Tell HN: SWIFT is not a payments transfer system | Hacker News Reader

Tell HN: SWIFT is not a payments transfer system

244 pointsatdrummond141 comments

As part of my startup's remit, I do a lot of work with clients' SWIFT payments. I feel like the media has done a very poor job explaining what SWIFT is and how exactly it works.

SWIFT exists not for the purpose of providing the rail(s) upon which money moves but to be the messaging system which coordinates said transfers. When sending funds (typically done via a document called a MT103, which is a text file with specific formatting requirements) one informs the recipient institution by sending a SWIFT message. When it is time for the banks to actually move the funds (which often times occurs significantly after the funds are made available to the client in question) this will be coordinated through a payment system. In most cases in Europe, this is TARGET2. However, as Russia is not a member of the Eurozone, many of their impacted payments will likely be on the STEP2 payment network, which is the only European-wide clearing house. (However, these days, it isn't uncommon for banks to work together to use multiple clearing houses across multiple jurisdictions to most efficiently move funds.)

From what I understand from banks and clients we work with who have a significant Eastern European nexus, Russia was already experiencing problems with the payment systems before the formal SWIFT bans were announced and that the effective shutdown of central bank settlements for Russia has served to bring international payments coordination for RU based institutions to a crawl.

Happy to add clarification where it could be useful.

View on HN

What you are implying here is that SWIFT ban is not that big of a deal as there are alternatives.

You are probably more educated on this, but I'm an Iranian and Iran was the only country so far that was banned from SWIFT. And it was a very very big deal. The banks could not deal with anywhere in the world and it become very difficult to send or receive money.

The EU tried to setup an alternative to SWIFT [0] so they could work with Iran but it didn't really work out.

[0] https://en.wikipedia.org/wiki/Instrument_in_Support_of_Trade...

A big complicating factor there was also the concomitant sanctions by the US and some of their allies. I don't want to imply that I don't think banning SWIFT wasn't important or impactful in either of these cases (it was) - I just wanted to make clear what precisely was happening, as I don't think most reporting was very clear. I'm sorry if it came off otherwise; I appreciate you dropping this comment as I don't want anyone getting the impression that I'm flippantly suggesting that Russia could easily go on with business as usual through alternative means.

> I don't want anyone getting the impression that I'm flippantly suggesting that Russia could easily go on with business as usual through alternative means.

One would obviously hope that the cumulative effects of the various sanctions make it difficult or impossible for Russia to continue with business as usual.

However, part of me worries about the abilities of the oligarch network.

I mean, if you look at the various flight tracking websites. Sure a lot of the commercial aviation traffic has been killed off, but there are still large numbers of private jets floating around in and out of Moscow. How is this possible ? Because the sanctions are only against Russian registered aircraft, and most of those private jets are registered elsewhere and/or are hidden behind charter companies registered elsewhere.

Quite possibly this extends to finance ? I imagine there is potential to find alternative financing routes. It wouldn't be able to last forever, for sure, but might last longer than the West anticipates.

It does to finance as well, however the intent is to also target 2nd and 3th degree actors because it is extremely common to hide the true owner of the money in Finance (tax evasion, sanction evasions, ...)

Here SIGINTEL can contribute a lot but also just better data pooling between the countries and more modern methods of detection.

This international commitment to combat money laundering has so far been lacking but it seems this invasion is giving the momentum needed to put such systems and collaborations in place.

It will still take some time though before it's in place.

In many ways, the oligarch jet movements are a positive sign: they're having to move around (or at least move their assets around) and incur both significant cost and scrutiny in the process. For all you know, a Bermudan registered Gulfstream moving from Monaco to SVO, Moscow, contains a pissed off oligarch watching his lifestyle and investments crumble.

Or, it could be a strategic hedge to ensure he retains control of the jet.

Either way, the sanctions are causing an effect, and by all accounts a significant one. Battles are won or lost by explosives, inexplicably awful things and largely brave men being sacrificed, by their choice or others. Wars are won or lost by economics.

I worry a little that we will test the limits of sanctions. Sanctions kill, yet are often not considered an act of war. We are used to imposing them on countries with little world power or ability to fight back.

If the sanctions are harsh enough, is there potential for a military response?

There's a risk of that if we don't abide by Putin’s demands to back NATO away from its existing positions as well as stopping expansion.

Binary “is there a risk” questions are mostly not useful, because the answer is almost always yes.

how is crypto scene in iran? with india's recent tax daylight robbery on crypto i am confident that a lot more people will join the informal network of owning and transacting in crypto. if i have to buy 1k worth of BTC or ETH for example, i can imagine paying a friend the cash and getting their wallet in return.

now translate that to international cross-border transactions and we have a bypass. do iranians accept crypto as in other countries (to be fair, even india is far far behind but the recent acceptance by imposing tax is seen in a good light overall)

North Korea is also banned from SWIFT.

> What you are implying here is that SWIFT ban is not that big of a deal as there are alternatives.

I don't think it's so much that, but the UK and US bans on dollar clearance for banking which happened a few days before the SWIFT block actually did 90% of the work that the SWIFT ban is getting the credit for.

It's like turning off the metadata layer two days after putting an axe through the ethernet cable.

SWIFT is a stone-age blockchain. Its messages have non-repudation properties.

The ban makes it much harder to make settlements.

I imagine the idea being that without SWIFT, any bank will require the funds to have actually moved before they are made available.

Unlike Iran, Russian banks has not been all disconnected from SWIFT. Only a few banks were. So effect is much smaller. In general, the sanctions on Russia are limited and not strict, the US hasn't gone anywhere as far as it could[0].

[0] https://twitter.com/M_S_Billingslea/status/14984295311951544...

Also, in Iran's case there's the FATF issue which doesn't yet exist regarding Russia.

So basically they are giving time to those that should be affected to move their funds and later when sanctions are increased it will only cause collateral damage.

