Overview
In this episode, industry experts John Pucciarelli, Stuart Smith and Will Thomey explore key market developments shaping the financial landscape in 2025. The discussion covers the evolving ISDA SIMM™ model as it shifts from annual to semi-annual updates, and the impact on firms’ workflows.
Beyond ISDA SIMM™, the conversation delves into major industry priorities, including:
- The rise of tokenisation and its potential to revolutionise securities transfers
- Critical shifts in repo and treasury clearing as regulatory mandates in the U.S., Mexico, India, and China drive significant operational changes
- Upcoming UMR and repo clearing requirements, emphasising the need for firms to proactively prepare for implementation.
Packed with valuable insights, this episode is insightful listening for financial professionals navigating these industry-wide transformations.
Listen to the podcast
Hello everyone, and welcome to another episodeof Ahead of the Curve. Thank you for joining us today.With us again, we have our co-heads or heads of business developmenthere at LSEG Post-trade Solutions. Stuart Smith.Hi there. -Will Thomey.Hello. -Thanks for joining us today.Stuart, you and I have been doing this together for quite a while.A few times. -Quite a few times.Will, I know this is not your first podcast,but I think this might be the first time you and I have done this.In this studio. -In this studio, especially.Today, what we want to talk about and what we like to doaround this time of the year, as we record this, it's January 2025.We had a big, busy year, 2024. What we're going to do for this podcast,anyway, is to look ahead to this year. What are some of the big topicsthat collectively we're looking at from a business perspective?What are some of the interesting things that are on the horizon, maybe,and really the things that you're both focused on for this year.We'll have a good conversation. We'll throw it back and forthbetween both of you, but I'll start first with Will.Will, I know your main focus is workflow. Stuart's is the risk area,but obviously, they all kind of co-mingle together.As we're bringing post-trade solutions together as one cohesive business,it's even more important this year than ever.However, we will look at things in a siloed wayish,and then we can talk about other things and how they fit together.I guess my first question to you is really, what is the outlookfor 2025 from your perspective? In the workflow space, I know we've talkedabout things like settlement, netting and tokenization,and a few other things, and I'll bring up UMR,which is always near and dear to my heart, in a little bit.That never goes away, apparently, but we'll talk a little bit about that.Then, Stuart, we'll talk a little bit more about, um, you know,what's on the horizon and the risk space. I know you and I have another podcastto dig deeper into some of those topics. Will, first to you,what are some of those things? Can you just take us through whatsome of the more interesting things are on your hit list this year?I think at a high level, we've been part of the London Stock Exchange groupfor a period of time now and that invokes the obvious questionsaround how did things start to come together and whatare the interesting capabilities that we have in collectively Post-tradeand how do we assemble that in some meaningful waythat's interesting to the market and our clients?The easiest way to project that, I guess, would be in a more operational contextof what we're doing in workflow. It's really the scenario of any marketwhere each counterparty to some form of tradingis calculating their own requirement and there is no authoritative data.There's no central calculation happening. Right. Effectively,then what they're doing is using Arcadia workflows to matchand commonly agree to transfer typically some form of collateral,but it can be other settlement requirements.If you go then step a little bit further into other areasof the Post-trade solutions business unit here, we have swap agent and trade agent,which are looking to build out services that are moreauthoritative data in nature. They're doing central calculationsthat effectively both parties are not doing their own calc,and it's authoritative data, and they're disseminating that outand articulating a settlement in a way. How does that all come together?I think that realistically, all we're doingis if you took a map of everything post-trade,the different markets that exist, and we're looking at how do we supportwholistically both models and how do we start to bring thingstogether? I think the most obvious way that we'll startthere is in terms of settlement. Everything is a silo to some extent,whether by market or type of settlement, and how do we start to bring those thingstogether into a more wholistic service that's catering for the nuancethat might exist by market? I think that's where we're at.You mentioned authoritative data. Easier for me to say.What are some of the benefits? I know that there are still a lot of legacy systemsthat do a lot of these things in a legacy type way,from a lack of a better word. The benefit to meis taking a lot of the friction out of it and ensuring settlement.Does that mean preventing settlement failures?I think the realistic scenario is a central calculationthat eradicates that you and