The Tech Update and Impact of AI on the Industry

As the array of technological solutions available to the muni industry widens, we will take a look at where the industry currently stands when it comes to the tech uptake as well as what latest technological advances are available to help the industry on its way to becoming more efficient.  Additionally, given that the muni market is not always transparent and liquid, will AI help here? Is AI impacting pricing analysis and how bonds trade?


Transcription:

Lynne Funk (00:12):

Good afternoon everyone. I'm Lynne Funk, Executive Editor at the Bond Buyer and I am delighted to welcome our panel of experts with us today for our technology panel. I'm going to introduce to my left we have Tyler Traudt, who is CEO and Co-Founder of DebtBook, and next to him is Will Kim, who is CEO and Founder of Muni Pro. We have Dan Silva, who is CEO and Founder of Adaje, and next to Dan is Gregg Bienstock, who is Senior Vice President Group Head of Muni Markets at SOLVE. And then next up we have Matthew Schrager who is Managing Director and Co-Head of TD Securities Automated Trading. And finally Matthew Smith who is Founder of Spline Data. So welcome guys. I'm delighted to have you here. We have a lot to cover, so let's just get right to it. As technology solutions become available in this market, we know it can lag to say the least, but these guys up here are going to talk about why it shouldn't and what opportunities are out there for this industry. So I think I'd like to kick off things first by asking you all, there's this nebulous AI question, what is AI? What does it mean? And there's a lot of different definitions of it. So I think it would be great is if you could all just talk about what goes into your models, what goes into your approach with technology when it comes to AI in the muni market, what's relevant now and what might be more so in the future? That's a big question, so who wants to kick that off? I

Dan Silva (01:50):

Sure,

(01:53):

Thanks Lynne. So at Adaje we use AI models to address a few different challenges in the muni market specifically, one of the key use cases for us is extracting essential information out of deal documents such as official statements and swap confirms. And this is really helpful not only from an elimination of a manual data entry perspective because let's face it, nobody likes doing manual data entry but also significantly reduces risks and catches errors pretty early. So we had an example recently where someone had an OS and we actually caught a maturity date that was next millennia 3054. And so got flagged as obviously an error and we're able to catch these types of errors early, which helps us avoid and our customers avoid potentially embarrassing and potentially costly mistakes, having to get them to avoid having a sticker, an OS and put out a supplement.

(02:55):

We're also developing AI tools to assist with idea generation. I think this is kind of an exciting area similar to how some people use ChatGPT to brainstorm or come up with first drafts where we have upcoming tools that'll help come up with different ideas around structuring to help folks tailor specific solutions to issuers credit or debt profile, subject to market conditions. This has really just to help folks spark ideas and help them move along towards more of a final product that could help lead to more innovative kind of ideas when folks are time constrained and have to put together structuring ideas and maybe response to an RFP and this just helps them get those ideas going.

Lynne Funk (03:40):

Excellent.

Gregg Bienstock (03:40):

I'll jump in next.

Lynne Funk (03:41):

Thanks Greg.

Gregg Bienstock (03:42):

So one of the things that just when we were talking about this panel, and it was actually, I was recently walking somewhere and there was something that was advertised and it called AI hype. This is not reflective of anyone on this panel, but I think a lot of what we all hear in the marketplace about what AI is or isn't, there's a lot of hype around it. A lot of it's just really technology. At the end of the day, when I think about what we're doing when we think about AI or what's called AI these days, so SOLVE is the company that acquired my company a couple of years ago. So we've been using natural language processing for years and what is that? What we're able to do is we're parsing unstructured data, so we're taking all these messages and we're able to parse that.

(04:23):

Is that AI? I don't know. I don't really think so, but some people think it is fast forward to where we are today, we are using ai, so we are in the business as some others are, right? And creating a predictive price. And so what is a predictive price? Being able to look at different buckets of data if you will. So trade data security, master data quotes, information taking almost 300 factors, building models behind that and then using those models to generate in real time predictive prices based on a specific size of a trade potential size. So that's more or less what we're doing right now. We also have a component of AI built into a new issue pricing tool. So one of the things that we're always looking for when you're doing new issue pricing and you're building scales using technology, you're looking for comps. Guess what? This is muni land. There aren't always comps, right? There aren't always primary or secondary trades that are comparable to whatever it is you're doing. So in those instances we've built some models that are again, leveraging kind of history, leveraging all the data that we have and being able to fill in the blanks as opposed to just doing straight interpolation or extrapolation. So that's just a little bit of what we're doing on the AI front.