Oligarchs don't have a ton of cash - the VAST majority of their assets are either in the Russian companies they own majority stakes in (which are down 50-90%) or overseas (which have been frozen and are actively being seized).

It's pretty hard to take all $80Bn from a foreigner. Especially without destroying a lot of value for other people.

> or overseas (which have been frozen and are actively being seized)

Except in the UK, where they fund the tory party so there's very little being done.

???

Boris Johnson has been the most outspoken voice outside Ukraine for increasing sanctions against Russia. He's already announced freezing assets of all Russian banks and 100 other people or corporations,[1] and he led the charge of getting Russia cut off from SWIFT.[2] Germany and Italy were the ones opposing it since they are dependent on Russia for energy.

[1]: https://www.bbc.com/news/uk-60515626 [2]: https://12ft.io/proxy?q=https%3A%2F%2Fwww.ft.com%2Fcontent%2...

What Boris Johnson says and what Boris Johnson does are two entirely different things.

https://www.opendemocracy.net/en/dark-money-investigations/u...

This was before any of the recent actions.

I'm a pretty big BoJo hater. But Londongrad existed long before BoJo - don't blame it on him.

So far BoJo has been leading the US and the EU in pretty much every form of support.

It absolutely did, but that's not the point I'm making. I'm pointing out that the oligarchs who fund the tory party don't appear to be under any sanctions at all. From this piece [0] "On Tuesday 22 February, the UK authorities sanctioned three oligarchs and five Russian banks" - all three oligarchs were already heavily sanctioned by the US and basically unable to operate in UK financial markets anyway.

He's sanctioning the ones who aren't funding everything he does, anyone giving him money is still absolutely free to do as they please. Just because that man says he's doing more than anyone else, doesn't mean he is. In fact, it's generally safer to assume truth is the opposite of what he claims publicly.

[0] https://www.opendemocracy.net/en/odr/uk-russia-sanctions-ukr...

Some banks lease other banks' tarnsfer mechanisms and I've read official statements that their card infrastructure was also affected.

I didn't get the sense that the author is implying that anything is or is not a big deal. I believe the author was just correcting for a technical audience that can understand it what SWIFT is. As the author said, the media makes it sound like an API for moving money around. There were some other posts on HN about the consequences of misunderstanding what SWIFT was, I recall one person mentioned being assigned a project to "integrate with SWIFT" which turned out upon deeper analysis to be impossible as assigned, because the people handing them the assignment didn't realize what SWIFT actually was and also had the same misconception.

I would be interesting to hear an further explanation in how this distinction is important.

Money in banks is in the end just information stored in ledges about who owns who money. If a transaction occur within a bank it just numbers of one account getting an relative change to an other account. If an transaction occur between banks, the banks has a relation and the bank own ledger get changed, in addition to the individual accounts. With a clearing house you add a middle man that takes up part of the risk if either bank would fail in their obligation to honor the ledger. The actually movement of physical funds and valuables is quite rare and don't general occur from single transactions.

If you can't transfer the information then the system of transferring information stored in ledges stop working. One can go out-of-band, but that assumes the other party is also willing to do so, and that there is no law/agreement against by-passing sanctions.

So how does it work when there are multiple of these systems (CIPS). Is double spending possible?

Of course double spending is possible -- the entire world economy currently relies on it.

Essentially trust. If your bank does not have relations with a foreign bank in a faraway country, both can use an intermediary bank that both trust and do have relations with.

> how this distinction is important

It's not. At least not in the short term.

The distinction is which messages we are talking about. When I (a US bank) send a message to the US Federal Reserve to please transfer $1,000,000 to another US bank, and the US Federal Reserve updates their records accordingly, that is an actual transfer of money.

When I send a message to the other US bank saying "Hey, I sent you $1,000,000. Of that, $10,000 is for John Doe, $10,000 for Jane Smith, $5,000 for Alice, $5,000 for Bob, ......" that's SWIFT.

To my understanding, the common method is that bank A and Bank B is both customers at the clearinghouse, in your case the US Federal Reserve. Bank A has an account, and Bank B has an account, generally a deposit serving as security.

A "bank day" is basically the sum at the end of the day that bank A own to bank B. No physical money get transferred unless the US Federal Reserve requires it, which if both banks has fairly similar outgoing debt to each other the accounts won't change by much. the US Federal Reserve don't see the "$10,000 is for John Doe, $10,000 for Jane Smith", only the sum of all transaction going from A to B or B to A.

Bank A tells bank B that "I told the US Federal Reserve to update your account by +$1,000,000 and reduce my own account by same amount, and you can verify this by talking to the US Federal Reserve and checking your account. The items that those +$1,000,000 represent are listed here. John Doe get $10,000, Alice get $5,000, ...". This occur through the SWIFT protocol. Bank B then do the same by first telling US Federal Reserve and then giving bank A the list of the individual items.

The US Federal Reserve don't really transfers money. They simple managed the ledgers for banks relation with other banks. What occurs is simply a ledger being updated about how much bank A is in dept to bank B, and wise verse. The federal bank will step in if a bank defaults, or I guess if the differences between A and B becomes too great to manage.

The US Federal Reserve's ledger *is the money*, in exactly the same way that the Bitcoin blockchain is the money.

Thanks, this comment has explained to me what SWIFT is more than anything else I've read so far :)

Heard a finance guy making this argument on radio news. His point was that it was essentially a telecommunication system, which could be replaced by telex or or other messaging service as a less efficient way of getting payments through, or at least important, large payments with cooperating banks and clearing houses. It ruining the coordination of day to day payments seems to have had the desired effect on the Russian economy though.

There have already been a lot of new rules put in place blocking transactions with Russian banks, which I guess is meant to take care of the "cooperating banks and clearing houses"

> essentially a telecommunication system

Indeed, as implied by the name.