I might calculate somethingdifferently, so it kills that problem. Now, realistically,in so doing, people have to get comfortablewith giving up certain risks like calculationsat times to a central service. That's not always the easiest thingfor the market to adopt, but I think we see this as a continuumof the bilateral markets will exist and operate as they do today,As we can insert more depth into swap agentand trade agent and similar services where we're producingauthoritative data to build that trust over timeand support wholistically, either model. However, I think that we obviouslyhave a want to help solve certain friction in the market,and authoritative data is one way to do that.That idea is not brand new, and I think it is a differentiator for usin terms of business. People are listening to this,hopefully our clients and prospective clients.I'm not going to hold you to any dates, but is thissomething that we're currently doing, or is this more of a look ahead,2025 and beyond? -We are literally active.While I'm here, I'm typically visiting clientsand starting to get some views in shape. There's an overwhelming amount of interestfor us to figure out how to use the best piecesof the different products that exist within the group hereand start to bring things together. Again, settlement's the most obvious thingthat is most likely to initially happen. It doesn't matterif it's authoritative data, central calculations of something,calculating a requirement that needs to settle where somethingwhere both firms are calculating their own requirement,and using an Arcadia workflow to mutually agreeto some form of transfer. At the end of the day,they still need to settle, so there are different waysthat that happens today. I think step one is almost lookingat the white space. What will say the commitment this yearis that we currently lack the direct abilityto support prime brokerage margin requirements.I think one of the questions is that people or firmsthat are most enthused about this netting oftentimes are the hedge fund community,at least on the buy side, and they will oftentimeshave some form of prime brokerage account. Not always will they,but it will be common. They look at it if I'm trading non-clear derivativesand repo transactions, and I'm also using my brokeras a clearing broker for clear derivatives and futures, and options.I might trade TBA, then maybe a little bit less common.Then, similarly, I would have some form of trading happeningunder my prime brokerage construct. If I have all of those silos,I'm literally settling. If I'm electing to use cashto meet all those requirements. I'm settling cash back and forthwith my broker multiple times a day. How do you collapse that down?That's always been an... Again, nothing novel here.This isn't something that the industry hasn't thought of before,but if you look at our collection of capabilitiesnow within this group, I think people legitimately thinkthat we have the right pieces to be able to bring together and to createa really interesting solution for the market.Yes, I think you're right. I think you just hit the nail on the head.It's not a brand-new idea. We've been talking about thisas an industry for quite a while, and I think there have been plenty of fitsand starts with getting something like that done in terms of netting.However, as you said, now we have the pieces.We have an opportunity to actually put that into practice.Not to architect it too much here, but the world lives in silos.Partially, there are some legal hurdles that exist there, but these are solvable.There are cases where operational netting agreementsthat will commonly be referred to, that these things existand between two different bilateral partiesthat people are settling net some form of different obligationsthat would naturally be more legal silos to some extent.There are instances of this. The issue ends up being,it's good for risk management. It's good in termsof reducing settlement risks associated with if I pay youand you don't pay me, obviously, why would I pay you?Netting that exposure down is an interesting thing to do.There's typically some form of funding like benefits,and not all firms operate the same way, but they usuallyhave some form of funding benefit. There can be other liquidity benefitsfor banks in particular. By the way, if we do it smartly,it should be a more automated solution so that you're doing fewer settlementsin a more automated way. Therefore, operationally,you should get some form of benefit and control out of it as well.How do you chip away at that problem? What's the scope of it?How do we start bringing some firms together?How do we close the white space down? How do we march forwardand actually start to do something interesting?I would say there's always more to come. I use that term a lot.I think there's more to come here for sure.Completely. -I think we'll probably come back to thisat some other point, especially when it's hopefully maybe...I'm not going to put a date on it, but at some point,we have clients actually doing this in practice.This early quarter, we sit here in January right now,and in Q1, it's interesting what the common scopethat people think the starting point should be.You could start from a few different angles on this.It is an interesting topic, but annoyingly,it's probably the thing