Always and everywhere banking has faced precisely this problem, which is the original reason for bank regulation and deposit insurance-type backup.

If you encounter an unregulated or irregular market participant doing anything that looks like banking, buyer beware-- on average, over time, you're not getting your money back.

I feel like cryptocurrency markets and investors are painfully “re-learning” the lessons of a hundred years ago.

It seems that reinventing finance works out about the same a rewriting a big software project. You think you'll get rid of a lot of nasty code, but then you realize all that ugliness was there for a reason and you're just rediscovering all the edge cases again the hard way.

But in doing so, you bring that latent knowledge back into the minds of living people who can rethink the solutions in the face of the modern ecosystem and clean out the cruft that truly has no modern purpose.

Do you think no one in finance has thought about financial regulations until crypto came around? I can assure you there were (and are!) many people paid quite a lot of money to “rethink the solutions in the face of the modern regulatory ecosystem and clean out the cruft that truly has no modern purpose”. One of the reasons securities regulations are “expensive” to comply with is because they’re constantly changing. Why are they constantly changing? Because we’re rethinking how, when, and whether the various regulations ought to be implemented.

The regulations implementing Federal securities laws are perhaps the most nimble, iterative, and “listen to your users” driven regulations in the US.

Certainly no one in finance ever thought about programmatic regulation of a currency. But your comment is of the type that has the premise "everything that could ever be thought of has been thought of before". Well obviously that isn't true. And it also isn't true that everything that has been thought of and not done is not doable.

Certainly? How certain are you of this?

I am not so certain.

I don't think that, no. Still, different people with different toolboxes will come up with different solutions.

I think the problem is that people think they are reinventing finance when the only thing that changed is the currency.

There exist algo stablecoins too though.

Highly underrated comment

Except that now its even worse, because we are throwing a whole lot of electricity and advanced manufactured goods (computer chips) into it, which currently has large negative impacts on the rest of civilization.

This is an interesting point, what does it mean when it is gray?

Comments get grayer as they are further downvoted.

"Extraordinary Popular Delusions and the Madness of Crowds" (1841) is still worth reading. You get to see the first versions of the classic large-scale scams. John Law's bank, the South Sea bubble, tulipomania. Scams go much further back, but mass-market scams had to wait for mass communications. Newspapers powered the first round of large-scale scams.

Most scams today are very similar to some classic scam mentioned in that book. There's not that much innovation.

Kids should be taught the classics in school - the short con, the long con, the big store, the pigeon drop, the badger game, the protection racket, the bait and switch, the Ponzi scheme... The list is not that long.

Personal finance should be a subject in schools, and scam awareness should be one of the topics covered. The long-term economic benefits would be huge.

At the most basic level, some entity is making statements about their financials (like “we have X in reserves”) and nobody can check this. If a liar can make financial statements and be believed, and nobody checks, they can do anything.

It’s odd that cryptocurrency was designed to work without trust and yet many people are so trusting.

> At the most basic level, some entity is making statements about their financials (like “we have X in reserves”) and nobody can check this

This is not a problem unique to stablecoin providers. Every company with a large number of stakeholders (e.g. public companies, bond issuers, companies that manage assets on behalf of their clients, etc.) faces the same problem.

The traditional solution to this is to have a widely-trusted third party, like a big four CPA firm, examine the company's books and operations and provide an audit. It's still possible to defraud auditors (there used to be a "big five CPA firms" before Enron blew up and brought down Arthur Andersen, and more recently Wirecard made EY look pretty bad), but it's a lot harder than defrauding people who don't get a detailed look at your operations. The reason to go with a big four accounting firm is that the firm's reputation with the public is substantially more valuable than any particular client relationship.

This is why substantial organizations have 3rd party audits and reviews.

All of this stuff has known solutions, but people don't like what the solutions would show.

> At the most basic level, some entity is making statements about their financials (like “we have X in reserves”) and nobody can check this.