“ Society for Worldwide Interbank Financial Telecommunication”

At the end of the day, the communication system was being used to make promises, based on an assumption that all the institutions involved in the transfer had the incentive and means to keep their promises. The latter is also being drastically undermined at the moment, making the emergence of an alternative ad-hoc "IOU" system less likely.

Yes, though the closure of dollar clearance services by US and UK banks also stopped the large important payments, at least via the West.

There will still be some facility to do it via China, but China's largest banks are also baulking (mostly because of the rapidly imploding value of the Ruble meaning they don't trust Russian banks to actually send the requisite funds than any great moral imperative I expect by the result is the same), and any Chinese services still willing to co-operate are going to extract a very, very hefty tithe from the Russians to enable it.

The assumption here is that Chinese banks operate with only a profit movie. Things don't work that way in China, the CCP could compel one or more Chinese banks to co-operate if it serves their geopolitical interests.

China would be keen on establishing and testing an alternative to SWIFT & CHIPS, the US stranglehold on financial systems, as they know the US would use it serve her geopolitical interests.

Looking at the number of countries sanctioned, there is no dearth of customers.

This reminds of Borg from Star Trek, more you use your weapons against them, faster they adapt.

CEO and co-founder of Routefusion, a global payments company. In theory yes you could just setup some other form of communicating between the banks, such as phone calls lol. What you will not get is efficiency, trust, etc etc. You dont just replace SWIFT, Ripple, has been trying for a long time, and are still not close.

I think nobody tried because there was simply no need. Now that killing financial access is used as a political tool in domestic and international conflicts, redundant lines could be established. Doesn't work in the short term of course.

As long as such sanctions are only being used against dictatorships that deliberately commit mass murder, I don't think there will be very much motivation to replace it.

Any country could be accused by that, it is the nature of politics. The US has gained access to transaction data with the threat of sanctions against Swift. While banks are naturally very forward with compliance, there might be interest in having redundant channels. Would also benefit the privacy of users.

I get the impression it's not so much that SWIFT itself is critical, it's that so many other banking services and activities are co-ordinated using SWIFT, and being on SWIFT is an indicator of reliability and that you or your organisation is an insider.

That would make sense. It's not the tech that's important here, but the network.

I’d also guess a lot of regulation, policy, insurance, and perhaps law is specific to SWIFT such that you can’t just swap over to another system or simply telephone the transactions.

Recommended talk by Mark van Cuijk from 2016:

A world without blockchain - How (inter)national money transfers works

https://media.ccc.de/v/33c3-8315-a_world_without_blockchain

> When it is time for the banks to actually move the funds (which often times occurs significantly after the funds are made available to the client in question)

Interesting. That makes SWIFT sound like a sort of digital Hawala system.

There are many more analogs to hawala, not specific to SWIFT, as many banks and financial institutions have accounts at other major financial institutions in a variety of countries, upon transfer by one of their clients, the sub-financial institution decreases their balance at the host institution by giving the host institution the money (or reducing a line of credit), and at the receiving host bank they simply increase the sub-institution's balance (or increase of a line of credit).

Final settlement honestly may never occur.

Cryptocurrency projects should be cloning these interbank exchange mechanisms, not making a single global ledger.

This is what many banks and “enterprise” think blockchain will do for them, yes. There are many solutions and some likely in play.

There are some memes about that too. Convoluted arrows between a bunch of banks, next to one arrow between two peers.

That would be called Layer 2. I wonder when cryptocurrency will evolve a SWIFT.

It should be layer 0, not layer 2. Full coordination is not a good default.

Look up how layer 2 blockchains work

> That makes SWIFT sound like a sort of digital Hawala system.

I chuckled, thanks a lot. But numeric money and banking exchange is already based on trust, trust of correct and consistent traceability of exchanges being written by both parties. Hawala is solely based on reputation, and has the very convenient side effect that there is no traceability whatsoever. Doesn't looks like swift to me :)

One thing I always wondered: is there ever a settlement between Hawala agents? It seems like someone could get so unbalanced that they would need to balance out.

Hawala can be out of balance within network. A lot of it was used for remittances etc however the agents had other businesses which helped them keep balance, laundering , high risk or payday loans etc outside the hawala system

Apparently money gets laundered internationally by under-invoicing via front companies such as import/export businesses, travel agents or foreign exchange shops.

https://www.treasury.gov/resource-center/terrorist-illicit-f...

Almost all banking systems divide into "clearing" and "settlement". There is an assumption that, like internet peering, the flows will roughly balance. So at the end of the day if there's $10m in one direction and $9m in the other, that gets settled as either a $1m transfer between the two accounts of the banks at the central bank (the modern, fast way), or if no such relationship exists, through a "nostro/vostro account" (Bank X literally itself has an account at Bank Y), or as an overnight interbank loan.

This was why the 2008 financial crisis was such a risk; it's perfectly normal for banks to lend hundreds of millions to each other overnight to keep the payments system open, at near-zero interest rates because the risk is zero. But what if the risk isn't zero? What if there's a very real risk that your bank lends $100m to a bank that doesn't open the next day and you have to get in the bankruptcy credit line? Payments would have to slow down or stop happening or dramatically increase in cost.

Hence the decision to lend hundreds of millions to every vaguely bank-shaped object to keep the payments system running. It mostly worked, the money was almost all repaid (in fact I think it made a profit in the US?), and incidents of customers queuing at banks or collateral damage to businesses unable to make payments through no fault of their own was minimal.

How does SWIFT play with ISO20022, is that the same thing but different message format?

Also I get it that SWIFT filled a need in the olden days, but with internet+messaging systems in general now available ww why don't banks use something more modern and less expensive? I am not hinting at blockchain. Eg they could use sth like email, telegram, some private incarnation of that, to send their MTx messages, or whatever.

