- Perspective on what AI is — and is not — and how it is expected to impact the muni market.
- How does technology aid in pricing and disclosure?
- How can technology help the industry do more with less?
Transcription:
Gregg Bienstock (00:09):
All right, thanks Lynne. Thank you for the Bond Buyer for having us in a great conference. So what solve, what do we do? Probably our core product is, well first off we're a data and software company. The core product is a product we call quotes. Think of pre-trade price transparency. For those of you that looked at the MSRBs, I think third or fourth concept release on the subject, right? They started looking at this SPAC in I think 2013 or so. That's about the time solve started actually doing what we call quotes, which is parsing structured and non-structured messages and being able to deliver that back to our clients real time in what I'll call a usable format. You can get a little more detail on that later. Kind of a second product that we've launched in that space is what we call our predictive price px.
(01:05):
And that is a real time size specific delivered however you want predictive price. So we're looking at where the next trade is going to occur, buy side by bid, mid and offer excuse. And then there's the kind of core muni business, which is essentially the diver application that we have. So serving the public finance market, debt analysis, debt service calculations, 15 C to 12 analysis, new issue pricing support, things of that nature. And then last but not least, retail time and trade disclosure. And in that regard, we actually have next week or the, I'm sorry, the following week we'll have a release coming out to help our clients in terms of the new amendments that are effective the first week of March. So it was a long trip down the elevator there or long, so.
Lynne Funk (01:57):
Perfect. Thank you Dan.
Gregg Bienstock (01:59):
Thanks
Dan Silva (01:59):
Greg. And there we go. Thanks Greg. And thanks everyone. My name isDan Silva. I'm the CEO of Adaje. We develop and offer a cloud-based AI powered debt modeling and management solution. Basically helps streamline workflows for advisors, issuers, bankers, everyone that needs to be able to model bonds and know what that might look like into the future. And also post issuance to know what it looks like from a debt modeling and management perspective for advisors and bankers. We help them more efficiently screen for opportunities, respond to RFPs and produce impactful analysis for their clients. For issuers, we help them really lower the bar to be able to do some really good capital planning on their own and also evaluate things post issuance for continuing disclosure and reporting purposes. And really our goal is to simplify workflows for folks in the public finance space to help 'em get more done with the fewer resources that they have these days. So that's us at Adaje and thanks for being here.
Lynne Funk (03:08):
Thanks.
Matthew Smith (03:11):
I feel like I'm staring into the sun over here, so if I'm squinty, that's why. So my name is Matthew. I run Spline Data. I founded it a couple of years ago. My background has been secondary algo trading. That's been pretty much the bulk of my career and I was doing that for a while and wanted to start spline to kind of make automation a little bit more accessible to the muni market. We compete with Greg on the pricing side, so we price muni bonds in the secondary market. We produce real-time curves, similar proposition as well. It's been minute, an offer for different sizes in the market, odd lots being an important distinction there. And then recently we got into the primary side. So taking some of our expertise from pricing bonds in secondary and applying that to new issue, helping to give a second opinion on where something might clear the market, which tends to be helpful in some of the foggier areas of the market, like Texas muds, some smaller competitive deals, things like that. So really we're trying to take all the data that's out there that already exist, boil it down into something that people can turn into money.
William Kim (04:34):
Thank you, Lynne. Thank you to the Bond Buyer. So I'm Will Kim, Founder and CEO of MuniPro. Well, we serve all the core constituents in the muni market. So we think of ourselves as a technology partner to issuers, advisors, banks, as well as asset managers. So for our issuer clients, what we're really doing is debt management, helping them with budget or financial forecasting, and then finally looking at pre and post pricing analysis for them. For our MA and banking clients, we're really doing that at scale. So we're covering every issuer in the market, providing them with all of the capital structures that are fully verified by credit. And on top of that, we're feeding that through to look at refinancing opportunities, look at secondary trading activity as well as advanced analytics like option pricing. And that's where we add most value to our asset managers where we're calculating option pricing for all secondary trades and normalizing option values over time in the secondary so that you can compare apples to apples. So very excited about our growth and happy to be here.