I've talked more about in my life in my 23 yearsor whatever in financial services. -You're doing it right now.We're still talking about it, so I feel compelled to solve for it,obviously in a practical sense, so let's do it.Let's do it. All right. Great.Thank you, Will. I know there are a couple of other things I had mentioned.Stuart, I know you're sitting there. -Yes.We can go back and forth, but I do want to finishon a couple of other things that I had mentioned in my intro.Will, we've been talking about things like tokenization.We've been talking about repo. In the US, there's a treasuryand repo clearing mandate, and then there's some UMR.Mexico has gone live. India is going live in the middleof this year, so far April first, and then, Stuart, China has just come outwith their rules recently and their roadmap,which extends into 2028. I'll come back to youon that maybe a little bit, but Will, just to finish offsome of those topics, tokenization, repo, and treasury clearing.These are the things that are really on our roadmapfor this year. Anything to add?Anything that you're seeing in terms of focus for the business,our client base, and if there's anything elsethat I'm missing? -I think that the UMR pointand repo clearing, and that seems like firmswill want to start to try to figure out how to go live.It isn't a requirement until the end of June of next year,but in the back half of this year, start to do something there.I think we need to offer a workflow to help people operationally,support the interaction of margin requirementsbetween their now clearing broker or sponsored brokerin some context for repo transactions. I view that as white space, effectively,that we need to fill in that white space as we kind of move into that time period.Like what we did earlier this year for the TBA margin.Completely, and I think comparably, UMR fits a similar fact pattern in thatmaybe it's a little bit specific to certain markets,but where we can be involved in those marketsto help the operational processes around marginingthat activity, obviously we want to be doingthat no different than where we've done it in other places.Tokenization, real quick, is just a more complicated topic.I think the way tokenization at the simplest level is,if you're transferring securities, one view is it's very cumbersome,and we can see this in T plus one and all the interestthat's gained in the fixed income and equity markets, but--What about substitutions as well? -Substitutions are more movinga security from me to you. I own some form of government bond,whatever this type of security is. It's not a trivial process.I think people who aren't familiar with it probably think settlementslike this instantaneous thing, but there's this long daisy chainand lineage of systems typically and operational teams,and sometimes accounts and involved in getting a securitytransferred from me to you. Now, one way to change thator sidestep that complexity is to onboard that asset in some capacityinto a digital ledger. Then once you do that,you entrust the ledger to be the determining factorof who owns what at that point in time. Then the transfer from me to youis simply, you know, a bit of data that's coming through a ledgerthat's saying I changed it from my ownership to your ownership,and you're done. -We're talking about a closed ledger,or any ledger or-- -The reality is--What is fit for this type of-- -There's been a lot of work on this.There are certainly companies that are out therethat are trying to solve this problem. Our view is moreif the industry pivots in that direction, I think we can helpthem adopt that type of technology by connecting them to these ledgers.I don't know if we'd ever go so far as to want to create our own.We talk about it sometimes, but realistically,that's a slightly different angle on things.If we're brutally honest, as much as we've talked about itfor a number of years here, there are very thin actual production use cases.You see it most in intraday liquidity repo,and that's where it's kind of actually gotten some real material usein a production standpoint. In the world of collateral management,it was always thought that that would be the next obvious placeto start or figure out how to work. -DTCC, but that's whatthey had worked on, right? -Yes, there are a few.J.P. Morgan did some stuff with their Onyx platform,and there are other firms that exist out there too.Again, not to plug people, but there are peopledoing interesting work on it. I think the demand is debatable.I think it's like we don't stand to win one way or the other.All we want to do is try to help firms automate.This is one approach where the industry could pivot towards tokenization.If they do, again, we would be, you know, fully helping as muchas we can to kind of connect them into these new capabilities.Got you. Thank you for that, Will. I appreciate that.Of course. -Stuart, you've been sittingthere very patiently. I didn't forget about you.We still have a lot to talk about in your world,although our worlds are all together, but we all have our focus.Just getting back to UMR, I had been doing some work in the Mexican market,just so people know, the recording of this is January,we are looking to do some events in Mexico.However, we have India yet to come, and we have now China yet to come.India is interesting. Your