Let me answer your second question - the recent implementation of SWIFT gpi (which is not being widely used by most Russian institutions so not relevant to the sanctions talk) is designed to increase the power of the SWIFT system, better automate payments and ease interaction between SWIFT and clearing houses/payment systems. So in that sense it would be far better than coordinating over a non-payments specific communication method.

Regarding your first question - ISO20022 currently requires a parser to convert ISO20022 information into a MT message. This is automated in SWIFT gpi and compatible softwares.

> Also I get it that SWIFT filled a need in the olden days, but with internet+messaging systems in general now available ww why don't banks use something more modern and less expensive? I am not hinting at blockchain.

Well, I won't be surprised to see blockchains like XRPL, Stellar and Algorand to be part of the ISO 20022 standard [0]

[0] https://finance.yahoo.com/news/iso-20022-cryptos-5-compliant...

Given that ISO20022 (MX) contains thousands of message types already, it certainly wouldn't be surprising to see a few more appear.

In addition to that, SWIFT has something called MyStandards, which basically enables financial institutions to create/share/use their own message formats. Insert XKCD-927.

> How does SWIFT play with ISO20022, is that the same thing but different message format?

The "legacy" SWIFT standard , called MT, was (retroactively) made into an ISO standard itself - ISO15022. Currently it is expected to coexist with ISO20022 on the same SWIFT network for 5 years, no idea which % of traffic is MT and which is ISO20022 though.

So anyway - The SWIFT network can carry either traffic.

As to why the SWIFT is not replaced.... it's not _that_ expensive, if you're a bank, and there is a huge network effect: If all your friends (other banks) are on network X, you really want to be there too, even if you maybe don't like too much how it's run and its cost.

Additionally, it provides a lot of convenience which would be really painful to reimplement - and it would be massively antieconomic to implement it unless you can spread the cost over as many participants as possible (so you're basically trying to create another SWIFT - and there are already plenty of alternatives, though none with the reach and trust SWIFT offers).

Said convenience is, for example, delivery confirmations, ability to replay any message up to 2 months ago, guarantee of every participant (bank)'s identity, non-repudiation. Oh and extreme resilience, which doesn't hurt.

But this misses the main value proposition, which I perceive as the trust in the institution: they've been doing exactly this job for longer than I've been alive, they're a neutral operator, simply carrying everybody's traffic and being truthful about what passes through their pipes. If you send a message through SWIFT, you won't be able to deny that, and the banking community will consider you committed to what you wrote. If you denied that you did, SWIFT would be able to replay the message (even without your consent, if coerced by authorities). And they will not agree to "lose" a message you regret sending - their entire value to the banking community would go to 0 in a moment.

Finally, they're a cooperative - participating banks actually own a stake, and profits just get redistributed (more or less), so it's not like they are sucking an inordinate amount of money away from the banking system and keeping it to themselves.

Most of the institutions doing a good job with 20022 anyway have jumped to gpi which handles it out of the box. I'd say we see about half and half but we work with a lot of smaller institutions that aren't on GPI and aren't particularly sophisticated.

thanks for your insightful answer! that really helps

SWIFT transitioned from X.25 to IP ~ 10 years ago, so they have made changes to keep up with technology. The SWIFT network is now a private IP network. Given the value of transactions that move across the SWIFT network daily, I think that many participants feel more comfortable with the added security that comes with a private network.

Thanks for this clarification. I honestly thought SWIFT was some kind of central coordination authority for all inter-bank payments. I have been bamboozled.

I don't feel like your definition is a bad one per se - just so long as people understand:

1 - there are other ways to coordinate payments!

2 - that the actual fund movement is handled through other means than SWIFT.

I contacted a lot of my Russian colleagues and none are affected by this SWIFT ban and are still being paid by EU companies. They receive their funds from private banks that are not under these sanctions.

I guess they might be affected once Russia starts borrowing the cash from citizens to fund the war.

Why borrow cash when they can just print more money? Same effect.

"Borrowing cash" is slightly less likely to lead to hyperinflation in the short term. "Slightly" and "short term" doing some heavy lifting in that sentence though.

But "slightly" might be useful to the Russian government given the game plan seems to be to try and complete the invasion before the Russian people mass revolt against their own government as the price of bread goes up 10,000%.

This has already started to some degree, starting from today, all russian citizens are banned from sending money over to their overseas accounts, all tax resident companies (and if you are a russian citizen, the only legal way to work for a foreign emplyer from russia is via creating a legal entity which is akin to a 1-man company) must sell 80% of received foreign currencies (this applies to money that was received starting from Jan 1st).

This would be because Russia no longer has that foreign currency, since it was frozen. If 60% of a totalitarian dictatorship's foreign currency is frozen, the dictator is going to make sure that all the foreign currency which the dictatorship does have access to belongs to him.

They may also be affected once the foreign reserves of Russia are frozen, which many of them now are.

I always thought that swift main feature was sorta like a DNS for banks. An hypothetical SWIFT consortium would assign unique short names to banks so to provide a baseline for authentication and identification.

This only based on the swift codes you have to add sometimes next to the IBAN.

That would make sense in the context of payments like SEPA, etc. where a unique code helps limit mistaken transactions - however, that doesn't mean SWIFT itself is actually being used beyond the identifying codes banks use on the system.

Where a SWIFT MT103 would come in is typically (but not always!) much larger transactions, such as moving $100m from a corporate account in one country to another.

https://news.ycombinator.com/item?id=24103753

Decades after a lot of the features in the new system should have existed in the US but I'm not going to look a gift horse in the mouth. Thanks for linking!

This posts comments is full of either people that don't have a clue or bots serving some agenda; a good solid 95%+ here is just noise.

Why even bother reading hacker news when its like this, or even commenting with any real substance.

Read some history (from multiple sources), educate yourselves, and stop feeding the noise. If you do that, you'll learn some useful things like how detente works, geopolitics, why the cuban missile crisis was such a big deal, etc. Those lessons directly inform events today.

