Thursday, October 10, 2019

Real-Time Gross Settlement

Cryptocurrency advocates appear to believe that the magic of cryptography makes the value of trust zero, but they’re wrong. Follow me below the fold for an example that shows this.

There are two fundamentally different types of systems financial institutions use to implement transactions. In a net system bank A batches up outgoing and incoming transactions with bank B, typically for a day, then the bank with more outgoing than incoming funds sends the other bank a single transaction for the difference. In a gross system bank A executes each outgoing transaction to bank B separately.

If, at the end of the day in a net system, bank B owes bank A money bank A has in effect made bank B an interest-free day-long loan. So the participants in a net system need to trust each other to repay these loans. In a gross system each transaction stands alone; there are no such loans and thus no need for trust. Instead there is a need for cash reserves. Capital must be available to cover the worst-case sequence of outgoing transactions. In practice this is much greater than the end-of-day difference between outgoing and incoming transactions. The advantage of lack of trust comes at the cost of the foregone interest on the greater cash reserves. The trust in a net system creates a pool of zero-cost liquidity.

In the days when transactions involved paper checks daily net systems were natural. But when electronic systems became common, it became obvious that a real-time, and thus necessarily gross, system was possible. Wikipedia explains:
Real-time gross settlement (RTGS) systems are specialist funds transfer systems where the transfer of money or securities takes place from one bank to any other bank on a "real time" and on a "gross" basis. Settlement in "real time" means a payment transaction is not subjected to any waiting period, with transactions being settled as soon as they are processed. "Gross settlement" means the transaction is settled on one-to-one basis without bundling or netting with any other transaction. "Settlement" means that once processed, payments are final and irrevocable.

RTGS systems are typically used for high-value transactions that require and receive immediate clearing. In some countries the RTGS systems may be the only way to get same day cleared funds and so may be used when payments need to be settled urgently. However, most regular payments would not use a RTGS system, but instead would use a national payment system or automated clearing house that allows participants to batch and net payments. RTGS payments typically incur higher transaction costs and usually operated by a country's central bank.
Wikipedia is wrong about the “high-value” part. Using my UK bank’s website, I can make small real-time payments at zero cost.

When central banks introduced RTGS, it had unanticipated results. David Gerard writes:
Izabella Kaminska takes you through how real-time gross settlement works in the real-life banking system — and how instant settlement turned out to be deadly to liquidity, and what this means for central bank digital currencies. “Banks needed funding not credit because without such funds in situ, real-time settlement could not be contractually achieved. That pre-funding need, however, would heighten the system’s sensitivity to logjams imposed by single institutions, and with it threaten total system gridlock.” See also Izzy’s Twitter thread of research — “What i find most interesting about all this is how it ultimately relates to the cost of pre-funding tx in any real-time system, and the degree such a set up will always be more expensive than a netting-based alternative that operates on trust.”
Kaminska's history of RTGS is a fascinating read.

The whole point of cryptocurrency systems is to eliminate the need for trust. Thus they have to be gross rather than net systems. Thus Bitcoin is gross, but not real-time because of the need to wait (typically six block times or one hour) to be certain that the transaction is irreversible. This delay, and the uncertain and potentially large transaction fees, have made Bitcoin useless for retail transactions. The Lightning Network was intended to fix this problem, by providing cheap, near-real-time gross transactions. But it is another example of the cost of lack of trust, because of the need to provide the channels with liquidity. David Gerard asks:
"How good a business is running a Lightning Network node? LNBig provides 49.6% ($3.7 million in bitcoins) of the Lightning Network’s total channel liquidity funding — that just sits there, locked in the channels until they’re closed. They see 300 transactions a day, for total earnings on that $3.7 million of … $20 a month. They also spent $1000 in channel-opening fees."

So $20/month earned by a $3.7M investment makes it worth the risk of being indicted for violating the Bank Secrecy Act? Because decentralization!
$20/month for 300 transactions/day is 0.2c per transaction. To cover 1% interest on $3.7M would need 34c per transaction. But raising fees by a factor of 170 would certainly reduce the number of transactions significantly, so increasing fees and driving an economic death spiral.

If running a Lightning Network node were to be even a break-even business, the transaction fees would have to more than cover the interest on the funds providing the channel liquidity. But this would make the network un-affordable compared with conventional bank-based electronic systems, which can operate on a net basis because banks trust each other.

7 comments:

David. said...

