Latest on the Tech and AI Front

Transcription:

Lynne Funk (00:07):

Everybody. We're going to get started here. I know we're the last panel standing in the way of cocktail hour. Fun fact, if you leave here, you can just cut to the front of the line and say tech panel and you'll get a drink. Just serious. Hi everyone, I'm Lynne Funk. I'm Executive Editor at the Bond Buyer. I'm delighted to welcome a panel of experts in technology. This is the latest on tech In the AI front. We have a great panel of experts who will discuss how technology and artificial intelligence is affecting the public finance industry. I would like to welcome Jennifer Fredericks. She is Sales Director at Solve Advisors Inc. And she's also the National President of the Board of Women in Public Finance. Next we have to her left, William Kim, who's Founder and CEO of MuniPro. Next we have, oh, I did get you out of order, huh?

(01:05):

Dan Silva, CEO of Adaje, Inc. Tyler Traudt, who is CEO and Co-founder of Debt Book, and Matt Schrager, who is Co-head of TD Securities Automated Trading. So welcome to you all. So glad you're here. I think one of the things that we wanted to kick this panel off with is artificial intelligence. It's just bandied about, talked about in so many different ways, and it depends on who you talk to in the industry of what it is. So I'm kind of going to open up to this panel to tell us what AI is in the muni industry, and I think maybe I'll kick it off with Dan Silva.

Dan Silva (01:42):

Thanks. Oh, there we go. It's working. Thanks, Lynne. Yeah, AI is kind of a loaded term. We had a discussion a couple of days ago and there's Chat, GPT, LLMs, that's kind of like the de jure hot thing, and they have some applicability in the space, but there's a bunch of other AI related technologies that folks have been working on for a long time. At Adaje, we use special natural language processing technology to pull information out of documents. That's one form. There's neural networks folks use with respect to figuring out the best price in terms of secondary pricing. So there's all these different applications and I think it becomes easy to kind of get them conflated and note we're talking about. And of course there's really hyperbolic statements being made about the capabilities of these systems. So I think it's kind of on us to be, as the folks that are developing it and trying to convince you all to use it, be really clear about what it can and just as importantly can't do and what it is, I don't know. Anyone else wants to add?

Matthew Schrager (02:53):

Yeah, I think the way that you framed that is really accurate. AI right now is a term that is ensconced in a giant hype bubble, kind of like in my opinion, the word blockchain from five years ago. But I think with a lot more real world impact to be had. But when people talk about AI, especially as it pertains to fixed income, I think it, as you mentioned, Dan Conflates a whole bunch of different types of techniques that have been used for, some have been used for literally decades and some are not really used at all today. And when we talk about AI, I think it's important to kind of disentangle those things, right? So AI, in my view, the new thing, the new hotness, as Dan mentioned, is a very specific technology called LLMs or Large Language Models, and they are amazing. That's ChatGPT, and that's all of the tools that everybody's heard about.

(04:01):

But realistically, I think the way to frame it is that these AI tools are part of a broader continuum of technology and machine learning. And that last part of it, the LLM AI part of it has barely, we're still figuring out what the use cases are, but there are tons of great use cases that have been utilized for literally decades in other parts of machine learning technology, et cetera. Some great examples on this stage. So happy to dive into as much detail as is helpful, but I think it's a really great way to frame it, Dan.

Lynne Funk (04:46):

And I think at Solve, we're using it for things like filling in the gaps where there are no comps, figuring out what we just launched. We're about 30 to 45 days out from being in full production, but we've got a beta predictive pricing model that's using AI. I think it's a little sketchy when we start throwing out AI to be the collective and all be all, I mean, we've been parsing data for 15 years. We don't call it, it's not AI, it's parsing data and I mean it's not going to make your coffee, but I think in the muni industry, it's going to start to replace some of these mundane things. I think about the way that I and some of you about the same Gen X Age group learned how to find data analysis. You were digging through Emma and all other data sources for hours and hours. You can do that in seconds now. Why not change the conversation? So we're actually analyzing the data and having intelligent conversations rather than calling data. I don't know about you. That's not what I want to start my career doing.