For those wondering why there aren't any SWIFT employees chiming in:

- They barely have internet access at their Office, it's a big hassle and it's all recorded.

- Very strong security culture, both digital and physical.

- The company actively monitors the internet for any mentions of SWIFT. Lot's of stories of people being threatened with lawsuits for breaking NDA. Even if they detail the technologies used on LinkedIn for example.

Lots of Security by obscurity, as usual. )

Obscurity is of course not sufficient on its own, but it still an additional layer.

I have always wondered, what combination of manual vs. automatic processes are happening when one makes a bank wire (actual gross settlement / instant wire transfers, not ACH for example in the US)?

Because whenever I have done wires, in addition to the bank/IBAN/account numbers, there are all these manual fields like "for further credit to", "for benefit of", etc. that are clearly having to be interpreted by someone. And very often errors come up and the funds are held or in limbo on one end or another.

Does anyone have a good explanation of what happens both on the initiating bank side and receiving bank side when a wire is done?

Does someone have to manually review an outgoing wire to allow it to proceed? Does the receiving bank have someone to review it inbound and route it correctly?

I imagine that this is why it still costs $25-40 to handle them.

Are you talking domestic US wires or international wires?

Oh, not too sure how different they are, but definitely even with domestic wires, many custom/human-review-required fields seem to exist.

With international wires, I am sure that even more complications exist with which currency it's to be settled in, etc.

And I have no idea/am curious whether the information about ownership of the account is transmitted with the transaction, to track whether the same person owns both accounts, etc?

Domestic US wires are (with the exception of flagged or manual telephone Fedwire placements) automatic through the Fedwire system, which is a RTGS (real time gross settlement) system.

Most Euro denominated payments (not sure I'd call them wires anymore) are also automatic, cheap and easy thanks to SEPA's wide proliferation.

Most of the time the extra info not actually required for automatic processing is in case the payment doesn't go through as intended, as well as for compliance reasons. As Elric notes above, international payments (outside a common currency zone with a system like Fedwire/SEPA) is where things get tricky/manual processing often comes into play.

I can't imagine it has a lot of manual work involved, exactly for the same reason you do: I haven't paid a cent for international or domestic wires in years. I'm in the EU, but I've done wires to Swiss or Lebanese accounts as well and no fees anywhere.

Maybe it's local regulations defining whether a human should be part of the loop?

Sadly that's very much a European thing. I never pay much or anything from my European accounts but even our very large, corporate client account stateside we pay I want to say $15 for domestic, $40 for international?

(International) payments frequently involve dozens of systems. Some steps are manual, even if you're not being charged for it. This is especially true if somewhere along the way the transaction is flagged as being suspicious (which can happen for any number of reasons, including innocuous ones).

Never underestimate how antiquated US infrastructure can seem lol. But to be fair European banks were always motivated to develop robust protocols for international bank operations due to the number of countries in Europe, even more so with the advent of the EU. US banks could talk to US banks, and I'd guess Canadian and Mexican banks would have just had to put the effort in to being compatible with the US system.

My armchair guess would be that for most banks and transfers, the happy path involved no manual human interaction but that there are various arbitrary situation where a human gets looped in. One of my banks seems super paranoid and have contacted me asking for supporting documents for each and every incoming international transfer >~800 USD (despite most of them being from the same account, held in my name).

I have worked in this business for 20 years at some of the largest banks, so know something about this. In short, yes, at the big banks, most wire processing is automated. Straight through processing rates are > 90%.

An explanation of what happens is at initiating and receiving banks is beyond the scope of anything I want to type.

At banks with high volumes, more processing is automated (see above) for low volume banks, not so much. Payments may fall out of automated processing for several reasons like poor instructions quality (unknown account number), complex correspondent/routing requirements, sanctions screening (OFAC in the US). For high volume banks, even with high STP rates, this can mean 1,000's of payments per day.

As for why wires cost so much, the answer depends on who is paying. The most frequent users of wires often pay very little for them because bank's provide wire services as a package in exchange for holding balances with the bank. Only retail customers and small businesses pay full price ($25-40) for wire payments, and bank's charge this much because they can. Retail is such a small part of the wire payments business that if you take you business elsewhere the bank does not really care.

One last note on pricing. Most wire originate from the same set of bank customers on a regular basis, so their processing is highly automated and the bank's risk and fraud systems are trained to expect these payments. Occasional payments from retail customers probably do require manual processing more often, so a higher price is probably warranted to some extent.

Have a look at Ukraine's military crowdfunding instructions, because it demonstrates a non-seamless transfer of money.

As a user outside of Ukraine you cannot send money directly to accounts in Ukraine (unless your bank handles it specially). However, what you can do is send money directly to the Ukraine central bank (!!!) and in the memo field, specify which account should be credited.

That is how I assume all interbank transfers would operate without SWIFT. Couldn't tell you why Ukraine uses this system since they're not banned from it.

I'm not an expert in this area, but from what I have gathered with respect to domestic Fedwire based wire transfers:

Fedwire is arguably the main settlement system used in the US between banks that participate in the federal reserve, when correspondence accounts are not in use. (Even other systems that claim to be settlement systems seem to often end up using FedWire for final settlement). The actual settlement is simply a debit and credit in the bank's accounts with the Fed. But while that does technically settle payment, in practice more information is needed, as the bank's accountants need to know what this was for, so they can keep their accounts up to date. Therefore metadata can be attached to the transfer. Without that it would be more complicated to figure out what exactly just got settled.

But when one is trying to send money not just between banks, but between customers of banks, even more metadata is generally needed. You are correct that at least in theory, the field for the target account would clearly who is getting paid. (Obviously the org receiving the wire will often generally need some additional metadata about what the payment is for, in much the same way banks do, hence the use of fields like 4320.)