Lynne Funk (05:38):
Excellent. I think as much as the tax exemption has been touched upon throughout today, I think a lot of people have brought up technology and I think that this is something that clearly we have a whole panel on. I'd like to delve into first though, can you maybe each talk about what exactly your work aids in the primary market and in the secondary markets and you can bounce off of each other. Who would like to kick it off?
Matthew Smith (06:09):
Start?
William Kim (06:11):
Okay, so will on the primary market, I think from the issuer's perspective, we're thinking before the bond deal, right? We're thinking about, okay, what do you really need to do to get ready for a bond deal? Think about the capital needs and structuring before that deal comes into market. When you're talking to a banking client, really they're trying to pitch more to cover more clients. We're helping them get their ideas out instead of spending time doing kind of the same grunt work, we're automating that process for them. So we're helping them get that operational leverage. And I think we're making the market more efficient in that respect, getting more pitches so issuers are seeing more potential ideas. And so that's where we help the market on the primary side. And then on the secondary, we're mostly doing analytics in terms of option pricing.
Lynne Funk (06:55):
Okay,
Matthew Smith (06:55):
Perfect segue. Perfect segue for me. So we try to take over when the bonds start coming to market, we're involved I guess leading up to deals and trying to give people an idea of where things will come to market. But we want to help people basically get the best pricing that they can, whether that be from the underwriting side or the people buying the bonds at the end of the day. So the way people are using our primary product at the very least is saying, okay, I've got this deal coming to market, or I've got these deals coming to market over the next few weeks. How can I manage the expectations with clients? And maybe that's giving them updates on how the pricing is developing throughout some of the more volatile market conditions that we've seen in the last few months. And then when that deal comes to market as it transitions into secondary, how is the price changing over time? And really we're trying to create an environment where it's easy to provide liquidity in secondary.
(08:02):
The beauty of Munis is there's so many of them, most of them are illiquid and I find it an interesting problem, but it is something that a lot of people struggle with. And you find that folks are focused on different sectors of the market or different states. And so ideally what our data is doing is making it easy for somebody who might be New York focused to start looking at adjacent states or something or bonds in other parts of the market and kind of trust that they're putting forward a price that's going to generate some p and l for them.
Lynne Funk (08:33):
Great.
Dan Silva (08:35):
So I'll start with issuers. We basically help issuers understand what refunding opportunities are and how to model up potential capital plans in a very quick way. So they can very easily say, Hey, here's my outstanding debt portfolio Adaje, tell me what can produce X level of savings across the entire portfolio and structure in a way that actually represents a real issuance. Whether that means there's other considerations like reserve funds or capitalized interest funds, et cetera, handle also investment of bond proceeds and model all of that out. And that also includes thinking about things like interest rate sensitivity. We have maybe a 100 million capital plan, but what does that look like depending on whether rates go this way or that way, whether my project increases or decreases in terms of dollar amount, whether any number of parameters change in a way that instead of running your own scenarios time, and again, you can just tell Adaje, Hey, these are the parameters I want to change.
(09:36):
Go and we produce for you dozens of scenarios in a few seconds. So that's for issuers very similarly, we do that for advisors and bankers that are producing those numbers for their clients, but it just goes a step deeper where they really want to fine tune the numbers to be ready to show those numbers to investors post pricing. So it just goes a level deeper. On the issuer side too, we help with internal loan management. A lot of issuers have maybe pool loan programs or internal banks especially. We see this on the higher ed side. Also, certain authorities that have recycled revolving loans and they really don't have a tool to, they model their external debt, but they don't have great tools for modeling all the loans that they're actually making to their constituents, to their different units and departments. We have a whole ALM system that helps them manage that alongside both their assets and liabilities and make sure that the programs remain solvent through the whole life of the program. So those are a number of the ways we help folks.
Lynne Funk (10:38):
Okay, excellent.
Gregg Bienstock (10:40):
I'll maybe approach this slightly differently. I'll kind of on the primary market side, think about it from kind of the cycle if you will. So starting with we're tracking all the outstanding debt for each obligor, we have our own obligor database that we've built that kind of serves as the foundational aspect of what we're doing. Once our clients decide that they're coming to market, they're using us, obviously I mentioned earlier with 15 C to 12 analysis. We've been doing that since around the time of MCDC. And then you get into new issue pricing and whether that's indicative pricing for a new issue or as you move towards final pricing. But the idea is to use our pricing platform where similar to what we're talking about here is right, you're entering a few data points and what we're able to do is construct scales for you within a matter of seconds, exposing vis-a-vis transparency, all the underlying comps that we're using, giving users control of that.