Matthew Schrager (05:35):

Excellent. Yeah, thanks Craig. And I very much agree with your characterization of AI as a term that gets conflated with basically anything to do with computers nowadays, which is unfortunate. Maybe I'll start by just trying to deconflate a little bit without going too many technical details and then I'll talk about how we use some of these techniques. So when people talk about AI nowadays, I think it often refers to basically ChatGPT, which is a particular form of machine learning, machine learning being the broader umbrella term that covers a lot of these techniques and the technology that powers ChatGPT is something called a large language model, LLM. And a lot of what people, when we talk about AI nowadays, that's what they mean, but also I think it gets conflated a lot with other techniques. The broader topic of machine learning, basically using computers to do things that used to be done by people.

(06:38):

That's not new at all. That's been being done for literally decades in finance and I think it's important to be able to talk about them both of those things, but be able to deconflate them. So at TD we're a little bit different than I think pretty much everybody on this panel. We don't have a product to sell. Our product is liquidity. We sell pricing to the market, but much like Greg, we care a lot about coming up with the right price for a security at any point in time. And we've been doing that since 2013 or so. And although, I mean ChatGPT LLMs didn't even exist back then, but we have certainly been using techniques that would qualify as machine learning back from the very beginning. Things like large language models ChatGPT, they're amazing for a lot of things, but the kind of core use case is sort of in the name large language. They're good at language type stuff. What they're not great at is math. And so when it comes to pricing coming up with a good, it's a very mathematical question as far as I know in the industry, not a lot of traction for that particular type of AI when it comes to pricing. But the broader topic of machine learning again has been in use for literally decades to solve the problem of coming up with better pricing for different instruments, which is how we use it.

Matthew Smith (08:13):

I feel like we shouldn't have sat together.

Matthew Schrager (08:14):

Yeah, you sat too close.

(08:16):

Two Matt S's right next to each other.

Matthew Smith (08:18):

And we used to work together.

Matthew Schrager (08:18):

Unfortunately we did.

Matthew Smith (08:20):

Yeah. So we're going to say all the same things probably I'm trying to sell something. I guess that's the difference. Yeah, I mean I echo that completely. I always say in the shortest form possible, AI is machine learning. Machine learning is math with some code. The math is not always simple. A lot of it's been around for a while, but there are some new models out there. The interesting thing about machine learning is that it's generally available to everybody. So if these models are available off the shelf, why isn't everybody and their mom just building their own pricing models, which we're getting in that direction. There's a lot of competitors in this space, but I think when we talk about AI or ML, it's available, but to make use of it, you generally need to specialize in something. And so much about how effective AI machine learning they're synonymous, how effective it is in the market is so dependent on how you frame the problem. And so we use machine learning, probably similar models to some of our competitors, but I think what really drives some of our success in being able to predict price we're a pricing company is how we're framing the problem. So I come from a muni trading background, specifically a algo muni trading background,

Matthew Schrager (09:49):

Which one?

Matthew Smith (09:51):

I think it used to be called headlands now tv. And taking some of that trading intuition to frame how we're answering or how we're utilizing machine learning gives us very different answers than somebody that might be using the exact same model.

William Kim (10:11):

So AI for us, we view it as an incredible tool to really solve problems at scale. So at Muni Pro we also use AI in the data gathering process where it's an integral part of how we get our information and that feeds through all of the products we service, but also we're using it now a lot for classification. So one example is when we automatically pull 10 years of financials for a client in order to build these proforma financial projection models for their budgeting purposes. Yeah, it's great that you get the table out, but you need to understand each line item and part of that tagging and classification process is done at scale as opposed to one person saying, Hey, this is a revenue item, this is an expense item, and exactly how that fits in. You use AI to take a first stab at that, flow it through the models, and now instead of spending days and days building a budget model, you could do that for 50,000 clients. In theory, we certainly do it for the thousands of credits that we cover. So I think it's just another tool in the toolbox. So ultimately it's how the sausage is made. It's not necessarily as customer facing as like a ChatGPT, hey, solve this problem for me. But it's certainly being used as part of our product delivery.

Tyler Traudt (11:31):

So really quickly, nice to be with everybody. My name is Tyler Traudt. You can think about DebtBook as a treasury management system for government nonprofit office of the CFO. We build debt management software, cash management software as our newest product. We're really excited about our vision and all the comments here make a lot of sense. A couple that I thought I would highlight in three different ways that we're using AI. The first is making our products more accessible to our customers. That if you talk to any individual working inside the CFO's office and a governmental or nonprofit organization, you'll hear a lot about the hiring challenges, the staffing challenges, turnover, loss of institutional knowledge, lack of productivity inside the organization. And they look at software solutions and they see 'em and everybody here has got really great solutions, but it becomes really challenging to go and capture the value.