USDC seems to be doing it just fine?

https://www.centre.io/usdc-transparency

> Top five accounting services firm Grant Thornton LLP issues attestations each month on the US dollar reserves that back the USDC tokens in circulation.

USDC does seem better, but the last attestation is from April and there are reports that they've been running late a while [1]. That's not as bad as the competition but still seems very bad, given that stability is what they do? Especially since they haven't explained what happened as far as I can tell.

On the other hand, it looks like USDC's market cap doubled in less than two months, so I guess that's good for them?

[1] https://news.bitcoin.com/usdc-attestations-run-late-raising-...

Attestations, but I believe no true audits.

Yup. If one can make money now by pushing off problems for the future, oversight is vital. Otherwise we guarantee problems in the future. Banks are a great example. As are investment companies; Ponzi schemes are one easy way for those to go wrong.

One non-governmental way to do that is audits. Which is why Tether's lack of serious audits is so concerning. They claimed to be cash-backed for a long time but wouldn't prove it. Once the NYAG force them to open up, it's clear they were lying.

Actually this is no longer true. As of recently, the crypto ecosystem produced Dai (https://makerdao.com/en/), which is a fully decentralized stablecoin. I think it represents one end-state of the crypto ecosystem, blockchain transactions but with decentralized systems keeping volatility near zero.

I've only skimmed the white paper, but this appears to be a "stablecoin" backed by crypto assets. Is that correct?

The white paper is complex, and the Emergency Shutdown price stability mechanism appears exposed to ETH price risk such that if ETH moves rapidly around the Emergency Shutdown, there's no guarantee that Dai holders could recoup on a 1:1 basis. Is that correct?

This seems to be not a dependable peg.

The Bank of England couldn't keep their peg, so what's the gold standard for a dependable cross-asset peg then?

DAI has been tremendously successful over multiple years and multiple bear markets and ETH price crashes; it can only do so much of course. If ETH goes to zero tomorrow, DAI holders will not be able to recoup, sure.

In fact, the real problem with DAI is that demand often exceeds supply, the latter being limited by people who want to go long ETH, so they keep having to fight DAI breaking the peg in the other direction. To solve this, they onboarded other collateral types, including the USDC stablecoin, which now unfortunately accounts for >50% of DAI collateral.

> The Bank of England couldn't keep their peg, so what's the gold standard for a dependable cross-asset peg then?

I'm not a crypto expert, but naively I would at a minimum expect uncorrelated collateral to be part of the picture. DAI runs on ETH, so collateralizing via an ETH mechanism seems like possibly not the best choice.

I would expect that the peg would depend on something that exhibits a low correlation with crypto asset prices and a very high and dependable correlation with USD, like money market funds or Treasuries.

Dai is backed by lots of different assets in the Maker Vaults. Most of them are crypto but they just added some real estate assets and the plan from day one has been to add as much diversified collateral as possible for safety and scale (Dai can't scale big enough to service the entire world on crypto backing alone).

How are non-crypto assets represented in crypto space?

It seems you would need a trusted party/oracle of some kind, which creates the same problem as centralized stablecoins anew.

Edit: indeed a sibling comment mentions they have USDC as a backing support which means they are as stable as that.

> It seems you would need a trusted party/oracle of some kind, which creates the same problem as centralized stablecoins anew.

If it is multicollateral, then yes, it somewhat involves trust, but you are dependent on a lot of various people betraying trusts along with ethereum price dipping, whereas a traditional stablecoin is entirely dependent on one entity.

Divirsification of assets does help reduce risk. However, you can't eliminate that risk without holding reserves that are entirely made up of the pegged currency.

A substantial chunk of DAI is backed by USDC. USDC has never been audited, and has been curiously late with its attestations at precisely the same time the supply of USDC has massively increased.

And the value of ETH backing DAI is dependent on Tether not being fraduluent and instead being worth $1.