William Kim (05:56):

I think that the way we use AI similar, we use it to extract data, but what really gets people excited is being able to interface directly with it. They imagine all the possibilities. I think AI is not at a point where it's going to take everyone's jobs, which is kind of the major fear out there, but it does open up new opportunities to achieve at scale, right? Whereas you had a couple people who could cover only a small subset of the universe. Now you can look at the entire universe of bonds out there. And in addition to that, you can look at, go deeper in terms of your analysis to really isolate exactly what term in the document that you're looking for or in the case of pricing, like Jennifer mentioned, to get a more accurate data. So even though there may be a lot of criticism around AI like hallucinating and whatnot, I find it actually improves accuracy because it gives that scale to the companies who are using it in order to do more with less and eventually get to a more accurate stage by combining AI with traditional humans tracking it.

Tyler Traudt (07:05):

Hey everybody. Tyler Trout, really excited to be with you. My company DebtBook creates treasury management software for office of the CFO public finance. So debt management software, cash management software, echo, everything that these folks have shared. When I first started talking to my co-founder, Eric Peltier, our chief product officer about AI, he immediately reprimanded me and told me I had to say AI ML every single time because I'm talking about machine learning and it's not artificial intelligence. He's a little prickly about it. For us, in the work that we're trying to do, we think about it as just efficiency. Is there a way that we can bring some efficiency in the way that our software, these folks software helps their organizations? And so very simple example of this, and it's really more on the cash side, is can a computer ingest all of the bank transactions that some of your customers or your teams cash management teams are using and help automatically categorize those transactions?

(08:03):

So if you've got a bunch of transactions from a bank account, can a machine go in and say, we've seen this before. We've seen it on this date, we've seen this description. We're going to go ahead and categorize this as this kind of transaction and not fraud, or this is really suspicious. We think this is fraud. So when we say AI, we say AIML and true, there's going to be nice steady gains and efficiency over time, but it's not going to be perfect. And today, at least where we are, we're just trying to find ways to incorporate it to make the data upload process simpler. One thing that would be perfectly satisfying for me as a former investment banker and municipal advisor is make continuing disclosure a little bit easier. My favorite days have been spent going through bond ordinances trying to piece together different covenant calculations that our friendly lawyers have spun into a wonderful maze for me, and so excited about potentially having something that I can load a continuing disclosure agreement into and have it try and summarize for me and piece together the additional bonds test. And maybe if that could happen in line with a financial statement, that can be pretty exciting. That's going to be really hard. It's going to take a really long time, but hopefully some nice gains here in the future.

Lynne Funk (09:14):

Excellent. So I think what, when we prepped for this, also, we came to the conclusion that each of our panelists really have different seats up here, but what they're trying to do is create solutions to problems. And it's hard in this industry where because different constituencies really face different challenges, but the end goal is oftentimes the same, make the industry function better and more effectively and efficiently and transparently. So I want to touch on three areas in the market, issuer, sell side, buy side, and how these folks kind of play into the different aspects and what they do and what they provide. So I would like to start off with the issuers. Oftentimes, I think they're the group that perhaps struggles the most from staffing levels to market expertise to cost limitations. I probably start with Tyler, but what exists today for issuers that did not say even a few years ago, and then anybody else who wants to chime in, let's go from there.

Tyler Traudt (10:17):

Sounds great. Thank you. So I started my career at Citigroup Public Finance Investment banking, RIPC. I spent eight years as a municipal advisor, and when I thought about my customers, I didn't really know what they were doing all day long. I needed financial information from them to do my job. So if I was doing financial planning and analysis work as a municipal advisor, if I was doing transaction execution, I really needed documents, I needed information out of those documents to do my job. And as these folks all know, the more time I've spent five years now working with treasury teams across higher education, healthcare, governmental, the more and more I've learned that their fundamental tech stack hasn't really been improved tremendously in quite a long time. So here's an example of this. We have a university customer debt management team uses our software to improve their operating efficiency, give them better visibility and their information, save them time with financial reporting, a lot of different benefits for them.