(12:24):

We hear that all over implementation needs to be really simple. We think that this solution could be really powerful and valuable for our organization, but getting to value seems really challenging. And so we talk about it and we say really getting started like AI should help us get started. And Dan mentioned this and others have as well. When you're onboarding a new customer, if you're pulling information from documents, being able to read and extract data automatically should get you to value faster. So that's a not very cool, but I think important one to mention. Additionally, for us continuing disclosure, we're building a financial reporting product that we're really excited about. And in financial reporting, being able to read and pull through a continuing disclosure agreement and set up these are the requirements, this is what your continuing disclosure says you need to do. And additionally you can imagine loading your last five years of disclosure filings and then creating a template off of that and saying, Hey, these values continue to change from year to year.

(13:25):

We likely think there's a task here. Is there a task associated with this one value that we have? And those are just a couple of examples in cash we think it'll be really interesting variance analysis. So you can imagine an organization that's looking all their bank transaction data, trying to understand do we have enough liquidity today, exactly the right amount of liquidity today. And then categorizing all those bank transactions and then being able to compare those bank transactions in your cash position to what you forecasted for the day or for the week and being able to run a program through that to try and identify where maybe you could tighten your forecast a little bit. So I would say at debt book, we think about it right now, getting started bringing our customers to value faster, making life a little bit easier. And gosh, if we could ease the implementation process, I know our customers would be more successful and everybody up here would be more successful too, for sure.

Lynne Funk (14:21):

Yeah. So none of you want to write a definition for the SEC of what AI is?

Dan Silva (14:27):

Hard pass, hard pass. Once they come up with it and we have to get regulated, then everyone's going to say they have a statistical model, not an AI model.

Lynne Funk (14:37):

All right, all we'll move on from that next year maybe. So when you think about the muni market, you all have these tools and sitting up here has sort of different approaches. We all know the muni market is somewhat slow to adapt or willingness. There's a willingness factor of do we really want to adapt, do we want to change? How does that affect the way you all build your product, how you interact with the market, how do you market your ideas? How do you sell this? Tyler, you want to start?

Tyler Traudt (15:16):

Let's work it back. It's like a draft. So I'll just share my experience generally on this and these folks will do a great job as well. I spent the first 10 years of my career in public finance investment banking, and then I was a municipal advisor for eight years and we would watch organization after organization come and present to us. And I always thought that the industry that we were working in would need to modernize and that we could benefit from better tools. I think that's really clear. I think everybody here agrees, but I never felt like the tools that we were looking at were really purpose-built for us. I felt like we were looking at solutions maybe that could benefit a different industry that is similar, but it wasn't purpose-built for us. And so when we started our company, we did that because we wanted to create purpose-built software solutions for these individuals.

(16:06):

And I think that innovation is hard and changing behavior is really, really challenging, but ultimately that's a decision that an executive inside of an organization is going to make. And if they see your solution and they see something that's going to materially improve their organization well beyond time and cost to put it in 10X kind of improvements over their current process, I think they will innovate. And I think they will look for those. I mentioned this a moment ago, but if you are in public finance and everybody here is in public finance, you'll know that there are really significant workforce challenges right now. There's a lot of institutional knowledge that is concentrated inside of the senior executives of the office of the CFO and government nonprofit. And those folks will retire. They will retire. And much of that knowledge has not been cemented inside of a modern system.

(16:57):

And a lot of it's in an Excel spreadsheet or a legacy homegrown system. And so I think there is absolutely a desire from these leaders to innovate, to modernize. That's very clear. And I think it's really up to the software companies, the technology companies to create solutions that deliver material benefits well beyond pain of implementation, which we just talked about and well beyond cost of the solution. And so everybody here has got a really wonderful solution and they'll talk about their point of view, their perspective, and everybody up here I think will be talking about how, well, I don't want to put words in your mouth, maybe you will, but for us, we want to focus on serving one customer type and listen to 'em and listen to 'em and they will tell you what to make and you make that thing, you continue to iterate and serve them really well and then the next customer will come and the next customer will come. So yes, building a technology company is incredibly challenging. It's incredibly challenging selling into organizations like governments and nonprofits do have more strict procurement rules that makes it more challenging. That's absolutely true for us and for the executives that want to innovate. Implementation makes it challenging, but it's really up to the software companies, it's up to the folks on this stage to solve that problem. And if you make something amazing, I think those executives really want to capture that value and I think they'll innovate with, you

Lynne Funk (18:18):

Guys can hop around however you want.

