Hacker News Reader

Comment by jacquesm | Hacker News Reader

Chances are they'll be dead if enough people act on that. So they're doing the opposite: reassuring people that everything is fine. If there is one common element that precedes every bank run and run on crypto exchanges and such then it is the 'everything is fine' phase just prior to the implosion.

View on HN

One clarification:

1. For a bank run on an actual bank, it is known that a bank doesn't keep all your assets on hand - they are loaned out, which is (largely) what allows you to earn interest. So it is known that if everyone tries to withdraw at the same time that there won't be enough money.

2. For an exchange/brokerage, your money/assets are explicitly NOT supposed to be lent out without your permission. So, in theory, if everyone asked for their assets all at once, the assets should be there.

Indeed.

But we already have a whole bunch of examples of (2) where the funds ended up in all kinds of places where they should not have been.

Re 2: this is not true if you're doing margin or futures trading, which was FTX's main business (and may be Binance's). This type of business is what characterizes an "exchange" (as opposed to a non-margin brokerage or a custodial bank).

Hence the "without your permission" caveat. Point being, when you lend your assets to brokerage, it's explicit, unlike a fractional reserve bank, where deposits are inherently lent out (and no need to comment that fractional reserve banking doesn't "really" work this way, I get it, but deposits are still part of the capitalization of the bank).

To be fair, it would be better if you explicitly lent money via certificate of deposits to the banks because it means their job is much less risky. The fractional reserve banking model isn't really something to imitate, it is a rather ugly hack that we cannot do without when people don't get enough CDs.

Perpetual futures don't require any margin. All it requires is matching buy side and sell side market participants and transferring money between their accounts as the underlying changes in value. The exchanges don't actually hold any underlying with perps, so there's no need to borrow money.

How does a perpetual future actually work?? What happens if 5 people buy the "sell" side and 1 person buys the "buy"? There's no way to redeem to the underlying, and I don't think the 1 person pays out 5x the price movement to the other side.

An exchange can hedge vs excess buy orders by buying the underlying, but how do they hedge vs excess sells?

I've been trying to figure out how to short BNB, and perps feel pretty much as risky as trying to do it on binance itself.

Each trade is bilateral. Futures, both regular and these "perpetual swaps" are zero-sum. A person buying is another person selling and vice versa. The exchange does not need to hedge the underlying because the exchange/clearing house is a neutral party in the transaction as long as the individual traders are solvent. What it is exposed to is traders blowing up and not being to cover their debts.

In case of regular futures, price discovery is aided by the fact that the futures eventually settle to either the underlying or its cash value at a specific point. In perpetual futures, price discovery is aided by "funding fees," which are periodic cash transfers from the side that is contributing to the price discrepancy between the future and the spot to the opposite side. E.g., if perp is below the spot, the short holders will periodically be charged the funding fee, which will go to the long holders, to encourage the shorts to buy/get the price closer to spot, etc.

If five person buys and one person sells, only one person actually has the buy order executed. The other four are just waiting in line.

Counterparty risk of exchanges blowing up is exactly why tether doesn’t have a gigantic short volume outstanding. Shorting stablecoins or coins that represent profit share of exchanges like bnb is incredibly risky and that’s largely on purpose, the ecosystem doesn’t want people shorting it.

In theory, the exchanges should lend their own money to margin traders.

If there’s a single case of an exchange using a customer’s assets to make bets isn’t that a fraud charge? Or with the unregulated nature of crypto is it not that explicit?

Dipping into customers funds for operating expenses or other uses (loans to associated parties, investment, leveraging) is a big no for an exchange, obviously, whether or not it is fraud is ultimately for a judge to determine but it sure looks like that to me.

The lure of those balances is invariably too great to resist.

> The lure of those balances is invariably too great to resist.

I think the folks at JP, etrade, and all of the other regulated exchanges would object to that characterization.

Regulated exchanges are obviously a different branch entirely, but the rest of them have serious problems keeping their hands off the customer funds. And as was pointed out to me upthread even the regulated exchanges may not have enough of a wall between the crypto they hold for customers and their operating funds to guarantee that those customer funds are going to be there if the company fails.

Time and again we find that exchanges have been dipping in and using money that wasn't theirs in ways that would never fly in a regulated environment.

Guess how most of those regulations came into existence?

these are the key points that get overlooked. i keep seeing that folks expect these brokerages to work like banks work.

The problem is that these brokerages do work like banks work but pretend that they are holding all of the reserves on the books.

They are leveraging and borrowing from the customer funds with abandon, just like a bank but without the associated regulatory oversight and rulebooks.

In the runup to the Lehman Brothers bankruptcy, the CFO Erin Callan would come on CNBC week in and week out to talk about their "fortress balance sheet" and it would always pump the stock a little bit. Was absolutely infuriating to watch, but you can't get too mad at the people doing it, regardless of how bad things get their incentives are 100% aligned with projecting confidence and praying they can hold on long enough for the market to turn back up.

This is true, but there is some availability bias. There are presumably cases where "everything is fine" turns out to be pretty accurate (at least for a while), but we don't keep track of those.

Reply on news.ycombinator.com