(11:19):

Their cash manager leaves. The debt management team is now doing cash management. The debt management team doesn't know how to do cash management, but the debt management team is doing cash management. And as we spent time with the treasurer, we began to understand that they really need a platform of products that's purpose built to help bring efficiency to the work that they're doing all day long. And I'll just be clear, I had no idea what they were doing. 10 years of experience in public finance, investment banking, and on the municipal advisory side. And so I'm really excited about the amount of investment that's coming into our space. Let's say I was on spring break with my kids and my family. Very recently I met a venture capitalist and I told him that I made software for governments and nonprofits. And you know what he said to me?

(12:02):

He said, why? He said, why would you do that? Sell to corporations. They don't have the same procurement rules. Your life will be much better. He teaches at Harvard and he's like, in my classes, I tell all of the students do not start companies that are selling to governments and nonprofits. And I knew that, and these folks here will know that. And it creates a massive challenge because we have a lack of innovation rooted out of a lack of funding for companies like ours in this space. But these folks will do a really great job of talking more about it. I'm personally really excited about bringing modern software to these organizations to help their entire team, whether you're managing your debt, which could be financial reporting, it could be payment process, making payments on time. It could be cash management, daily bank reconciliation, cash forecasting, trying to understand how much cash we're going to have in six months or into the investment management side of it.

(12:58):

There's just a tremendous amount of work that these organizations have to deal with every day. They haven't modernized for a number of different reasons. Lack of investment is one of those things, and they're just grinding through spreadsheets and old processes, and it's a massive, massive challenge. So lots of investment, all good things for this industry, and we certainly want to lift them up out of the operational, eliminate the not so fun stuff all the time. Accounting can't, it's not always really fun. We want to eliminate a lot of that work for them so that they can go be more strategic and do more important work.

(13:34):

What we bring to our issuer clients is an independent capability, and that's what our issuer clients really rave about. The main thing we do for issuers is financial forecasting. So we can take their existing financials, automatically digitize them, and then look 5, 10, 30 years down the road how all their debt issuances look, their refunding opportunities and how it impacts their debt service coverage ratios and all these metrics that normally they would have to go out to their FA or their MA I should say, or their investment bank to get. But having that capability in-house where you have not just the debt management, but also the financial forecasting and the ability to really track your secondary trading to be more informed as you go into a pricing or as you're presenting to your board. And so I think that that technology wasn't there previously and it was really manned by folks on the ground. But as you know, it's difficult in terms of hiring folks. A lot of people, there's a lot of attrition in the muni space. So folks when they come to us can have that technology and not be so reliant upon all their other financial partners

Jennifer Fredericks (14:41):

At Soft. We do a little bit with issuers, not a ton. We have an issuer disclosure management system. We're not a dissemination agent, but we do provide that. And we do have a few issuers who also, there's a quote from them, they like to be as smart as their bankers, so they've selected some of our products that give them additional insight just so they can kind of watch things the same way their bankers do.

Lynne Funk (15:06):

Okay.

Dan Silva (15:08):

Our customers at adage are pretty much split half and half issuers and then advisors and bankers, those folks that you have to go through to get these deals done. I was a municipal advisor like Tyler was at one point in my life, and everyone in our space has to use bond optimization software. And it's pretty interesting that the state of affairs is such that the predominant bond optimization software when you're trying to model out these transactions was developed in the early nineties and is still predominantly used pretty much everywhere in all the major investment banks, all the municipal advisors, a lot of issuers in fact use the 800 pound gorilla in the room. So we're the first to start actually converting folks off of that platform. And I think you're going to see this throughout, it's a recurring theme, but a lot of folks are realizing that these legacy systems, that there's better ways of going about doing the job.