Matthew Schrager (18:19):

I would agree with that. I think the talking point forever is Munis are slow, Munis don't innovate quickly. And sure there's been some truth to that over the years, but at least in the part of the world that we deal with, which to be clear is primarily the secondary market trading of these securities. I just think it's decreasingly true nowadays. Honestly. I mean I think the has left the station. If you look at split up the market however you want by a DV or average daily trading volume of a security or AUM at asset managers, a huge portion nowadays of all the trading that gets done either by volume or especially by trade count or of the AUM in the market is dominated by players that have a lot of technology.

(19:11):

You look at the liquidity provision side of the part of the world that we come from providing liquidity to the secondary market. I don't know what you guys would estimate, but it's a very high percentage of trades are done by firms like us that provide algorithmic liquidity using machine learning or AI or whatever you want to call it. And that's been true for a while now. And if you look at the asset manager side of the equation, I think we're going to talk about SMA a little bit. I mean SMA accounts now account for something like a quarter of all a UM in the market up from zero effectively like 10 years ago. And the way that they've achieved that scale is 100% through technology and some of these large scaled SMA players have really impressive technology platforms. And look, I think ultimately I think to Tyler's point, markets are responsive to a good product and they are responsive to price and technology at the end of the day improves pricing and improves firms like ours ability to provide services at an effective price. And I think markets in various ways are and will continue to respond to that over time.

Dan Silva (20:30):

I would just say that the primary I think is a little bit slower than the secondary to adopt.

Gregg Bienstock (20:35):

I was going to make the same point,

Dan Silva (20:37):

So I wa,

Matthew Schrager (20:38):

Anyone in the audience?

(20:39):

We're getting there though. We're going to talk about that in a little bit too.

Dan Silva (20:41):

So I was like Tyler and MA before this life and I think all of us, when you start something like this, you have to have felt the pain a little bit yourself to think, okay, yeah, maybe something better could exist and then you go out and try to do that. But I was a municipal advisor, had to use a lot of the old legacy modeling software that's still is predominantly used throughout the industry. We're trying to change that, but I think we're seeing folks are becoming increasingly open to change or realizing like, hey, there are better ways of doing things. We've used this for 30 plus years, but maybe we should start looking at other things.

Matthew Schrager (21:18):

We're in the second inning.

Dan Silva (21:19):

I think that's starting to change. So I mean one of the things we've hit in terms of Lynne, you were saying how do we go about marketing and what do we have to do from a product development perspective is for us, we have all these nice bells and whistles and new technology, but one of the things that we've really seen as folks really want to see the old legacy features and capabilities still exist. You can't take something away from folks. So one of the things we've been laser focused on is making sure that while folks want to do away with the old systems that we're replacing and one of the things that they really want to see as us completely cover the breadth of capabilities that old system can do. And we're dealing with pretty complex things. It's not just plain vanilla new money level debt service.

(22:07):

We're having to cover things across the spectrum of reserve funds, capitalized interest funds, net, net funded escrows, premium capital appreciation bonds that are subject to coverage constraints and some pretty sophisticated capabilities around structuring. And I'm happy to say we're pretty much there, but that's really how we make our product stand out is for us being able to do the breadth of structuring capabilities for our clients, they need to be able to handle those real world complexities and I'm happy to say we do that a whole lot easier, make it a whole lot easier for them to do that with our tools than those old legacy tools.

Gregg Bienstock (22:47):

I'll just jump in just real quick. I think my experience, I got a little more gray hair. I do have hair than these guys. So if we're in the second inning, I started for preseason or something, but one of the things that we learned early on is that you got to talk to you, all right, we all have great ideas, but if you don't think it's a great idea and it doesn't meet your need then matter, we're going to be left with it blaming on ourselves. That's it. That's nothing more you can do. So when we think about what we can do with technology, is it something that is truly needed by a target audience, whoever it is that you're trying to solve the problem for? And then what we also learned early on is that it can be cool, but if it's cool and it doesn't meet the need or it doesn't fit in terms of the work style or the work out, the work need, kind of what you're describing here, it's not going to sell, right?

(23:41):

We've got to make your life better. So we have to create efficiencies and that's what technology does. We have to solve a problem for you. Are we solving a regulatory problem or are we helping you make money? And so when we think about all the things that we have done from a technological perspective and then wherever it is we're going next, it's always with the idea that are we helping you solve a problem? Are we helping you become more efficient? I was talking to someone earlier and at the end of the day it's like if I can take a task that took 30 minutes and get it down to one second, that's efficiency and that's going to help you do your job better. And then it comes down to data integrity and data quality. So that's how when we think about what we're going to do in the market, that's how we address that.