(11:35):
So the idea is to bring you to that point, I call it indicative pricing because the fact of the matter is that I think at the end of the day, each firm, each of you in this room, the underwriters at your firms, there's a certain appetite and there's a certain gut feeling, and that's going to help drive what you all do as well. There's a tool that also can look just historical deals, excuse me as well. And then something that's near and dear now seemingly to SCC examiners is the post issuance analysis. So being able to look at where your deal price and being able to, in a matter of seconds, identify where those bonds have traded in the secondary market, that's all part and parcel. And then when you get into the secondary space, I think similar to what Matt said, the objective is how do we help you make more money?
(12:23):
So the quotes platform, what we're doing there is we're taking this mass of messages that are out there that as we speak to folks, what we know is that they don't get a chance to look at all those messages. And if we can all of a sudden take those messages and turn those into information and make you that much smarter, be it on the buy side or the sell side. So on the sell side, it's an information asymmetry that you're solving for on the buy side, you're solving for the issue of I got too many damn messages coming in and I can't look at 'em, and so what am I missing? And then similarly, again, as Matt said, I guess we're competitors. It's like the Toyotathon, right, that we just keep running at each other. So I had to put that in there.
(13:07):
But in terms of what the predictive price is, we're trying to use all the data that's out there. We're trying to use security master information, ratings, information, trade data and quotes, data to construct. And we have to construct an and built an AI driven model that's going to give you a sense of where the market is going to occur for virtually every bond that's out in the marketplace. And as Matt noted, right, there's plenty of illiquid bonds out there, so how do you go about doing that? And then lastly, there are tools within the platform that permeate the idea of being able to identify comps based on structural and credit characteristics. So again, a similar point to what Matthew mentioned.
Lynne Funk (13:44):
Okay, so you touched on transparency, Greg, and when I say transparency, everybody says that this market needs more transparency. I guess from all of your seats, how are your firms making transparency better, whether that's through issuer disclosures, the ease of finding it, whether it's pricing, whether it's just data in general, big and broad, but have at it?
Gregg Bienstock (14:12):
I just think that the biggest issue that we face that we have faced over the years, so first there was just an absence of the data out there and then Mark Kim mentioned this morning how far things have come and there's more data available, it's better structured, but it's still not great. So if you're trying to pull stuff from Emma and you're trying to extract data, you can do it now better than you could in the past, but I just think it's a matter of what we all do up here. There's more data out there, but how do you take that data and make it into information for all of you in the room? I think the data to support transparency is there. It's just how do you take that from a raw form, if you will, and put it into something that's usable. And I think that's the objective of each of our companies in one form or another.
Dan Silva (15:00):
Yeah, I mean, just to kind of go off of what Greg's saying, we're not necessarily a data provider, but I think we work with folks and companies like Greg's and to basically synthesize that data, make it more useful. So we live in an industry that for decades, data has been super fragmented. There's a lot of manual process. And quite frankly, the tools have also been pretty antiquated for a very long time, aside from our products of course. But for us, it's bringing in all these data sources to be able to actually make it actionable, put it into an actual analysis. It's great to know pricing, but okay, what does that then now mean for my bond deal? What does that mean for my refunding analysis? What does that mean for my debt service? Can I still make this project work with the net revenues that I'm projecting out?
(15:58):
And so for us, it's just synthesizing all this data into a way where we can quickly put them into actionable models and have very nice reporting so that you can say, okay, given the latest data, given all these sensitivities I need to account for, and all these complexities, real world complexities that exist in the market, having to meet certain coverage ratios, having to account for all the different costs of issuance that exist, come up with real plans of finance that are useful, that you can then report to your constituents, to internally to your management, and for those advisor and bankers out there, produce interesting analysis for your customers. And I think automating that whole process and integrating everything is just saving folks a ton of time and a ton of cost.
Lynne Funk (16:38):
A lot of data, and there's a lot of data out there. How do you make it good and how do you present it?