(16:10):

I think someone mentioned efficiency. At the end of the day, we're just trying to get things done a little bit faster or a lot a bit faster than what we are used to traditionally. So for our issuers, I think there's also an aspect of democratization of access to tools like this. You don't need a PhD in finance to actually model up your next bond issuance. And so we try to reduce that threshold in terms of accessibility, being able to use a piece of software and say, okay, I want to think about what a new money deal might look like, a new commercial paper transaction. How might this look layered on top of my existing debt? I want to model a tender. It should not be rocket science. We try not to make it rocket science. So I think that's like technology can really lower that barrier to entry to being able to have more control like Will was saying over your own processes, over your own data, over your own view into the future as you layer on these new transactions.

Lynne Funk (17:10):

So I'm going to stick with issuers for a minute here. We don't want to go too far down into the FDTA, which for those folks who in the room who know it's SEC issuers have to comply by 2026, it's a big concern for issuers. And I think I'd like to ask, I think probably Tyler might have some thoughts on this to start, but how perhaps state governments might deal with this new regulatory regime, but then I think I'm going to pivot to overall disclosure more broadly and how that's kind of advanced in this market. Transparency and data is often brought up as lacking for muni. So it's a twofold question.

Tyler Traudt (17:50):

Well, Lynn, coming off the legislative update, what if we consider reading the FDTA here and try to figure out who could not fall asleep in the crowd? We could do that. You can do that. So we're obviously learning a lot about the FDTA and it's going to be a lot of work for many individuals in this room, many borrowers, office of the CFO, the professionals that are going to be absolutely needed here. And we're going to learn a lot more hopefully really soon as the SEC works their way through the rulemaking process. My assumption, operating assumption as of today is that everything that's filed into Emma is going to have to be produced in a machine readable format. I might not be right about that, but that's I think what will ultimately happen. I think when you really think about that, and many folks here have enjoyed many hours filing things on Emma and pulling information down on Emma is probably one of the favorite pastimes of folks in this room.

(18:52):

You'll respect the tremendous variety in the types of information that's being filed on Emma, an offering document, A POS, you're going to have an audit that's going to be filed on Emma. You're going to have material event notices, you're going to have budget documents that are being filed. It's a real challenge, quite frankly, thinking about the work that's going to have to happen. And as we see it today, there are going to be clearly many solutions, a number of different solutions. There will be some organizations that have invested in audit compilation, software tools, organizations that have got financial reporting software at Workiva as an example of this. Of course, that document will be able to come out in a machine readable format. So that's pretty straightforward, but that's just the audit. What about the rest of the disclosure? And so you would think that there are going to be a number of solutions where you can take an Excel spreadsheet, a Word document, don't change your process today loaded into some software that a combination of people and a computer will go through and help you comply with that rule.

(19:54):

It's going to be a team effort. Absolutely. We've got a couple years or so to really get ready for it, but hopefully it's an opportunity for all this financial information finally available in machine readable format for some pretty cool software applications. I'll say just one thing and then I'll stop hogging the mic here. If we are only going to allow or ask of folks to take an audit, don't change your process and let's get that into a machine readable format so a computer can grab it and we can analyze it and present it, and we can do a lot of different things with it, which is I think where most of the market will go, that's fine. But we're not addressing the fundamental challenge of financial information being available. In my belief, it's a little bias. My belief is lack of modernization, lack of investment in the fundamental systems that treasury teams have across all of their different functional areas, and whether it's payments or cash or investment or debt, all of those different areas, it means that these organizations and these individuals, they're struggling. They're struggling and they need to find more time to do more strategic work. And so finding a way to help them invest in the actual creation of the audit, that to me is a stronger longer term solution. Although I fully acknowledge that in the short term, a lot of folks are going to look for an easier button and they're going to hit the easier button. And we would certainly like to be able to help folks on both sides of that.