Matthew Smith (24:23):

I agree with everything that's been said. I think adoption is not a muni specific problem. I don't really think it's a problem in general. It's natural as a human to be wary of something that's not familiar to you. If you're camping and you hear strange noises outside of your tent, you think the worst. And I think it's not really different when people come into your office and start swinging around their machete and telling you how to do your job. That's just kind of a bad way to go about business. So from the very outset, we've tried to produce data products using all these fancy models and whatnot, but the outputs are very, very familiar. So our pricing product is bid and ask and it's by size and it's things that in the secondary market people are familiar with looking at. On the primary side, we're coming out with a product that will spit out a scale just like you would get from an underwriter or an eventually from an underwriter. So we try to create or incorporate new technology and do it in a very familiar way. And then there's a very specific reason that I went into pricing rather than anything else. And I think it's the value proposition is very exponential and no offense to software companies, I couldn't do it.

(25:56):

Pricing was very familiar to me and it's easier to go in and say rather than we'll save you X amount of dollars or X amount of time the money you're leaving on the table, here's how we can show it mathematically and we have something that can help you achieve it with your current workforce. And for me, that's a lot easier of a sell. So all power to you guys for rising the tide of a workflow. It's really hard, but it's very, very important.

Tyler Traudt (26:30):

Don't make me start walking through our value proposition and talking about that, but yes, selling money is a good thing to sell.

William Kim (26:36):

For us, I think throughout the conference I keep hearing the same three themes, right? It's like you've got aging infrastructure, you've got hiring challenges, and you've got revenue constraints. I think the same three themes exist with the clients that we serve, whether it's a bank, an asset manager or an issuer. And what we're trying to do is show them you can modernize all of the tech that you have that was maybe built 20, 30 years ago. If you're an asset manager and you built an option pricing model in 2006, like hey, you can get with the times and see the full breadth of development that's happened since then. If you're a bank and you're like, Hey, I can't hire enough analysts who can do enough work, hire muni pro, get 80% of the data collection process out of the way in the click of a button and then finally it's with revenue constraints. You got to show that we're cost effective. It can't be a dollar cheaper than hiring someone. It's got to be 10 times or a hundred times impact where you're really reaching everybody at an extremely cost effective way.

Lynne Funk (27:45):

So I kind of want to move into something you brought up the secondary market, right? Matt Schrager here that it is ahead of the curve at this point as algo muni strategies have become more common in the secondary market. Is there a way, how's that new liquidity potentially come into the primary it? How's it affecting how money gets raised, how it gets put to work?

Matthew Schrager (28:15):

Sure. Maybe I'll take a stab at this one. So again, our business from the outset was really focused on providing liquidity to the secondary market and we've become a very large liquidity provider to the secondary market. After we joined TD, we were a separate company. We got acquired by TD back in 2021 as a standalone company. We had essentially no visibility, no ability to participate in the primary market, but after we joined td, we started surveying the landscape. And what strikes you when you really look at the primary market to the point that a lot of people have made on this panel is a lot of it looks a lot like it did 20 years ago, especially on the liquidity provision side. And we thought that a lot of the same techniques that we've applied to the secondary market in terms of understanding what a bond's worth and who's going to buy it and how do we interact with all those people who want to take the liquidity that we provide. A lot of those same techniques we thought should be applicable to the underwriting of debt as well.

(29:34):

And the municipal space, unlike corporate credit actually has wonderful concept of a competitive primary market. And the competitive primary market has the great advantage that it is competitive, meaning that it's actually relatively easy for a new entrant like us to put our hat in the ring. And so we started, I want to say a couple years ago, something like that product basically built from scratch, but using a lot of the same algorithmic techniques and modeling approaches and AI and machine learning that helped us understand where bonds in the secondary would trade. There's a very tight correlation between that and where they should get underwritten. And so we started participating in the primary market in the competitive new issue space, but in a very automated algorithmic kind of way. We literally have two people working on it, but that's the beauty of technology. We've gone from having underwritten zero deals in the history of our business to this year, year to date we've underwritten hundreds and every one of those deals that we won competitively by definition is a deal where the issuer got a better yield than they would've otherwise. And so I think a lot of these techniques and the same benefits that technology and automation can bring to the secondary market, I think those same arguments apply to the primary market and hopefully the results are starting to show that.