thoughts there because it's quite different.They're using the central clearing mechanismfor SIMM calculation. Then, we'll get into...I'm using this as a segue into talking more about SIMMand optimization in general, but I figure we can talk about this first.What are some of your observations? I know there was a delay.For India, it was November. For the first go-live date,it's now moved to April, but they do have a different type of setup.Yes, if you take Mexico quite a traditional UMR rollout;it feels like a similar set of rules, a similar set of requirements,and rolling out, following largely the normsthat have been done elsewhere with some tweaksthat make sense for their geography. However, when you look at India,they've taken a very different approach, where the CCIL is going to takea much more active role in helping people calculateand manage initial margin much more as a centralized infrastructureprovider. Which is really interesting because that's not the wayanybody else went. I think it was always an option.Iran is the same, but in the end, most of the time,people have gone down the bilateral route using third party infrastructure providerswhere necessary. It'll be interesting to see how that goesand what the benefits and limitations of having done that are.I think that date is still pretty aggressiveon the first of April. -I agree.However, we're sure that that is the date they're going liveand the reason the FAQs and other pieces have come outto fill in some of the gaps. I think there are still some questionspeople have around some of those mechanics and how that will work in the early days,but it'll be fascinating to watch it go liveand see how that approach works. -Yes, I agree.We've been following this for quite a while since VM went live,and now we have the IM rules, and the constructin the Central Clearing Corp of India is very interesting.We've been talking about doing some more educationwith our clients on how that is going to work,how they can potentially leverage our solutions to help as well.However, I think it is a unique situation, let's put it that way.We'll see. I think as a business,because we are the standard in IM reconciliation andcalculation. We should probably play a part, but we'll see.We were talking this week, actually, that we feel it's probablyabout time to start doing maybe some SIMM refresher courses.Probably five or six years ago, a lot of people went on those courses,learn all about SIMM, and became quite expert in it.There's that natural turnover of people, and there's a lot of new peoplenow in that world who maybe didn't get some of that backgroundbut would benefit from it, but then also those new geographies.I think it's worth understanding some of the mechanics of the calculation.If you're coming from a VM background. It's a big jump allof a sudden to ISDA SIMM. It's maybe not so much of a jumpif you're coming from a risk background, but I thinksome of those education courses might be quite helpfulto let people really understand how that calculation worksand some of the challenging parts of it as well.Some of the changes that we're going to go into really quickly,but I'll finish off on that. Well, if you have any comments on that,happy for you to jump in as well. Again, as of the taping of thisand it's always dated, and you can look at a datewhen you watch it. Hopefully, people are watching it.January. From Mexico in general, in terms of education,we are working with our clients directly in the regionto do more education and give them updates.We do have a working group that is still running evenpast the December 31st go-live date. To continue that I thinkwe're going to be partnering with ISDA as we've donemany, many times in the past, and they are also goingto probably join us. We don't have exact dates\,but maybe we'll put a link to the working group on this podcastto see if people want to join. Our working group, obviously,is only open to clients, but if any clients are watching this,we'll give them an easy way. -You have to be a special breed of personto not be a client and want to be part of one of those working groups.We always get some good numbers on our podcast.We're doing really well here. Thanks to the crewfor making us look good. Anyway, that's part of some of our effortsthat we've always done and we'll continue to do,but you're right. Look, we're not going to bearound forever. We've got to pass on what we have learned.I think I heard that from a Star Wars movie.I think Yoda said that. Anyway. It's a joke.Staying with SIMM now, in 2025, there are some big changes happening.Break it down for us. -People may have heard us talk in the pastabout recalibration cycles. The last one was oneof the biggest recalibration cycles we've seen.We think around a $100 billion change in exposure across the whole industry.That's a whole year's worth of changes in the marketplacebeing wrapped up into recalibration and published.I think for a while, there's been this feelingthat that doesn't really meet the original requirementsthat people had for this model, and there's alsojust too long a time frame to have for recalibration.There's been this push to bring down the calibration windowsto make them more regular. We're going to seethat come into fruition this year. We're going to see two-week calibrationsthis year, and these two