I suspect that the receiving bank's handling is usually mostly automated (especially with larger banks), but that there are some safeguards that sometimes stop things for manual review both for correctness concerns (like making sure the name of beneficiary is even remotely close to the name associated with that account number) and for security considerations (flagging suspicious transactions to have further verification). And of course, if things are incorrect and things are not caught by the safeguards, it requires human effort to fix.

The receiving organization's steps seems to often be at least somewhat manual. (But this can vary, orgs that receiving thousands of transfers a day likely automate more than those that receive just a few.) I get the feeling some organizations with the capability of automatically processing most incoming wires simply don't, because having a person look at the incoming wire, and review what the automated processing will do before clicking an "approve processing" button can catch other mistakes, some of which could be a pain to unwind manually.

> I feel like the media has done a very poor job explaining what SWIFT is

The media, in general, does a very poor job explaining anything and everything, but due to the fact that the vast majority of people have no clue whatsoever, nobody ever really notices.

That's why it's so easy for mass media to manipulate people. Nobody ever really notices that either.

Except for people who actually know, of course. Like yourself.

But then you, OP, you, turn the page, keep on reading and believe everything else that's written:

https://effectiviology.com/knolls-law/#:~:text=Knoll%27s%20l....

Where can you learn more about this stuff?

For example, I've always been curious how Swift is routed at the lowest layer. Does it go through the normal internet? The telecommunications infrastructure? Custom fiber-optic cables connecting almost every bank in the world to some kind of central office?

It's just VPN and MPLS connecting banks with SWIFT.

Rumor is, they still keep ISDN gateways operational as well ;-)

I believe that SWIFT is now a private IP based network. That means it runs on dedicated comm channels, but uses the internet's IP.

SWIFT is:

50% payment messages

43% securities messages

The rest are a mix with a lot being treasury transactions.

I work on the investments side and use SWIFT for security transaction settlements. For example, when we trade a security (stock, bond, derivative, etc) we send an ISO 15022 formatted message over SWIFT to the custodian telling them to expect delivery (for a buy) or to deliver (for a sell) a particular security on the settlement date.

If we don’t give them this message then their settlement group will DK any securities being sent to them that don’t match a notification. (DK = Don’t Know) which means the trade will fail and you’ll have some really pissed off people that will want you to make them whole if they lost money as a result.

SWIFT is used for lots and lots of things, not just payments.

I've heard SWIFT called "the WhatsApp of the banking world".

It allows for efficient communications between parties. You can use other forms of communication to arrange for the movement of money, but they're more of a hassle.

And many institutions do not support them because why would they - everyone is on SWIFT except for murderous dictatorships.

> I feel like the media has done a very poor job explaining what SWIFT is and how exactly it works.

Strongly disagree. The media are for laypeople. You wouldn't appreciate complex breakdowns about how proteins in SARS-CoV-2 work, too much depth.

I also disagree with the implication that a SWIFT ban is meaningless.

SWIFT is a notification of intent to settle. Modern banking (read: the electronic age) moves MUCH faster than money does. You can see this in ForEx. By the time the money arrives, it's worth something else. This is risk exposure and is partly why most banks like to backout into their native currency. Not always the case, but still a useful line.

Banks exist to provide 2 major services.

1: Trust. Doesn't matter what so-and-so claims, their bank are the ultimate gate keepers. If the bank says it's paid, that's a legally guaranteed promise.

2: Liquidity: You actually need to have money to be able to send it. Banks will let you use "their" money while your money is still actually arriving (settlement).

As you rightly pointed out: Banks make this money available to you before it actually arrives. The only way they can do that is because of #1: They trust the settlement will happen.

If there's no trust, there's no releasing of funds. If there's no SWIFT message, there's no incoming settlement. TARGET2 can be used without SWIFT, but good luck finding a lot of banks that will handle CIPS in place of SWIFT.

But by FLATLY removing Russia from SWIFT, they have successfully killed transactions in BOTH directions. Money can't come in, money can't go out.

There are arguments to be made about the efficacy of these decisions (nobody believes the vast majority of Russians voted for Putin, right? Cuz they certainly did not so pressuring them to once again not vote for Putin seems pretty stupid, but the west can't do literally nothing. Difficult problem) but this was an effective move to further crush the ruble.

Let me address the last paragraph of your comment.

Vast? No. Majority? Most likely.

You do not need to take my word for it, you can look for 1st hand accounts from people who were P. election observers from the opposition.

I'll at least agree that nobody believes that the party in power had the majority votes nationwide in the 2 most recent national elections.

As for the current events, most recent official (official being the key word here) nationwide polling shows 68% support for the "special war operation", which I do not believe for a second, considering how many have either relatives, friends, or at the very least business contacts across the border.

Please could someone explain me (ELI5) how this all works with a very practical example for a layperson to understand? Say, I want to transfer £500 from the UK to India. How will this transfer work in detail behind the scene?

1. Well, doesn't this make SWIFT sanctions the less important part, relative to getting "clearing houses" to ban transactions from a certain party?

2. With SWIFT in place, why do banks even need clearing houses? Is that merely to save the need for bilateral verifications that transfers have been registered?

3. At the bottom line, the "transfer" is just an agreement to decrease a fixed-with decimal value on one side and increase it on the other... we're not talking about physical movement of paper or bullion, right?

4. It's still a problem that SWIFT lets the EU (or EU+US) turn unilateral sanctions into global sanctions.

1. Yes but for said clearing house transactions to occur you would need to get counterparty institutions to agree to another communication method for coordinating purposes. The loss of the already agreed upon oracle is problematic.

2. A number of reasons but mainly to do with how funds interact with central banks and their capital requirements. Your later point that funds don't PHYSICALLY move means that it's less about each bank recording the appropriate addition and subtraction (that's easy enough) and more about how it is recorded, so transactions can be properly audited and transfers that would cause systemic or institutional issues can be curtailed as much as possible. There's also the issue of banks being able to more easily manipulate books using off ledger funds if Central Banks/payments systems weren't involved but that's again a systemic risk issue and not really relevant to our discussion.