Dan Silva (16:42):
Exactly.
Matthew Smith (16:44):
This is why I got into the business actually not too long ago. I was going to say a while back, but a big thing about big improvement in BALs curves was that they included some of the comps and the observed things that contribute to their curve. And people would always kind of look to those MMD reads through email to see what prints they're looking at. I think you can kind of split it into two different areas of transparency. There's transparency of inputs to your model, and then there's transparency of the outputs. And I think that as models get more complex and it becomes, there's more data being used into pricing and individual bond or an individual deal, what people are going to start to care more about is the transparency of the outputs. You're trying to sell me a product, how do I know it does what you say it does?
(17:50):
And so one of the things that we wanted to do from the very beginning was be very forthcoming about how well it does what we say it does. So if we claim to predict the next trade very easily in our system, you can kind of go and see how we perform in different parts of the market on individual QIPs, everything about that. If you want to see receipts, if you want to see the data we use, we can't show it all because it's a lot. We use thousands and thousands of bonds to come up with a price for one single one. But really I think what matters is how we perform where you care. And so we've been trying to shift the conversation at least in that direction. On the data side.
William Kim (18:30):
I think it's an interesting question. So for our customers, I think we're providing two parts to transparency. We're giving them more transparent data in the sense of data quality. There's a lot of garbage data out there. So we do an incredible job verifying all the debt and all the trades and making sure they're all right. And so our customers are extremely happy. They get that kind of more accurate view into the market. But on the other side, in terms of the public disclosure and the public mission of transparency, the work we do at issuers helps them commit to providing continuing disclosure. So for example, we'll automatically produce continuous cashflow reports for some of our larger state clients. And since it's so simple for them to produce, they're like, okay, I'm happy to include that in my continuing disclosure knowing that I got to do this every year on an ongoing basis. So providing those reports I think not just helps our customers, but also the broader public.
Gregg Bienstock (19:30):
I just want to just jump in for a sec on this just to continue on the transparency point. Matt raised a really good point. It's being able to show your work, if you will, when asked and being able to show the value of what you do. And so when I think about that contextually, we've been asked to do that for a very long time. So if we're doing a 15 C to 12 analysis, it's not just show me what we call stoplight report, but it's, let me see what's under the hood when we do our new issue pricing, when we build a scale for you, it's not like here's a scale and trust us. It's the idea of being able to look at not just the comparable issues, but the comparable bonds and the comparable trades that are supporting it. And it's the same thing now when we're doing what you're doing right? It's the predictive price. It's being able to back test and deliver the results back for the market segment or the bonds that your user or users care about to be able to show we're doing a really good job or we're not. And at the end of the day, that transparency is critical to everything that we all do from just a delivery perspective for all of you.
Matthew Smith (20:36):
And for the record, Munis are already pretty transparent. It's kind of a privilege to be able to get a feed from the MSRB with all the prints. There are plenty of markets out there that it's not obvious where trades are happening. And so I think we're in a very fortunate position from a source data standpoint. And now as vendors, we have to deliver in ways that give people a chance to trust the data. That gives them a reason to buy into automation and see the benefits.
Lynne Funk (21:12):
I want to move into, I guess it would be efficiency, transparency, and then AI, the recent headlines from Deep Seek, which I assume everybody in the room has heard. I guess I'm really curious about what does that mean? Are there implications for this market? What does it mean?