You think a joint collaboration by Coinbase and Circle with monthly audits has "never been audited"?

https://www.centre.io/usdc-transparency

It hasn’t, no. Those are attestations, not audits.

You won’t find the word audit on that page or in the reports.

I thoughts an attest is a form of audit. Can you explain why you think this is insufficient?

And aren't these reports going to take time to complete? Every report in the past took about a month to complete and I don't see why that's a red flag.

Former auditor here.

An attestation offers considerably less assurance than an audit.

An audit is the most comprehensive type of assurance. Often called positive assurance. A clean audit opinion means the auditor collected sufficient and appropriate evidence to form an opinion on the financial statements (or reserves in tether/usdc case).

On the other hand, an attestation or review is a form of negative assurance where auditors state that nothing has come to their attention to indicate that subject matters or financial statements contain a material misstatement. In this type of assurance, auditors do not give an opinion; they simply say that financial statements look "reasonable".

Unlike positive assurance, auditors are not required to obtain sufficient and appropriate evidence to form an opinion. Instead, they only need to review if there are any problems with financial statements or subject matters.

Thanks! How would an attestation work with a fraud. For example, suppose a company simply produced a false bank statement.

Would an attestation have no ability to verify that the statement was fraudulent? In other words we must trust the entity undergoing attestation in order to rely on the attestation, and the attestation merely certifies there is no error of math or logic in what was presented.

Good question.

With an audit, the auditors get a representation from management that they will provide the truth etc. The auditors would also get third party evidence eg. from the bank providing the audit client's account. For important things you would always get third party evidence from banks, custodians, etc or even just go and check to see if physical things exist!

With an attestation or limited/negative assurance engagement, there's no third party evidence. Instead, the auditors just rely on what they are given and whether it looks reasonable. The auditors would state in their "report" that only limited evidence was gathered and not enough to form the basis of an opinion.

Basically, limited / negative assurance is not really that useful in most circumstances.

Regarding fraud - auditors are not expected to find/uncover fraud under any type of engagement, which is a common misconception.

The biggest audit firms won't go any where near tether, and this alone, tells you quite a bit :)

Thanks, that’s what I figured. And that’s very interesting about not even audit or assurance finding fraud.

> Basically, limited / negative assurance is not really that useful in most circumstances.

So what exactly can we glean from USDC having attestations? It’s certainly a step up from Tether in that respect but I’m also not sure it tells us all that much.

Or maybe a better way of asking is: how exactly would you prove that a stablecoin was backed?

To prove a stablecoin was backed you'd probably do the following:

1) review the processes and controls which operate the business to check they were operating correctly for the period under review

2) interview the various key stakeholders to assess competence and get representations

3) perform substantive testing over the collateral balance for the whole period. Eg daily bank reconciliations. Get third party confirmations for EVERYTHING.

4) perform a contingent liabilities review and a legal review.

5) see if there are any related party transactions

6) do a going concern assessment

The key thing would be to check existence, completeness and valuation of collateral and existence and completeness of liabilities (issued tokens).

Depending on what the assets are that would entail different procedures. For tether I would want to see their whole CP portfolio to perform a thorough credit risk and systemic risk assessment. Do some modelling to understand valuation implications under various scenarios.

Is worth noting that it's not feasible to do this on a monthly basis because it's so onerous. Hence why probably they just do monthly attestations. I would expect that the legal entity which issues the tokens and holds the collateral is audited at least once a year.

Never knew my audit knowledge would ever be useful/interesting :)

Cheers

They seem to have caught up somewhat for the April report, but March was extremely delay. News article below.

I’m not an expert, but I think an attestation just examines if a statement makes sense. My accountant did one for my revenues and the percent that were in USD. I sent them a spreadsheet with my revenues from various sources and calculations showing total USD.

The accountant verified that my spreadsheet said what I said it said. However, they did not actually verify the info underying the spreadsheet beyond examining some screenshots of customer addresses I provided. They samples a handful at random.