3. Correct.

4. I agree but I don't think that's going to improve any time soon and I've yet to see a system that doesn't trade one point of centralized failure for another. I do hope cryptocurrencies can solve the distributed exchange issue because that's actually one very attractive (potential) feature of crypto for me, as someone who has seen first hand how these systems can be abused to target people - deservedly or not.

"I feel like the media has done a very poor job explaining what SWIFT is and how exactly it works."

Could it be replaced by Email+GPG? Or ZeroMQ with Curve crypto?

As long as every institution that's required to make the payment happen buys in (and it is compliant with whatever regulations are involved for sender, recipient and intermediate institutions) one does not have to use SWIFT explicitly. That said, gruturo above (in our discussion about the associated ISO standards) makes a good case for why most banks default to SWIFT.

So banks refer to SWIFT for instruction on how much to transfer where? And anyone who hacks SWIFT will be able to fake fund transfer instruction?

One would need some inside baseball knowledge to effectively hack SWIFT. Not only do you need 1 (or more) BO or Senior BO PINs to verify the send, a hacker also needs to properly format the send without including any details that will tip the recipient off that it is fake.

That said, things are changing significantly with SWIFT gpi payments and the much more automated nature of the new system may actually make techniques such as fake sends more, rather than less, effective until solid mitigation strategies are attached.

> Not only do you need 1 (or more) BO or Senior BO PINs to verify the send

That doesn't sound impressive at all by modern standards. At the very least they should have some kind of challenge-response scheme.

That would be easily implementable in a bank's SWIFT gpi software as that system gains traction. Probably doesn't add any value in the current system as the recipient as no way to confirm the challenge-response scheme was actually completed as would be implied (beyond, perhaps, coordination between sending and receiving institutions outside SWIFT proper). Although I could be simply not sufficiently creative enough.

The bigger point I should have made is that most send fraud doesn't actually occur on SWIFT and is conducted using falsified documents to give the impression of a correct send, typically to get funds released before the fraud is revealed. It relies much more on social engineering than any kind of actual systems hacking skill.

Plus:

There are some additional hardware/special componentes one has to use to connect to SWIFT, those boxes arrive pre-configured at your datacenter.

Yes, thank you - I should have been clear that SWIFT terminals are kind of like Bloomberg terms in that they're not going to be easily cloned at home. Better off attempting the social engineering tactic for most criminals.

Security by obscurity

IT, even when huge sums are spent on it, is still seen as a cost center (rather than a competitive advantage) at the vast majority of banks.

However, as the other reply said to you... I've really not seen any evidence the SWIFT system isn't decently well constructed. When attacks (such as Bangladesh) have happened it has been due to not following best practices as established by SWIFT and other institutions.

Well, no, they are just controlling the whole line of delivery for the data packages; its not uncommon in critical infrastructure, that you have to deploy some special hardware stuff for whatever security reasons.

For sure, the typical HN-mentality is: "there is no security and since this a dumb bank/financial-service, they are just trolling and they dont know what they are doing" - no, let me tell you that you are wrong with this assumption: SWIFT is pretty secure and there haven't been any larger (successful) attacks on the network itself (hint Central Bank of Bangladesh losing 90m in a CEO-scam is not a problem of SWIFT, same for similar cases)

> is not a problem of SWIFT, same for similar cases

Perhaps it was not in SWIFT's domain of responsibility. But it was for sure a problem for them and it's why they started CSP.

Memories of Manually Initiated Fund Transfer training come flooding back!

I'm sorry for putting you through that.

> And anyone who hacks SWIFT will be able to fake fund transfer instruction?

Yes.

https://en.wikipedia.org/wiki/Bangladesh_Bank_robbery

Yes, although Bangladesh' central bank was notoriously lax in their operation of their SWIFT handling. This was well known and a big reason why some of our clients would not use Bangladeshi recipient banks due to a lack of trust in the central bank.

There is a Darknet Diaries episode about this heist, which is very recommendable as most of this podcast's episodes: https://darknetdiaries.com/episode/72/

> And anyone who hacks SWIFT will be able to fake fund transfer instruction?

More common is to hack any bank connected to SWIFT, and send messages on the SWIFT network over that bank's SWIFT terminal.

If SWIFT is the messaging system and STEP2 is the clearing house, why is SWIFT-only versus SWIFT-and-STEP2 being sanctioned?

Really no need to sanction STEP2 or other payment networks at this point as central banks/intermediary or recipient banks themselves have also been instructed as well (either formally or informally) not to engage with Russian institutions. They're not processing payments with the vast majority of the global financial infrastructure at this point.

You didn't actually think they would really hurt the interests of their true patrons - transnational corporations, big industry, high finance ?

This SWIFT ban only really affects the usual IBAN to IBAN transfer of the plebs and SMB's. Large payment processing and gross settlement transfers will just change to accomodating intermediary banks in China within a couple days, and that'll be that. Further fund movement will be done through the Russian-Chinese CIPS.

Never forget EU commissars are not elected officials, they are named by an totally informal group called "Eurogroup", to which supposedly only 15 EU presidents participate. However, as it's not formal, it's completely opaque how they get chosen.

CIPS is three orders of magnitude smaller than SWIFT. Even for large institutions that reduced network size introduces tons of overhead & risk to their transfers. You don’t just pivot on a dime with bank corridors.

But the other sanctions are an important part of the mix as well. No matter how you receive your payment instructions if your local government has banned the money movement it won’t go through (modulo money laundering & dodgy banks). This is we’re Chinas lack of sanctions might help. But what do you do as a large organization with the money in a Chinese bank now? You still can’t pay for EU goods with it.

The sanction mesh very effectively makes “high finance” for Russian corporations and international corps doing business there slower and more risky. Will it touch the trillion dollars Putin and gang have looted & hidden? Probably not but what could?