Matthew Smith (21:41):
You've been waiting for this one? I have mean LLMs have been kind of disappointing from a quantitative standpoint. Everything that they are pointed at is generally speaking qualitative, which has some really good use cases for whether it be parsing oss or I'm sure Dan and Will have plenty to talk about here. I think the more important takeaway from that headline is that advances in technology are a lot cheaper than they have been, and it's a lot more accessible to the broader market. And so I think we're in this kind of golden period where we have all of these tools available and now with more and more competitors in the space for a good reason, there's more data that you can apply that technology to and do it in a way that's additive to your business. So I think things like deepsea coming out or things like ChatGPT in general is great because it makes it really easy to code. If you're code curious, you can throw something in there and get a starting point. I would never commit any code from ChatGP to our own code base, but it makes technology a lot more accessible. And I think that that is the biggest value add that the muni market's going to see from LLMs. And
Dan Silva (23:04):
For us, I mean it's going to have a pretty significant impact on a new offering. We're about to I guess announce today, but yeah, now let's go. So let's do it. But for us, it's great when there's new LLM models out there that are cheaper to run, it lets us do way more than we would've otherwise. ChatGPT is not cheap to use if at scale it's relatively cheap, but deep seek is way, way, way cheaper. And so for us, we've always believed people in public finances are better tools that move a lot faster. And I don't think this is particularly revolutionary in concept because some of our customers have brought it up to us, but we're announcing our own basically Muni GPT. It's a industry's first AI powered basically financial analyst where you be able to, instead of running your own numbers for hours, you can basically say, Hey, Adaje, throw together, run a model for a tax exempt bond deal that's 30 year final maturity, $50 million new money and by the way, refund all the maturities that exist for this issuer that produce at least 3% savings on an option adjusted basis.
(24:16):
And you'll get your first iteration of that analysis completely modeled up, and then you can take iterations from there and say, Hey, maybe I wanted to change the cost of issuance, or you made an assumption on the reserve fund change that. And so we're building out this whole new way of basically interacting with your modeling software that folks just haven't seen in the past. And we're pretty excited about it. I think it's going to be a game changer. I think for a long time the complexity of dealing with some of the older softwares that are out there, which we will go unnamed, have been a hurdle for folks, new analysts entering the industry for issuers, trying to model their own deals. And I think this will lower that hurdle pretty significantly.
(24:57):
And I think advances like deeps seek and probably further advances that are going to happen here as soon as somebody like Deepsea proves that this is possible, I think you end up seeing other folks come out of the woodwork and show, okay, well we can do it in America too. So I think you're going to see more LMS, cheaper, easier to run, make advances of being able to integrate that into industries like ours. Also cheaper and easier to have in terms of products. So we're thrilled to announce we're going to have our own little Muni GPT, and it'll be available later this year and can't wait to share it with our customers. And if you'd like to be on the wait list, come speak with us. We got a booth out there.
Lynne Funk (25:33):
Yeah, this is against the rules. You were supposed to give me a heads up of any kind of breakthrough. I did not know about this. I was not privy.
Gregg Bienstock (25:40):
I have nothing nearly as exciting as that. So I'll just add to what was said here. So I think every time there's some new technology out there, the initial cost is very high. And then once you have competition, and I was talking to someone earlier today about this competition is really important. It forces us to do better. It forces us to do better for whatever it is we're doing, but also to find a way to do it in a more economical way so that we can deliver it to you all more economically. And it is a cycle that continuously happens. And so while with regard to this specific technology from China, I have some concerns, especially as to the integrity of some things that are in there, maybe search for or ask about Tiananmen Square, but go more broadly and just conceptually think about the idea that anything in our lives that we've experienced as new technology over time. When it first came out, I remember my first friend who had a big screen tv, the flat screen, which was that thick at the time, it was like $10,000. You can go in for 600 bucks now and it's paper thin and the technology just keeps evolving, keeps getting better. And I think that's at the end of the day, what I would take away from this is that it's forcing the competition. It's forcing us in this country and around the world to just do better, smarter, faster.
Lynne Funk (27:01):
Okay, Will I know you didn't get to answer this question then.
William Kim (27:04):
To add to what Greg was saying, I think competition is incredible. So we looked at deepsea the first day it came out. We're like, okay, let's sign up, let's see if it's real. And then by the time we got around to actually kicking the tires on it, open an eye announces new model with significantly lower cost. So we don't have to move anywhere, which is great for us. I don't have to move shops. But not only is it cheaper, it's not like it just saves us money. I think it increases our capabilities because with the same amount of dollars we can cover more. So what does that mean for the muni community? It means, hey, instead of just doing select pages in an OS that we're not sure on and feeding it to an LLM, maybe we do the entire financial statement. Maybe we do every single document ever uploaded because costs keep going down. And so that changes kind of what we're able to offer. And at the end of the day, expand transparency, expand availability, and make things cheaper for everybody.
Lynne Funk (28:09):
Has it moved? I mean I think this is, maybe the answer's obvious. Has it moved a lot faster than, and when I say it, I mean technology from your seats a lot faster than you anticipated?
