $1 USD to refill the reserves. And that's damn convenient, wouldn't it be? USD-backed investments are pretty risky; we could be in a down market and their AAPL shares are in the red; and its USUALLY the case that demand for USDT, for stability, would be high when other markets are in the red; when it rains it pours. They need to pay taxes on the ROI of those investments. Things like bank loans or external investor funding usually also have expectations of ROI. Those are all costs; and it would be REALLY convenient if they could just ignore those and mint the tokens, who cares, the money will flow back in.

This actually works most of the time; it's literally called Fractional Reserve Banking, and everyone does it. But during black swan events, it can break down; especially when you mix in typical finance bro greed, but that isn't even necessary for something like USDT to break down.

There's no perfectly safe way to create a reserve-backed stable asset. The best hope in a deflationary context is that if the price of USD:USDT goes to something like 1:1.01, people preemptively come to Tether, give them USD for USDT, then arbitrage it back to 1:1. But because the logistics of doing that are different than just exchanging crypto-to-crypto (latency, KYC, etc), there could be a demand mis-match. Tether can't let the price stay above 1:1 for long, because it actually devalues the USD basis of anyones' investments denominated in USDT, which is a loss-of-faith event that can spiral to be even worse. So Tether steps in.

If the Tether is pegged to the USD, who would ever sell one for $0.99?

I don't think it's super important to my question though.

Pegged is a meaningless term if nobody is supporting the peg.

As for why would someone buy or sell for other than $1.00? Plenty of reasons, if a market opportunity to make 10% in a crypto investment pops up, damn right people will liquidate tether at a slight loss

It’s been sitting at 99.9 for several days now.

Depending on how many people want to sell, how quickly they want to, and how many market makers are willing/able to arb it back up to $1.00, you might just take $0.99 instead.

This is the problem. I don't believe all the tethers were sold in exchange for $1. If Tether wasn't receiving $1 for every tether then everything else falls apart.

Say they have 50% in cash, 30% in other easily convertible assets, and the last 20% in more speculative instruments.

What if they figured - hey, let's put those 20% on the market. Any returns, we keep, no-one needs to know. The rest we can use as reserve to keep the 1:1 ratio. Hell, investing 20% of xx billions on any fund or security that beats inflation is going to make you filthy rich - especially when there's only a handful of employees.

The disaster, of course, happens if/when any of the markets they're exposed to takes a nosedive, and they either can't prop it up fast enough, or get problems with paying.

Probably just the cynic in me, but I wouldn't be surprised if that's how things play out internally. Without any solid audits, it's hard to say. For all I know they have a very high % of USD reserves, or they could be Bernie Madoff reincarnated.