7 ms·

Out of curiosity, are there any reasons to prefer slower settlement? (Since this seems like a no-brainer, but clearly hasn’t happened already)

dan-robertson 6y ago
Some ability to net—that is to buy at one price, sell at another, and only settle the difference—is probably desirable
- freeone3000 6y ago
  Sure, but end of day settlement (like is practiced on TSX!) has all of the reasonable benefits of T+2 settlement, while still taking place in an understandable time frame.
  - throw0101a 6y ago
    If TSX = Toronto Stock eXchange, I do not think that is correct: * https://www.tsx.com/news?id=533&year=2017 https://www.tsx.com/news?id=533&year=2017 In Canada it's done by CDS Clearing and Depository Services Inc. (CDS) and is currently T+2 just like in the EU (2014) and US (2017). A 2015 whitepaper discussing the (then) proposed change: * PDF: https://www.cds.ca/resource/en/174 https://www.cds.ca/resource/en/174
tomatocracy 6y ago
There are reasons to prefer settlement is not instantaneous - eg it makes fat finger trades easier to unwind. There are also probably some special cases like creation/destruction of ETF shares, ADRs, IPO greenshoe etc which take a bit longer to work through. But I think the main reason for these changes being slow to occur is because noone wanted to break the system in the process - the settlement process for share trades in the US is/was viewed as quite fragile due to the way it had evolved over time.
- atombender 6y ago
  Adding support for fast settlement doesn't necessarily mean removing support for slow settlement. The customer placing an order could simply mark trades as "fast" or "slow".
  - totalZero 6y ago
    ...And then we would need arbitrageurs to keep the fast NBBO in line with the slow NBBO, so that people in the two pools can trade with each another.
    - atombender 6y ago
      Why? This only affects settlement, not execution. Execution is already "real time".
      - totalZero 6y ago
        Because the trade has two sides that have to agree on settlement terms.
Barrin92 6y ago
I'd say the first reason is staring us right into the face, it disincentivizes actors on the market from engaging in exactly the kind of nonsense that is divorced from fundamentals we've seen over the last few weeks playing out on Robinhood.
- Miner49er 6y ago
  Was that not just the market correcting itself? The stock was extremely over shorted, and the price rose to shake out the shorts. I think short squeezes need to be allowed to happen properly. If we don't want that, we would need to limit short selling.
  - tylerhou 6y ago
    Why do you think the stock was “over shorted” apart from the fact that short interest is generally not that high? Or — what do you think are the bad consequences of “over shorting?”
    - Miner49er 6y ago
      > Why do you think the stock was “over shorted” apart from the fact that short interest is generally not that high? It was the highest shorted stock on the market. That is "generally not that high"? I think the fact that retail investors were able to cause a short squeeze on it, that cost shorts many billions of dollars, is an indication that it was over shorted. > Or — what do you think are the bad consequences of “over shorting?” There needs to be balance to everything; one of the naturally occurring risks of shorting is a short squeeze. Messing with that would ruin the risk/reward balance of shorting. Shorting creates more longs, and therefore drives the share price down. The risk of a short squeeze is a good way to prevent people from opening an extremely high number of shorts, and artificially driving the price down this way.
      - tylerhou 6y ago
        Short squeezes are not good for the market, period, but they only happen in extraordinary circumstances. And actively encouraging people to cause short squeezes in order to reduce the potential for future short squeezes seems... counterproductive. If you want to reduce the number of people shorting a stock, creating artificial short squeezes would work, yes, but then I again ask: why do we want to prevent people (or hedge funds) from shorting a stock so much? What does that accomplish in the bigger picture? The logic is almost like: we want there to be fewer car crashes, and more cars = more crashes. If we ourselves cause crashes, drivers will be afraid to drive, so therefore we will have less crashes.
KaiserPro 6y ago
yes, because there are two parts to the trade: 1) the contract 2) the paying of the money There are ancillary bits as well like unwinding of mistakes. The problem is this, moving money fast is hard. It can be done, but its far cheaper to settle up at the end of the day, when the total inflow/outflow is calculated. But crucially there isn't enough liquid capital to service this level of actual transaction.
totalZero 6y ago
Slower settlement gives the exchange and clearinghouse time to process the trade, it gives firms time to borrow cash overnight in the repo market to pay for trades made by the front office, it gives short-sellers time to borrow and deliver shares, and so forth. T+1 means overnight (options settle this way), so T+2 gives everyone in the world (sometimes you trade from one region for a legal entity in another) a little more than a full business day to tie everything out. I see delayed settlement as part of the reason why we are able to trade instantaneously; execute first and address the details later. EDIT: Also, the way equity trades presently clear is based on a net short and a net long, rather than breaking out individual trades as would be necessary for instantaneous execution.
- barnabee 6y ago
  These are terrible reasons. Time to process the trade should be zero if systems are designed properly. If you need to borrow cash to trade you should do it first, etc. If some participants want to stay slow, bear more risk, require more capital, etc... let them. But the rest of the financial system shouldn't be held back because of them. Time for instantaneous settlement, across the board. This will never happen if we leave it to banks, exchanges, clearing houses and the rest (some of them even have a vested interest in it not happening). This is one of the clearest use cases for blockchains [1]. It answers the problem of how next gen financial infrastructure will get built, and frankly the savings and benefits to be had are astronimical. [1] These will probably be public blockchains, but it's possible a "proof of authority" real time settlement chain or similar operated by a consortium could work, too, as an interim step/for some cases
  - totalZero 6y ago
    A San Francisco approach to a New York problem.
    - barnabee 6y ago
      In the UK we do the tech and the finance in the same city. My observation is based on over a decade working on trading and settlement systems at multiple major international banks, exchanges. The infra is creaking at the sides, it is too expensive to fix individually. We're talking hundreds of millions, even billions, at every institution, that they internally decide, repeatedly, that they cannot spend. Every so often a bunch of them try and work out how to build an industry utility for faster, better settlement but can never get past the politics. So it won't change in a big way, not from the current vantage point. But the world is changing. More countries and industries getting into the financial system, changing retail demand, more complexity for small businesses working internationally, etc... so I don't think it ends with "it's too hard to properly fix this" so let's not. I am pretty convinced at this point that the world will fix it, as I've described, but from very different starting points to major New York and London institutions, and they'll drag the rest of the industry kicking and screaming along with them at some point... once there's money to be there in size, the big banks and funds will follow.