If you want to complain about something the carve out for energy financing is the big one. It shows that the EU wasn’t really ready to sacrifice to hurt Russia. Which is understandable in a real politik way.

> if your local government has banned the money movement it won’t go through (modulo money laundering & dodgy banks)

That's my point and OP's : they didn't prevent it, just the messaging through SWIFTnet. TARGET2 and STEP2 still fully work.

Also, you can transfer batches of messages through FTP cronjobs and automated import at night, like it was done 20+ years ago. It just slows settlement from 1-2 days on average to 4-7 business days, depending on the correspondent bank.

Also, you can easily automate use the plenty free MT103 fields for additional routing data ex-SWIFTnet.

It's something which is often done - source : been there, done that, in my youthful days.

> “high finance” for Russian corporations and international corps doing business there slower and more risky.

Yes, anything in regards to transnational money. More expensive, too, but what's that in this inflationary world. A drop in the bucket of Moloch.

> Never forget EU commissars are not elected officials, they are named by an totally informal group called "Eurogroup", to which supposedly only 15 EU presidents participate. However, as it's not formal, it's completely opaque how they get chosen.

EU commissars are not directly elected officials, no, but they are also not "named by a totally informal group". 'EU Commissars' (typically called an European Commissioner) are nominated by their member state, one each. The European Parliament then votes on the whole. So they are still under democratic control: through the various chosen governments of the states, and through the EP.

The Eurogroup -- a discussion body for the 19 states that use the Euro as currency -- had nothing to do with the SWIFT ban. (it does have a bunch of issues from a democratic perspective, although it seems like most of the members are just their country's finance minister... which makes sense for a discussion forum about the Euro?).

Mostly agree with what you wrote with the following precisions :

> "named by a totally informal group"

I used the term "informal" from the legal point of view. I'll have a box of cookies of your choice delivered to your home if you can find me the treaty basis, or any other legally formal basis for its existence.

> "EU Commissars"

That was tongue-in-cheek from me, as related to Bolchevik buraucratic komissars - lovers of all things centralized.

> had nothing to do with the SWIFT ban

The Eurogroup didn't have it, correct, but the people it chose did very much so.

Similarly, what's that about waging war, which happened in the same bureaucratic breath yeasterday ? I don't remember J.Borell or Ch.Michel having been chosen as commanders in chief.

> So they are still under democratic control [...] (it does have a bunch of issues from a democratic perspective)

In the end, I'm partly Swiss, so I abhor my democracy being representative with a separation degree larger than n=1, or directly myself.

And what exactly is "money" and how is it "transfered" via messages?

Time for russia to stop playing by the wests rules. This is a watershed moment for crypto and will be pointed to as a historical example of when the US started to lose influence over the world stage.

what russia is doing is terrible and should be stopped. But this incident is going to show that the US and the EU can no longer control the worlds trade. technology has evolved past these financial systems mared by corruption, false data and glacially slow payment transfers. Who needs swift, clearing houses and forex exchanges when you have Terra?

Side tangent, the first nationstate to adopt crypto will have the biggest benefits.

Yes, because the administrations of authoritarian warmongers absolutely LOVE difficult to trace currencies. I'm sure Putin himself is setting up a Coinbase account as we speak...

Yes actually.

China is spending alot of time rolling out its digital currency among it's populace. What is difficult to trace about it? It's literally a public ledger...? If you institute a state controlled wallet, not only can you control flow, color of money, etc. but you can perfectly see the data in real time which is a tremendous improvement over the current infrastrucure. You can actually answer questions like, how much does the average citizen spend a month on XyZ with a simple query among your wallet.

> I'm sure Putin himself is setting up a Coinbase account as we speak...

You are cut off from swift and other internationally accepted banking measures. How do you get foreign currency into your nation and how do you transfer payments internationally? Crypto.

yes, let's use the attempted destruction of a sovereign nation by a neighboring dictator as an opportunity to promote your favorite cryptocurrency

War is nothing new. Sanctions are nothing new. The ability to evade sanctions via crypto is new. Simply pointing out that these sanctions are useless and if you use etherscan, you will see a huge rise in usage in russia and ukraine right now.

> This is a watershed moment for crypto and will be pointed to as a historical example of when the US started to lose influence over the world stage.

Just a month ago China and Russia made a 30-year deal to sell/buy gas through a pipeline. Settled in Euro!

https://www.reuters.com/world/asia-pacific/exclusive-russia-...

Russia wouldn't trust the yen. People in China can't freely exchange to other currencies (or move money out of the country). The yen is being held together by a "captive audience" (or a subjugated people).

And China won't trust the Ruble, because well, it's obvious, right?

The idea that you'd make a 30-year deal settled in bitcoin/eth is plain crazy. Trust is hard to build (it takes time), when neither Russia or China has that level if trust, what makes you think crypto would get there tomorrow?

I never said anything about trade deals.

Russia is being blocked from the wests financial system by making harder to transfer funds. Crypto can make international transfers.

This was just an example that trade between countries like Russia and China often happen in dollars (or sometimes Euros). Not yen or rubles.

You think they would use bitcoin/eth instead?

I don't fully understand banking systems, but if China buys something from Russia in dollars. They are not paying with truck loads of cash. They might pay with IOY in a Chinese bank. But Chinese banks might not want that liability (buying dollars could be expensive in the future; and certainly Russians won't trust a Chinese bank.

They are probably moving dollars between a US account belonging to a Russian bank, and US account belonging to Chinese bank.

Moving large amounts of dollars (or euros) by other means is very hard.

Gold, bitcoin, eth, etc. are all more volatile than dollars (or euros), and harder to move.

Actually, this should be bad for crypto's reputation, because crypto has no way to cut off bad actors that everyone wants to cut off.

Reply on news.ycombinator.com