have let us know the timings for those now.We're going to see one go live at the start of July,and the second one go live at the start of December.They're going to split out recalibrations, so simply changing the underlying datathat goes into the model and recalibrating itusing the same methodology from the model itself or any calibration methodologiesthat they use to generate that model. Now we're going to seethis two-tier system where we have calibration changescoming half-yearly. For the moment, I think some peopleare suggesting may go even more frequently going forward.Then these longer time frame model changes each year,coming out aligned with the December release.What do you think? Again, just always putting myselfin the client's shoes. What are some of the challenges for them?What should they look out for? How are we helping?Yes, actually quite tough. Those timeframes are quite different.It sounds, "Oh, it's not going to be that different.We are halving the time, maybe that's going to be okay."Realistically, your time to consume a new versiongoes from three months to four weeks. When you put it like that,it sounds like a really big change, and it is a really big change.A lot of those people are using vendor systemswith install packages, four weeks is an incredibly short amountof time for someone to create a release, ship it to their clients,install through UAT, and get into production.Have some contingency. -Well, yes, forget contingency.It's too quick. -It's far too quick,and some of those vendors may have 100 clientsthat they have to release patches to in that same four-week period.That is tough. We're lucky we're a SAS vendor,and it's one platform. We can put a time frame in.We're still writing it down and making surewe've got the thing to go by the numbers, but much tougherif you're doing an install system and you've got your own modelrunning internally, not to mention all the modelsthat hang off it as well, your XVA models and other things.I think a lot of people look back at those processes,seeing if they can tighten them up, seeing wherethey can cut out some dead time. We've released extra toolsover the last couple of years in anticipation of these changescoming to make it easier for people to do impact analysis,to make it easier to do back testing, and other things.Please do make sure that you're ready. Make sure thatyou can do that time window, and if you're struggling,by all means, talk to us, and we can tell youabout some of our experience and how we've approached things.Yes. That's a great message. As always, we're plugged into the forums and are fully updated as any good vendor should,but that's good to know for anyone listening who might have some questionsaround what it means for them. If I'm an end-user client,what are these changes for the SIMM recalibration going to mean?This really comes from the recommendations,from the regulators. This is the reason why they're doing it.To be more nimble. I would probably put money on it that,as you said, it might change again, and we'll see how that goes.There are no plans to do that. I don't want to start to scare anyoneor speculate too much, but as we know, things don't stay the same for too long.The changes seem to have been put together very well.[crosstalk 00:24:29-00:24:30] -I think the communicationwith the industry has been very good, so kudos to it.A lot of flexibility in the future about how they want to structure,if they want to make another change. They've done the structural changesthey need to do now to enable themselves to be more nimble in the future,so now they've got more flexibility. -Absolutely. All right.Well, more calm. That's July is the next one coming.That'll be here before you know it. I can't wait.It'll be summertime and golf. -I can't wait.Stuart, anything else? What are some of the other big thingsyou're looking at for 2025? We've talked about thingslike IM optimization. We're talking a lot about SIMMfor House IA, and we've talked about that recently while we were here.Then, maybe we'll just throw a monkey wrench and talk about AI.We have to. It's an obligation to talk about AI.We'll just talk about it a little bit before we close,but what else is on the agenda? Obviously, this is not exhaustive.I don't want everybody to think that we're...This is only what you're doing. We're getting the highlights down.I think optimization is important. It's incredible when you see how muchof an impact those firms have on IM levels globally.They're pretty important in the amount of IMthat gets posted and making those bank business models work.IM optimization in general is really driven by data quality.The higher the data quality, the better the optimizationthat you can get. We're doing some work with a lot of banksto try and improve the quality of SAC data that we can see.Really interested to see how we can use data standards similarto what we've used elsewhere to see if we can bringmore consistent data that we can use more widelyand really help banks get that data quality high,as high as it is, hopefully, for ISDA SIMM.Then they'll get more benefits from the optimizationthey get off the back of that. -Awesome.We can start to see how 2025 is going to look.Hopefully, there aren't any giant surprises, but you never know.Market disruptions could change everything.We shall see. In the