William Kim (28:20):
Oh, absolutely, yes. I think even the most optimistic people were like, oh, maybe at the end of this year something like this would occur, but it's much faster than anything we've ever imagined. I think it's because of that. You see this market impact. I used to be a power banker, like utility banker, and everyone's like, oh, we got to build all these new power generation stations, which I think will still happen, but it definitely made waves. In terms of projected power demand, just given how advanced it's,
Lynne Funk (28:48):
Is there any risks to that, to how fast it's moving to your businesses, to the industry? Or is it all gravy?
William Kim (28:56):
I mean, you're talking to the tech guys, it's good for us.
Gregg Bienstock (29:01):
I don't think mean, first of all, I think things, they move faster and faster and faster. There's constant innovation going on. And again, look at what we're doing up here, but as a tech provider, and I'll say, I won't speak for all of you, but I think I might in this regard, we don't put something out there and the providers of technology that we use aren't putting stuff out there that isn't tested. And it's your lifeblood. If you screw it up, it's bad for you. And so if those providers that we're relying on who are accelerating things for us all, if they're thrown out in beta, they're saying it's beta because they don't want to screw up their reputation. And I think it's the same thing for all of us. So innovation is better, it creates the efficiency, it creates cost benefits for all of us. So I just look at that role, that kind of snowball if you will, as something that's really beneficial to the market and to each of us, our businesses, which ultimately benefit you all.
Matthew Smith (29:56):
And you can't really stop it, right? It's going to move forward regardless. And so I think everybody in some capacity is a risk manager. It's the user's job to evaluate whether that's a risk to their career if they use that data and it turns out to be wrong, there's some underlying force at play, right? I think it's called Moore's Law, where the number of transistors you can fit on a chip doubles, I don't know the timeframe, but you get the broader point of that. And I think it's the same pretty much everywhere. The more tools that are available, the easier it is to make more tools. And sometimes I wonder, 30 years ago, if you wanted to build the same thing, how long would that take? How much would it cost? The cost of building has gone down dramatically, but the value has kind of remained. And so the return on investment just goes up year after year after year after year. And at some point that ROI on buying some data and trying to automate or doing something different doesn't have to be automation. The ROI exceeds the potential risk of doing.
Gregg Bienstock (31:09):
And to your point, I think if any of us when we started our businesses had to have a data center, we probably would've never started our businesses because the amount of money just associated with that is just massive. And all of a sudden there's this thing called the cloud.
Dan Silva (31:25):
Thanks Bezos.
Lynne Funk (31:31):
So I actually would be remiss not to ask you about Washington and the uncertainty going on in terms of tax policy, you name it. I was trying to remember when the Deep Sea headline came and I can't remember. I know it was recent, but the headlines are just constant. So from your seats, where do you see perhaps you planning out your business? Do things change? I don't even, who wants to start?
William Kim (32:01):
I can start. For our issuer business, obviously our issuers are incredibly impacted. Their costs are going to skyrocket if tax exemption disappears, which is a terrible thing for the country and for them in particular. But the advantage of using MuniPro in that context, it's like you have less resources, so you have to be more careful about how you use your allocated dollars. And so that becomes more compelling to use a lower, very cost effective solution like MuniPro. So we see opportunity there even in that negative scenario, obviously we don't want that to happen. We want issuers to keep tax exemption. I think on the bank side, I don't have a crystal ball, but if the whole market goes taxable, the question will be does that mean that everything's done on the institutional desk where there's $8 take downs and that kind of thing? I don't know. But I think regardless there's a very low percentage chance of it happening. I'm hopeful, but I dunno.
Matthew Smith (33:10):
Something like that is where some of the risk lies. If you make choices on who your vendors are and if your vendors aren't able to adapt to particularly the political volatility that we've seen recently, let's say munis lose tax exemption, totally against that. For the record, I think it does wonders for our community communities. Some vendors might not handle that very well. For us, it would be a bummer because we wasted all this time making muni curves, but at the end of the day, we just swap it out with the treasury curve and the model still works. So there's risk in going with a startup, but there's also a lot of value in going with a startup and that they're usually pretty fast to adapt and able to keep up with the market. That is changing more and more quickly.