dan-robertson 6y ago

Some ability to net—that is to buy at one price, sell at another, and only settle the difference—is probably desirable

freeone3000 6y ago
Sure, but end of day settlement (like is practiced on TSX!) has all of the reasonable benefits of T+2 settlement, while still taking place in an understandable time frame.
- throw0101a 6y ago
  If TSX = Toronto Stock eXchange, I do not think that is correct: * https://www.tsx.com/news?id=533&year=2017 https://www.tsx.com/news?id=533&year=2017 In Canada it's done by CDS Clearing and Depository Services Inc. (CDS) and is currently T+2 just like in the EU (2014) and US (2017). A 2015 whitepaper discussing the (then) proposed change: * PDF: https://www.cds.ca/resource/en/174 https://www.cds.ca/resource/en/174

tomatocracy 6y ago

There are reasons to prefer settlement is not instantaneous - eg it makes fat finger trades easier to unwind. There are also probably some special cases like creation/destruction of ETF shares, ADRs, IPO greenshoe etc which take a bit longer to work through. But I think the main reason for these changes being slow to occur is because noone wanted to break the system in the process - the settlement process for share trades in the US is/was viewed as quite fragile due to the way it had evolved over time.

atombender 6y ago
Adding support for fast settlement doesn't necessarily mean removing support for slow settlement. The customer placing an order could simply mark trades as "fast" or "slow".
- totalZero 6y ago
  ...And then we would need arbitrageurs to keep the fast NBBO in line with the slow NBBO, so that people in the two pools can trade with each another.
  - atombender 6y ago
    Why? This only affects settlement, not execution. Execution is already "real time".
    - totalZero 6y ago
      Because the trade has two sides that have to agree on settlement terms.

Barrin92 6y ago

I'd say the first reason is staring us right into the face, it disincentivizes actors on the market from engaging in exactly the kind of nonsense that is divorced from fundamentals we've seen over the last few weeks playing out on Robinhood.