Gregg Bienstock (31:06):

I'll add to that, I think you're absolutely right. I think it's principally a behavioral shift that you ultimately look at. We have platform that's about scale writing so different than what Matt's talking about, but conceptually it's the idea of being able to aggregate based on structural and credit characteristics, comps in the marketplace. And what we've learned over time is, and your model works for you guys, it works really well for you guys. What we've learned over time is that every one of you that are pricing a deal are doing it slightly different. Everyone has their own special sauce, so you want to play with variables, you want to be able to make a determination so we can generate for you in seconds based on what you put into our system, we're going to generate a scale for you and we're going to show you everything that's behind that, the comps and then you're able to look at it.

(31:56):

Is it based on primaries only? Is it based on primaries and secondaries? Is it based on are you extracting retail trades from that because you don't want to use them in there? Is it your transaction and negotiated deal and therefore you only want to look at negotiated comps? There's so many variables that go into the equation and what we have found as we've worked with clients, both bankers, underwriters and some issuers who are using our platform is that it's the access to the mass of data that is out there and the ability to extract an answer essentially with an easy button but based on the unique requirements of each of you in the room. And so I think you'll hear from, I think Matt's going to give a different perspective here, but there are tools out there to take what was that manual process and for some it still is a manual process and will continue to be, but to be able to create indicative scales at the click of a mouse, it exists today and it's going to continue to be refined based on your collective feedback and responses to what we all produce.

Matthew Smith (33:01):

I think this conversation gets a little bit heady quickly in terms of customization versus centralization and I think everybody would agree to an extent. Customization is good. You want it to match specifically what you're trying to answer. I think in terms of the product being useful to every participant in the market, there needs to be some sort of consensus and I think the easier it is for everybody that's involved in the issuance process to have access to that data, it makes the decision a little bit more simple kind of through and through. So our product is similar in the fact that you can select competitive select negotiated, but it's more so, hey, this is where you should start from and then keep your opinions not in a bad way to yourself. Take spread to that kind of consensus and incorporate your view into that and then keep it yours.

(34:03):

And then that's kind of what separates underwriters, that's what separates how municipal advisors are speaking to their clients about some of the scales that they're getting back. In terms of the broader question, does this secondary liquidity bubble up to primary logically? Yeah, I don't think it's going to better pricing in this secondary market is really going to spur some massive increase in issuance. If it's easier to raise money, people logically will raise more money. I think it's a hopeful yes that it bubbles up, but I don't think it's as straightforward as if you build it, they will come. I think there's some policy aspects to that that I am absolutely not equipped to opine on.

Gregg Bienstock (34:56):

I just want to add one other piece. It came up if you were here for the pre-session or some of the other sessions we heard from Dave Sanchez and he talked about pricing and responsibilities for pricing. So whatever it is you're doing, that's what they're talking about. It's an SEC priority, so whether you're looking at something that Matt's firm is producing or the ability to access historical scales or do those things, avail yourself of this information because really important to not only getting the best prices that you can, but also making sure you have good information for yourselves or your clients.

(35:33):

All I got.

Lynne Funk (35:34):

That's Okay, so I guess you kind of touched on this, but yesterday's or not yesterday's yesterday, yesterday's CDIAC panel on pricing. There was talk of as this market moves to more competitive issuance or if it does rather, what does that mean for your customers? Are you in a position to help?

Matthew Schrager (36:00):

Yeah, I'll take a stab at it. I think some overlap with the prior question, but I mean look, so there's competitive, there's negotiated, there's benefits to both, right? Neither are going away, but I do think that there are great benefits to the fact that competitive new issue is competitive. The negotiated process is very, and I'm certainly not an expert in the space, but is very relationship driven and that this is not to diminish the importance of relationships. They do matter, they always will matter, but a byproduct of how that part of the market functions is, and I could speak to this authoritatively, it makes it very difficult for new entrants to break in. Those relationships don't come easy and by contrast, you look at the competitive space again, we were able to start from nothing and become by deal count like the number two underwriter in the space within the space of a year or two and forget whether that's a good story for us or not.

(37:04):

We don't matter. The market shouldn't serve us. The thing that matters is that we produced better outcomes for the issuers, they got better yields, they got better financing, and we're just one firm. But to the extent that the competitive process allows firms, not just us but firms, more firms like us to compete on a more even playing field and produce better results for issuers, I think that's a great outcome. And so my guess is to the extent that that story is true, it's certainly negotiated is not going to go away. But to the extent that story is true, I think you will see continued movement in that direction.

Dan Silva (37:46):

On the workflow side over, I feel like there's the pricing and the money side and then the workflow guys definitely different.

Matthew Smith (37:55):

For the record, I did not mean that in a derogatory set.