Dan Silva (34:00):
For us. We model taxable bonds right now, and so a lot of our customers are not just modeling tax and bonds in Adaje, so it's kind of lends itself to either, but it would also be, I think a bummer for the market in general. So,
Lynne Funk (34:18):
You might want to use a stronger word in this.
Gregg Bienstock (34:25):
I'll just kind of add there. I think i's been set on a bunch of panels. So I think it's a completely different environment. It's unpredictable for me is probably the best way is you wake up in the morning and it's like, holy crap, what today? And it's different. And so trying to predict what's going to happen. I think it was a gentleman at lunch who said basically Trump's got basically a year, and so he's shaking things up in his first a hundred days. He's throwing a lot of stuff out there. Is it tariffs? Is it negotiating point? Is it something else? So there's lots going on, but when it comes to the tax exemption, I guess the way that I look at that is I don't have a crystal ball. I don't know what's going to happen. I am really concerned. I know some people are just kind of like, oh, it's not worry about it.
(35:15):
Same thing was said about advance for fundings. And so I do worry about it and I feel very important for our industry. I tried to say it before and it's really important that we get the message out there and not just speak to each other about it. So I kind of get on a little bit of soapbox here, put the story, get the story, write an editorial, get it in the Wall Street Journal. Use your resources that you have, make people aware of the importance and the value of the tax exemption. All that said, something happens. You lose partially, you lose all of it. There was another word used earlier, resiliency. This is a very resilient group of people and resilient industry. And I do think the public finance market, the way that we issue is different even if it becomes taxable than the corporate market. And so there's going to be a place and we will adapt. And that's part of what our business models, you made the point, right? We're agile, we're smaller companies, we're startups, we're smaller companies. We've been around a little longer. But the idea a startup,
Dan Silva (36:18):
You still call yourselves a startup?
Gregg Bienstock (36:20):
I don't. Not anymore. My old company was, but no, but seriously, I think the whole idea is that there's this resiliency and there's this adaptation that will occur. And my guess is a year from now these three guys, maybe even me too, we'll all still be up here, but talking about technology and how it's going to make your lives easier.
Lynne Funk (36:42):
I sure hope so. Are there any questions out in the audience right, as of yet? Oh, can you guys talk? Can you wait a second just for the mic? I can speak loud enough. Alright guys, just for the recording,
Audience Member 1 (36:58):
And maybe this is just that Dan or just trust in the municipal market is
(37:04):
trust in the municipal market is what has been built on, whether it be from an FFA with an issue or from the trading desk. I mean Matt talks about the profit motive behind what he was doing in the algo world and what he's doing with spline, he's making money off it, he's making things more efficient from a standpoint of the rest of the municipal market, which has been reticent to change because of maybe it is a trust factor. Maybe there's a three letter word company that has been around for too long that isn't intuitive and should be replaced. But can you talk about how do you gain that trust?
Dan Silva (37:36):
Yeah, I mean that's a question we deal with all the time. Trust is huge. And I think what we say is to folks is just put us to the test, right? I mean, there's no mystery in bond math, sorry to say. The same inputs in are going to produce the same outputs out. Now there are some nuances in optimization, the routines that are actually run behind the scenes when you're saying, okay, these are all my inputs, now tell me what's a good debt service or amortization. And I'm happy to say that we find lower Ts than that three letter company. So there's that. But yeah, new tools are always a little bit scary, but we all use computers now. We're not still using calculators and scratch pads and all that. I think we're going to move forward eventually. I don't think that all of us are going to be using that three littered software 10 years from now. I hope not. That would be a bummer for us. But also I think pretty miserable for the industry too. I think things move forward. I think things get better. I think things get easier to use and we're just looking for folks to be open-minded, try the software, put it to the test, and let us gain that trust.
Gregg Bienstock (38:55):
I'll just add to that. I think there's one other piece that is part and parcel of what you're saying. And we talked about it before as transparency. If what we're delivering to you is transparent and you can understand what we did and how we did it and see the result and see the underlying pieces of the puzzle, that's how you're going to gain trust. That's another piece of that equation.