Miner49er 6y ago
Was that not just the market correcting itself? The stock was extremely over shorted, and the price rose to shake out the shorts. I think short squeezes need to be allowed to happen properly. If we don't want that, we would need to limit short selling.
- tylerhou 6y ago
  Why do you think the stock was “over shorted” apart from the fact that short interest is generally not that high? Or — what do you think are the bad consequences of “over shorting?”
  - Miner49er 6y ago
    > Why do you think the stock was “over shorted” apart from the fact that short interest is generally not that high? It was the highest shorted stock on the market. That is "generally not that high"? I think the fact that retail investors were able to cause a short squeeze on it, that cost shorts many billions of dollars, is an indication that it was over shorted. > Or — what do you think are the bad consequences of “over shorting?” There needs to be balance to everything; one of the naturally occurring risks of shorting is a short squeeze. Messing with that would ruin the risk/reward balance of shorting. Shorting creates more longs, and therefore drives the share price down. The risk of a short squeeze is a good way to prevent people from opening an extremely high number of shorts, and artificially driving the price down this way.
    - tylerhou 6y ago
      Short squeezes are not good for the market, period, but they only happen in extraordinary circumstances. And actively encouraging people to cause short squeezes in order to reduce the potential for future short squeezes seems... counterproductive. If you want to reduce the number of people shorting a stock, creating artificial short squeezes would work, yes, but then I again ask: why do we want to prevent people (or hedge funds) from shorting a stock so much? What does that accomplish in the bigger picture? The logic is almost like: we want there to be fewer car crashes, and more cars = more crashes. If we ourselves cause crashes, drivers will be afraid to drive, so therefore we will have less crashes.

KaiserPro 6y ago

yes, because there are two parts to the trade: 1) the contract 2) the paying of the money There are ancillary bits as well like unwinding of mistakes. The problem is this, moving money fast is hard. It can be done, but its far cheaper to settle up at the end of the day, when the total inflow/outflow is calculated. But crucially there isn't enough liquid capital to service this level of actual transaction.

totalZero 6y ago

Slower settlement gives the exchange and clearinghouse time to process the trade, it gives firms time to borrow cash overnight in the repo market to pay for trades made by the front office, it gives short-sellers time to borrow and deliver shares, and so forth. T+1 means overnight (options settle this way), so T+2 gives everyone in the world (sometimes you trade from one region for a legal entity in another) a little more than a full business day to tie everything out. I see delayed settlement as part of the reason why we are able to trade instantaneously; execute first and address the details later. EDIT: Also, the way equity trades presently clear is based on a net short and a net long, rather than breaking out individual trades as would be necessary for instantaneous execution.

barnabee 6y ago
These are terrible reasons. Time to process the trade should be zero if systems are designed properly. If you need to borrow cash to trade you should do it first, etc. If some participants want to stay slow, bear more risk, require more capital, etc... let them. But the rest of the financial system shouldn't be held back because of them. Time for instantaneous settlement, across the board. This will never happen if we leave it to banks, exchanges, clearing houses and the rest (some of them even have a vested interest in it not happening). This is one of the clearest use cases for blockchains [1]. It answers the problem of how next gen financial infrastructure will get built, and frankly the savings and benefits to be had are astronimical. [1] These will probably be public blockchains, but it's possible a "proof of authority" real time settlement chain or similar operated by a consortium could work, too, as an interim step/for some cases
- totalZero 6y ago
  A San Francisco approach to a New York problem.
  - barnabee 6y ago
    In the UK we do the tech and the finance in the same city. My observation is based on over a decade working on trading and settlement systems at multiple major international banks, exchanges. The infra is creaking at the sides, it is too expensive to fix individually. We're talking hundreds of millions, even billions, at every institution, that they internally decide, repeatedly, that they cannot spend. Every so often a bunch of them try and work out how to build an industry utility for faster, better settlement but can never get past the politics. So it won't change in a big way, not from the current vantage point. But the world is changing. More countries and industries getting into the financial system, changing retail demand, more complexity for small businesses working internationally, etc... so I don't think it ends with "it's too hard to properly fix this" so let's not. I am pretty convinced at this point that the world will fix it, as I've described, but from very different starting points to major New York and London institutions, and they'll drag the rest of the industry kicking and screaming along with them at some point... once there's money to be there in size, the big banks and funds will follow.