Daniela Gabor has a whole lot more detail on the problems caused by Real Time Gross Settlement in a guest post at Alphaville.

David. said...

Izabella Kaminska summarizes the five main points of the G7's report on stablecoins in A G7 stablecoin report that tells it how it is.

David. said...

Izabella Kaminska's The stablecoin anathema takes PegNet, the latest and spiffiest stablecoin, to the woodshed:

"The rise and popularity of stablecoins is subjecting the crypto world to mass cognitive dissonance.

On one hand, stablecoins are to be adored because they were born out of the cryptocurrency movement, are distributed on blockchains and — in the case of Tether at least — support the entire crypto economy.

On the other hand . . . argh . . . they’re centralised and . . . double argh . . . they’re supported by evil fiat systems and . . . triple argh . . . they’re kind of an admission that people really just want fiat (albeit wrapped in a regulatory arbitrage advantage)."

David. said...

Martin Wlaker's Do six per cent of financial transactions sent via the Swift system really fail? is a detailed takedown of yet another example of deceptive marketing of cryptocurrencies:

"Making a factual error and building a complex but wrong story based on it, happens in all organisations, though most reputable organisations go to great lengths to check their facts before making public statements. Something that should have been straightforward for Ripple given that their “senior vice president of customer success” is Marcus Treacher, a former board member of SWIFT and their global head of banking is Marjan Delatinne, a long-term former senior employee at SWIFT. What is more disturbing is that announcements made by Ripple are repeated and often exaggerated by numerous supporters of Ripple and XRP on social media."

David. said...

One-third of a billion pounds a year is a nice little earner for the bad guys. Izabella Kaminska's The real story behind push payments fraud reports on the epidemic of fraud (£354m in 2018) in the UK's faster payments system, to which I referred in the post:

"With push payments customers initiate transactions on an irreversible basis, unlike direct debits which can be challenged by the payee, and that’s a key part of the problem."

Just like cryptocurrencies! The problems have two fundamental causes, implementation:

"While the European payments directive specified that any new payments system would need to collect a recipient’s name, it apparently did not specify that the information would have to be processed, which allowed the legacy card system to form the basis of the new infrastructure. Accordingly, names are never processed, meaning it doesn’t matter if the name in a transaction order doesn’t match the account numbers given. The transaction gets processed regardless.

This alone has helped criminals dupe unsuspecting victims, by making them believe they are sending funds to themselves when they are not."

But also inherent:

"There is always likely to be a trade off with any frictionless payment advance due to the architecture of secure-systems and the requirement for someone somewhere to bear the risk of non-payment or fraud. Simply speaking, the cost in the payments system is, and will always be, related to security, screening, compliance and credit guarantee. There is no free lunch."

David. said...

Izabella Kaminska's work on RTGS continues in A not so freely-floated market. She quotes Carolyn Sissoko, economics lecturer at the University of the West of England, who points out that:

"the banking system developed precisely in order to address the problem of providing unsecured credit to support netting as part of the settlement of payments."

Kaminska comments:

"This is a fantastic point. The presence of credit-providing banks (which have the capacity to net off exposures between them) allows the system to make gigantic liquidity savings, which otherwise would not be made in a capital-markets exclusive system.

And that by and large is a function of a trust-based financial system. Trust spares the system liquidity. Since banks trust each other, they don’t need to liquidate everything to the full gross level just to execute transactions that ultimately only change the composition of wholesale asset ownership or custody on a marginal basis."

David. said...

Izabella Kaminska's By Jove! Crypto has discovered netting is a must-read descr4iption of the problems the Lightning Network faces:

"1) The system is massively credit dependent.

2) The system is extremely liquidity draining.

3) The system’s dependency on pre-funding is likely to be a huge deterrent to most users (do consumers really want a system that requires them to tie up funds with retailers they may or may not use?)

4) The system is hugely exposed to the coincidence of wants problem, especially when it comes to getting funds out of the system and back into convertible bitcoin. This is because there is no mutual counterpart everyone has in common or has channels with, but also because there is no compulsory clearing cycle.

5) To overcome the coincidence of wants problem the system will have to rely on an expansive pathway of pull-payments arrangement. This opens the system up to many as yet unknowable credit related risks.

6) Those funding the system and taking all that risk — unlike in the conventional depositor-dependent banking sector — get no return or reward for funding or providing liquidity to the system, or for having their funding locked up."