In USDC’s case, I think the auditor would look at a bank statement and say “the bank statement on May 31st indeed says Circle has $X” and Circle says this money is theirs for backing USDC.

Stuff they wouldn’t verify:

* Was the money there before that specific minute of the day?

* Did it remain there after?

* Was the money from deposits, or was it from a loan or some other source? (Bitfinex did this with a prior attestation, mixing up Bitfinex’s money and reserve funds)

So most people would assume these attestations mean “At all times USDC had backing of basically all of their tokens by $ in a bank account, free and unencumberer” but the attestations don’t examine that claim at all. They examine a very specific moment in time, and don’t examine the source of the funds.

In an audit you might actually examine the accounts at a time not chosen by Circle.

https://news.bitcoin.com/usdc-attestations-run-late-raising-...

Centralised stable coins maybe or yes. But there are also decentralized stable coins like DAI that have smart contracts to ensure a level of backing

> decentralized stable coins like DAI that have smart contracts to ensure a level of backing

Which has been collateralized by centralized stable coins including tether, if i recall correctly..

Dai is not collateralized by tether.

Isn't the DAI better describes as semi-stable since it holds reserves in Ether and other cryptocurrencies but is pegged to the dollar. This exposes it to systemic risk in the way that stablecoins backed by reserves in their pegged currency are not. The degree of excess in their reserves does cushion against a lot of violatility, but that cushion is finite and could easily vanish in a cryptomarket crash.

AFAIK it survived ETH price crash from $1400 to $90 .

Have any details on this? I find those numbers implausible without a context or a citation.

You can just look at the market cap and price history of DAI through the last market crashes

I did, it was not clear to me what was being referenced. If there is a good example of DAI surviving such instability, a reference discussing how that happened seems like a pretty reasonable requests given the strong claims that are being made.

FYI skip to 2:10 if you don't wanna watch a skit

I started using SponsorBlock (https://sponsor.ajay.app/) a bit ago. I was skeptical that the crowdsourcing wouldn't reach enough but I find it works for most videos.

It essentially tags parts of videos and it's pretty customizable - it only highlight a certain type, have it skip automatically, etc.

I already pay for YouTube Premium and am not a fan of ads.

The whole video would have been better if he had built an argument from the ground up instead of starting with the idea that Tether is a scam and laying so many attempts at humor and irony on top of it.

Or at least it would have been clearer and more educational; I haven't seen his videos before and maybe other people watch him for entertainment.

Why would they kill the golden goose though? Think of how much money they could be making by simply holding deposits...

The amount of money a bank (or loosely equivalent crypto entity) can make holding deposits and making loans out of them is far lower than the amount it can make by printing new fiat (or crypto tokens) out of thin air and loaning them out (/selling them.)

Usually, this works fine, they make a lot of money, and everyone's happy. When there's some unexpected macroeconomic issue, which empirically happens with some regularity, they have massive problems as loans default en masse (or everyone tries to redeem their token for USD at once.)

The other difference is that loaning out to unconnected parties /= selling the tokens and keeping the uncollateralized bit as profit. Banks with outstanding loans might have issues with the timing of repayments, they might more rarely make so many bad loans that their expected future income is less than they loaned out, but they haven't simply pocketed the money depositors paid in.

And it's critical to note that Tether has, in fact, refused to process any redemptions for any US client since 2018.

I’m wondering why non-US intermediaries couldn’t do it?

According to their stated policy (https://tether.to/fees/), a non-US intermediary would have to:

* Pay them a nonrefundable $150 fee.

* Wait for your account to be verified, if you're approved - they have "sole discretion to approve or not approve accounts"

* Send them a minimum of 100,000 USDT

* Pay an additional $1000.