Dan Silva (37:59):

So I know for our advisory clients the competitive is becoming a bigger thing. I think that's a theme we've seen and they're seeing more competitive deals and kind of a push towards competitive deals versus negotiated and one of the things we're seeing is speed and being able to evaluate bids and take in those bids and look at them quickly, award the correct bid and as a, there's a little bit of stress to that, making sure you're not making any mistakes and comparing them and rewarding the right person. We've actually been approached by a number of our clients asking for tools to help facilitate that. So I think there's kind of a demand and an interesting thing to do from a workflow perspective to make it more efficient to process the bid, manage the bid process. I think in the future there's potential to even have folks submit bids into a particular platform to make it easy to just do side by side. Right now, a lot of times it's pdf f bids agreed, there's no reason for that nightmare. It should be automated in terms of being able to do easy apples to apples comparison of these bids coming in. It makes probably your life easier. It probably makes the MA's life way easier. We know that for a fact and also avoids potential mistakes like making the wrong award. So I think there's a lot of value there just in terms of time saving, time reducing risk and making that whole process a lot easier.

Gregg Bienstock (39:22):

I'll just quickly add, I think the negotiated side of the business serves real value. I'm not going to spend time talking about, you all know what that is. I think whether it's what Spline is doing or whether it's availability, easy access to data, that's going to help to the point that you made Matt, right? In terms of just it's going to help generate better pricing when there's better information. One of the things that we've heard over the years is we need more data, but more data is great if you have tools to be able to utilize that data and that's what it's about. So at the end of the day, if you get more data and you can access it and utilize that to get better pricing for your clients, what's going to make the difference, whether it's on the competitive or negotiated side said.

William Kim (40:09):

I think part of that previous panel was about transparency and making sure that issuers get the best outcomes so it doesn't have to always be competitive as a tool to do that. I think in the negotiated space for our issuer clients and MA clients specifically, we do a lot of break pricing analysis. We show what was the original credit spread versus where it's trading immediately in the market and kind of tracking that over the next month or couple months. So that technology is there for the negotiated side, but on top of that there's a lot of situations where negotiated is the only answer or the only appropriate answer, right? You're seeing enormous amount of tenders in the market. I don't even know how you would do that competitive. You have to take all these bids and you have to structure around it so that it's sensible for clients and I think bankers provide that incredible service and on top of that, they're showing not just the refunding for the tender directly, but comparing it against the hypothetical, what if we waited until the future? They show the option value analysis, you're not getting that in a single bid item. So I think there's a lot of value that the negotiating market plays.

Dan Silva (41:13):

I don't think negotiate is going anywhere, anytime. I don't think anyone up here saying negotiate is going away. That's definitely sticking around.

Lynne Funk (41:19):

80% of the market. Yeah, I don't think it's leaving anytime soon either.

Tyler Traudt (41:23):

Just a quick perspective, I thought it was worth sharing listening to these good comments earlier we talked about AI enabling easier implementation. That was the perspective that I shared, being able to go and capture the value. We talked about leaders and you talked about maybe a lack of willingness to change. Maybe our market's a little challenging to change behavior and those things are both true. As I think about the leaders though inside those organizations, the leaders that we get to talk to, that these folks get to talk to, there's absolutely no doubt that those folks want to do great work. They want to do the best possible work to produce the best possible outcomes for their organizations and their communities. I think that is universally true every day they are presented with so many diverse challenges. The treasury teams are not just doing municipal finance, this is a small part of their job.

(42:21):

They have a million things to do and because of I think a lack of modernization, which I think is in part brought about by a lack of investment in the space, which I think folks here are trying to change that bringing more purpose-built products to them. I think a lot of those treasurers and those CFOs would say, yeah, I'd love to spend more time doing the more strategic work, but I'm trying to do the work that has to happen every single day. I've got some fundamental operational work that has to happen. And if you're doing it in an Excel spreadsheet, if you're doing it in a legacy system, you're experiencing turnover at exactly the same time. So now you're trying to ramp your new employees on said Excel spreadsheet or legacy system. It's just going to be really challenging for a leader to elevate to do the most impactful work that they could possibly do.

(43:08):

I think that's true, and there are probably certain CFOs in this room that are saying, I don't really have that problem and maybe they're gifted, they don't have the workforce challenges that some folks do, but those exist. And so I'll just say something that I'm excited about. I'm really excited to see these conversations in two years, three years, five years, because I believe that as these CFOs adopt software to materially improve their operating efficiency, to improve their visibility into their data, to make their teams more resilient, to be able to ramp new employees here all the time, we want our people to work with the data, not on the data. We want them to use it to move the organization forward. They're going to do more strategic work and they're going to produce better outcomes. And I think that negotiated competitive. I've got no dog in that five at all. It sounds great. I think ultimately CFOs are going to spend their time working with their lawyers and advisors and bankers to come up with the best possible financial outcome for the organization, but if they don't have any of the time to do it all day long, they're going to have less time to go spend doing some of the impactful stuff that's there. So I'll just be really excited to see those organizations and how they operate in the future and how that impacts some things like this.