US, obviously, we have a new administrationcoming from a rag and industry perspective,which is near and dear to my heart. There are going to be some changesat some of the agencies. We've talked about thingslike Basel III endgame. Who knows where that's going?Then anything really associated with the SEC, Gary Gensler stepping down.Whoever's going to have that, there might be changes.That affects everything we do could even affectsome of the things we're talking about here,but we shall see. This is probably a good overall viewof what we're going to be facing this year,and we even know about some things in 2026.We'll probably come back next year at this time, hopefully,if we're all still here and doing well to talk about 2026.I know I mentioned AI, what are we looking at right nowfrom our business perspective? Obviously, we have a chatbotthat works off of ChatGPT and that has been our first little toeinto the waters, but disruptive, helpful, and destructive.Well, I mean, you don't want to ask me what I thinkbecause I think we're all going to be sitting on our couchesat some point with no work to do. However, maybe we'll be long gonebefore that happens, but you always need ideas.I think so. I think for the interim, in my opinion is going to be very helpful.Y I only heard about ChatGPT from my daughter,who's a college student, not less than two years ago.All of a sudden, it's being used for so many things.It's almost replacing Google for Google searches,and it does help me write, actually. -Definitely, this yearwe've been using the chatbot internally to help us navigate regulations,navigate our own documentation, and it proved invaluable for those things.I think this is the year that we'll bring that to clients,make it more available more widely. We're really looking forward to thatand getting some more feedback. We did some demos recentlyat some client events, and people loved it.People just crowded around the machine and kept throwing stuff at it.It was great to watch. -Multiple languages.Multiple languages, yes. We forget that sometimes,how global we are as a business, and how important it is to hit someof those other languages, it's not always so easy.We're such a niche. I mean, it's the trainingthat we can offer in some form of capabilitylike that is the uniqueness. That is the difference.That's right. I think some of the crossoversbetween my world and Will's world where we have risk calculationsthat we then try and reconcile so we can drive a workflowoff the back of it. I think this is the areawhere it could be really interesting to automate as much as possiblesome of those processes to make it easier to get data from one firm to another.I think that's an area that we're going to tryand jump into this year. Very hard to know what's possible,but I think it's really exciting. Somewhere where we could havea real impact in terms of the efficiency for firms,the amount of manual intervention that's needed,and how much more we can automate. -Probably one more use casethat's a natural language model type of use case,which is we have a relay service that is trying to take an email,and today it requires the email to be fairly structuredin the composition of the email so it can be read into our workflow product.That's effectively just helping to automate.If my counterparty is not on the network, I can still get the automation benefitsby plugging email into the tool and we call that service relay.However, the requirement is that it's fairly structured in the wayyou need to represent data in that email, so it can be understood as a margin call.However, the language model bolted into thatobviously should be able to help. If I can understand a personreading an email and the natural heartbeat of a day,and what that client is telling me, equally the language modelcan do the same. I think the difficulty there is one,how do we do that? How do we do that in a controlled way,and all of those things? Then equally, the risk is how do you knowthere's something that might not be fully accurate,some level of concern that my interpretation might not be right,and serve that up to clients so that they can take action,which oftentimes would obviously be sendingan email back to my client around, "I don't quite understandwhat you're saying here." However, it's an interesting use case.I think we're currently figuring out what would be the right plumbingfor us to go after that. -Yes, we're scratching the surface.Large language models, generative AI, it's all coming.It's obviously coming to the world of finance as well and to our business.Not to mention how we build tech. I think over time that'll hopefully helpus deliver things quicker and quicker and quicker into the market.Absolutely. All right, guys, I think we'll leave it there.2025 is going to be a busy year. Sounds like a very good year.I look forward to seeing how all this stuff comes to fruition.Thank you both for joining us today, and we'll see what 2025 brings.Thank you guys. -Thanks, John.Thank you, John. -Thank you allfor joining today's podcast. I hope you foundit interesting, I know I did. You can find this podcastand all our other podcasts on Spotify, YouTube,and any of your other favorite streaming services,and at acadia.inc, our website. Thank you for joining us today.Thank you to my guests, and we'll see you again soon. Bye now.