To be clear, it's likely that all of this is a lie and they simply aren't processing redemptions for anyone. I'm not aware of any documented story from the last few years of someone successfully performing a Tether redemption. The US policy is just easier to talk about because it's explicit: they say, in black and white (https://tether.to/faqs/), that "no issuance or redeeming services will be available to these users".

There's no evidence that Tether does redemptions for anyone.

If you ask any trading desk in the space they will tell you they personally regularly do this.

What bank did the wire transfer come from?

"When there's some unexpected macroeconomic issue, which empirically happens with some regularity, they have massive problems as loans default en masse (or everyone tries to redeem their token for USD at once.)"

"Nobody knows you're swimming naked until the tide goes out ..."

Because killing the golden goose could mean a several billion dollar payday for the insiders?

That's exactly how fractional reserve banking started...

Which is backed by a governments, laws, and regulators…

If my deposit account gives up the ghost, I know the FDIC or the Fed will make me whole (or SPIC for my brokerage account). Who will do the same for Tether and other stablecoins?

Not good!

>the FDIC or SPIC

You mean they will print more money to fill your loss (dilution net effect zero)

>the Fed will make me whole

Says who?

So, I'm really unsure how to have these sorts of discussions with crypto proponents. The US government's securities (UST), as well as the US capital markets, are some of the most trusted financial assets and mechanisms in the world. Treasuries, in particular, are considered one of the safest asset classes to hold. As a crypto proponent, you might not believe that, but collectively $22 trillion of capital (and those responsible for its management, sovereign and institutional alike) believes that's the case. If the US government fails, crypto is likely to be of lesser value than say, ammo, food, and fuel.

This is in stark contrast to offshore crypto operations who refuse transparent auditing of their reserves.

The FDIC's Deposit Insurance Fund has a balance of over $119 billion as of March 31. No need for money to be printed. https://www.fdic.gov/analysis/quarterly-banking-profile/stat...

They're not "backed". It's simply not there. They're backed by the law that if you arrange a "run on the bank", YOU will go to jail.

They're not backed by an imaginary law, they are backed by an outstanding portfolio of expected loan repayments, the ability to borrow all the short term cash they need at lower rates, and ultimately you're backed by the FDIC even the people the bank loaned out to all default.

How do you go to jail if "arrange a 'run on the bank'"? There's no law against it, so far as I'm aware.

As for the backing, let me go through this in more detail (see https://www.bloomberg.com/opinion/articles/2021-06-10/maybe-... for a decent coverage of how this works). There are three categories to keep in mind here: assets (which include loans, cash, precious metals, etc.), liabilities (i.e., the value of all of your deposits), and capital (which is money to support the assets that's actually the bank's owner's).

Regulation requires liabilities to hit no more than about 90% of assets. They also require the bank to stake about 8% of its own capital to support the assets, on a risk-weighted basis. The risk weighting reflects the likelihood of assets crashing in value: a typical loan might have a 100% weight, so $100 in loans requires $8 in capital. Safer loans (e.g., residential mortgages) would require less capital. Stuff like cash or T-bills have a risk weighting of 0%. Riskier things include stocks (which tend to go down big when they go down), requiring maybe $24 per $100 of stocks, or cryptocurrencies... which require $100 of capital per $100 of cryptocurrency.

What this actually means, when you combine stuff together, is that you really do have something like $100 in safe things for every $100 in deposits a bank has.

> but inevatably greed kicks in and you start thinking what if you minted coins without the backing.

Yet another tired implication of Tether printing coins without backing, so let's break down the NYAG's announcement of Tether's fine. The fine was for:

1. Buying a bond (= lending out their backing) in a related party transaction, and

2. Not updating their customers that Tether backing had changed from 100% cash into a mix of cash+bonds.

The guy with the handle bitfinex'd was an extremely prolific screamer about fraudulently unbacked tether and folks seemed to believe him. If you still believe him, it now means you believe that NYAG Leticia James looked at Tether's books and gave them a pass on fraud in order to collect a few million dollars in a fine. Not bloody likely.