Lynne Funk (44:26):

Great. I want to give one chance here if anybody has any questions for our panelists. Any questions? Alright. Well I would ask Tyler, I think you kind of just answered the question I was going to ask from each of you now.

Tyler Traudt (44:42):

I'll say it again. If you want me to say it again, I'll do it.

Lynne Funk (44:44):

Say it again. You want to say it? I would ask perhaps the rest of you to kind of we're running short or running up on time here, what would you leave this audience with? If there's one thing that they really should be thinking about, say in the next year till we're up on this stage again, would that be?

Matthew Smith (45:05):

I think as each year goes by, all of us are going to be asked to do more with either the same or less. So I think generally folks are doing themselves a disservice if they're not at least thinking about really succession planning. It's been kind of the way that a lot of other markets have moved and I think there's a huge weight on us. We keep saying data and I just have flashbacks to a science teacher I had in the third or fourth grade and she was explaining the difference between data and information and it's like, that's great. We have all these random numbers that mean basically nothing to us. It's our job to turn that data into something useful, something actionable. And I retract part of my previous statement where I think issuance won't increase because of better pricing. I think it will. I think the workflow guys will do it, making it more obvious where there are places to refinance or raise money in opportune times. Thanks.

Gregg Bienstock (46:20):

I'll just chime in. I guess in terms of what, looking ahead, when we started our company 14 years ago, we had an idea based on what we were told by certain market participants. And since then we've used that as the evolutionary tool for our business. And so I'd say it's really, we're all going to have ideas what we think, but it's really incumbent upon you all to keep pressing, pushing us as supposed innovators. We innovate based on what we see as an economic opportunity or an efficiency opportunity, but that comes from you guys. The other thing I just want to go back to early on when we talked about ai, the hype thing there and ChatGPT and stuff like that. What I would say is as you go forward and are exposed to all these tools, use it responsibly. There's a lot of crap out there and so validate especially something like ChatGPT, garbage in, garbage out. Same thing with data coming upon all of us and our companies and what we do to make sure the data that we're delivering to you, the information that we're creating from the data is good and sound. And if not, we got to find a way to make sure we do it better. So just use it responsibly.

Dan Silva (47:45):

I think that there's some really cool software out there right now and I think it's pretty powerful and I think folks, it's really easy to kind of stay with inertia and do things the way the old way you've always done, but I think that's leaving a lot of money on the table and I think there's, I'm of course incredibly biased, but I think across from the pricing side to some of the workflow stuff these guys and myself are doing, I think there's a lot of efficiencies to be gained, a lot of risk mitigation to be done. I think you can do a whole lot more with a lot less than you might realize, especially considering some of the staffing issues Tyler was mentioning. So I want to keep pushing that drumbeat of folks being open to new innovation and trying to test out new products. Give us your feedback. Some of the best ideas we get is from our customers on a customer driven development. So that's what I believe folks with.

William Kim (48:54):

I think it's really easy to get caught up in the day-to-day. It's like there's always the next bond deal, the next budget cycle, the next board meeting, but ultimately sometimes you feel like you're running on a treadmill, right? You're like you're running and running and running, but you're not really moving forward. I think investing in technology, while it does take some time to review all the kind of options that are out there and companies that are providing these great solutions, that's what will move you forward and set you up for success for next year and the year thereafter. So take the time to check out what we've got. I think we've come a lot further than you might expect. I think technology's evolving extremely quickly and a lot of your peers are doing so as well.

Lynne Funk (49:37):

You want to leave with us with anything?

Matthew Schrager (49:41):

I don't have some grand moral of the story for everybody, but I think use the word inertia. Inertia is the enemy. It's very easy to fall victim to the way it's always been, but your peers and competitors, not all of them will. And the ones that are willing to overcome that, I think history has shown in market after market will be the ones who win. And so I guess if there's some grand role to be had, I think it's be excited about and willing to try out innovations and it doesn't have to be scary. These things can be really, really helpful, really powerful, can really help your businesses, your organizations. So I would just encourage innovation.

Lynne Funk (50:33):

Excellent. Thank you all so much. Really appreciate your time and your insights. This is obviously a discussion that's going to continue and I look forward to continuing it with you. Thank you everyone. See you out there.

Matthew Schrager (50:43):

Thank you, Lynne.