Matthew Smith (39:14):
Yeah, I mean for us, for the most part, it's like where the team's experience is. We're not coming out of the blue and saying, oh, we can turn your business around without actually knowing how your business works. Everybody at Spline has some association with Munis in their past except for two, but we're training them. But it's a whole different ball game. I would imagine if you went in and tried to pitch something on what we do and you didn't fully understand their business and how it could actually help them. So I think familiarity is really important in telling the right story and establishing that trust because there's nothing more annoying than being told how to do something by somebody who's never done it.
Lynne Funk (39:58):
I would agree.
William Kim (39:59):
Yeah, echoing those comments, I think track record and visibility into what you're doing builds up credibility. So we've been here for six, almost seven years now, and all of our clients have resigned and we have this great consistent group of folks who trust us, and that word of mouth gets around. And so I think that's how you really build that. It's time plus the quality of the product and kind of the insights into it. When we produce our debt profiles, we tell you, Hey, it's different than the financial, it's by maturity, a tender cash fees, et cetera. So you can actually trace back all of the differences. And so I think that's important both time as well as how your product shows or proves out that you're right.
Dan Silva (40:43):
I would add that I think pretty much everyone at this table has been in the muni industry for their almost entire, if not entire careers. I have, I was a municipal advisor before this, and I mean, I know you were a banker, right? I feel like we've all kind of lived and breathed this whole industry, have dealt with the old technologies, know what's involved, know what the workflows look like, right?
Lynne Funk (41:10):
Any other questions
Audience Member 2 (41:10):
Over here, Lynne Okay. This is more about the predictive pricing. And when you're looking at tomorrow, it's going to be a great day. Non-farm payrolls at eight 30 and one second, we're going to know what's going on in the treasury market. How does the predictive pricing in Munis work when you don't have a lot of data set to be able to pull from, how quickly does it adapt to be able to come up with that pricing so then you can act or react to it as a trader, as a portfolio manager, maybe to sweep some bonds that are on the ECNs that have not been marked up because people are doing it the old school way. How do you do that in a predictive way?
Matthew Smith (41:47):
I think that's most of the secret sauce when it comes to muni price modeling. It's not the same as corporate bonds where you can look in the same qip or the same issuer and get prints five seconds away from each other. Munis are very, I think what makes Munis extremely interesting is that you have to borrow so much information from other muni bonds and the more muni bonds that you can borrow information from, the closer you can get to actually being on top of how the market's changing. You can combine that with outside sources of data. If you talk about ETF implied yields, right? There's these equity traders that are trading what they think the NAV of the ETF is going to be based off of what other factors are doing or what other things are doing in the market. Treasuries is a good example of that, and it's not always a hundred percent correlation, right?
(42:38):
I mean, the day after election day, you have a 20 BIP sell off in treasuries, and then we're able to pick up a 10 BIP sell off in Munis, and they sold off a little bit more throughout the day. But being able to take that information and turn it into Muni specific information is what makes these models special. And again, it goes back to the transparency of outputs, right? It's wonderful if I can tell you, oh, we use a million a million bonds to price this one bond, but how did that perform relative to where it actually printed? If a bond prints at 9:00 AM where was I at 8:55 saying that was going to print?
Gregg Bienstock (43:18):
I'll tackle it slightly differently. So the beauty of, and I think Matt, you guys do the same thing, but the beauty of what we do is it's real time. So as the data is coming in, the trade data comes in. So within three seconds of a trade being reported to the MSRB, it's in our system. It's being contemplated by our models. Same thing with quotes. Quotes. As soon as they hit our system, they're coming directly from clients. We're able to bring that information in and the impact of what happened in the jobs report or anything else that's going to change a price, it's going to change a trade, it's going to change. A quote is going to be able to be run through our models literally as that data comes in. So the data gets updated and the model runs constantly. The model also gets retrained hourly. So you're looking at the idea of, I think our model's contemplate something like 295 different factors that are going to go in there and everything from did that specific bond trade, did it get quoted? Or as you start to build out, the tree of the bond hasn't traded in two years, but you need a predictive price for it. How are you getting there? But the idea is that the models contemplate all those factors and then as responsive to the information as it's in real time.
Lynne Funk (44:35):
Okay. Any other questions? Well, I would love to thank the panel that went very fast, just like technology. Super fast. You guys were great. Thank you so much.