Jennifer Fredericks (21:23):

I think at the most basic level, I think we owe it to issuers. We owe it to the market to be able to compare apples to apples. The first thing I am hoping for FTTA is that we start using standard taxonomy. I mean, I can sit in 15 different rooms and seven of those people are going to use the same wording and it's going to mean something completely different. And I think being able to compare things, being able to see the county I grew up in rural Indiana, be able to have their bond compared to another county somewhere else and actually give it the audience it deserves. It's what we owe the market.

Matthew Schrager (22:06):

And I can maybe speak to this from a slightly different perspective. So very briefly, just to set some context, we at TD Securities automated trading, we are focused on liquidity provision primarily in the secondary market and the types of data that we're talking about here. On the topic of transparency and disclosure more broadly, we have a very algorithmic, very quantitative business. We do thousands and thousands of trades a day in the municipal bond market. And the only way that you can build a business like that is if you have access to good structured data that is available in some sort of an automated format. And a good example here, and this is probably something that a lot of us take for granted at this point, is the existence of Emma or the MSRB feed, just trade prints, right? There was a time where that didn't exist.

(23:14):

Trace would be the analog for corporate bonds. And in absence of that data, it's impossible for businesses like ours and probably a lot of yours to exist. Okay, why do we care about that? Nobody cares about our business in particular. The reason that we care about that is ultimately it allows for to hit on a word that you're probably going to hear a lot in this conversation, much greater efficiency in the market overall. And a stat that I like to quote probably far too frequently, but so we've been in this business since about, call it 2015 or so. And I think one of the most important metrics of efficiency in trading in secondary markets is transaction costs, bid, ask spreads. And if you look at the size of bid, ask spreads in the muni market, I don't know if a lot of people realize this, but over the last call it 10 years, particularly in the smaller sizes of transactions that make up 97% of the trades in this market transaction costs have declined by something like 50 to 80% depending on the sizes that you look at.

(24:39):

And I think you can draw a direct line from the existence of high quality data like the MSRB feed, for example, through to that level of increased efficiency in the market, which is a great outcome. Right. Speaking about the FETA in particular, obviously we're not using that data right now because it largely doesn't exist, and it would certainly be an integration effort on our side to bring in these new data sources, but there's no question from our perspective that were we able to access this kind of financial data in more standardized formats that it would allow us to further improve liquidity in the market, which is ultimately, I think a good thing for everybody

William Kim (25:34):

With respected A-F-D-T-A. I think our issue is really with the implementation of it, it would be a great thing for the market if we had all these issuers digitized their financials. But when we looked at one of our clients and we digitized all of their financials automatically, one of the main or major pain points was the restated financial is due to GasB's change in lease accounting. So almost every single issuer had to restate the prior years and change the different categories of the debt. And so looking at proposals from XBRL and others to question this, how often will it be able to be updated and can you upstream changes to the taxonomies for every single given issue, right? There's 50,000 or so active issuers, that's extremely hard to create one consensus in terms of a singular way to look at things. So I urge the MSRB and the SEC to consider when you're actually implementing this, how is it going to work going forward and not just today

Matthew Schrager (26:43):

And one point in time.

Lynne Funk (26:45):

Right. Let's pivot here for a second then to not for a second, for a few minutes to the sell side. And I'm curious from all of your seats, how, Matt, you touched on a bit, but how has it aided in workflows and getting primary market deals done and then secondary market trading of those deals? I'll open it.

Matthew Schrager (27:07):

Sure. Yeah, I can speak about this for far longer than any of you probably want to hear about it. Some of you are nodding and smiling in the audience technology. Again, the panel is labeled ai, but we should talk about technology more broadly as we've discussed, is an inextricable part, an essential part of at least how we go about providing liquidity to the secondary market? Like I mentioned, we are a very algorithmic quantitative automated business. We've been that way from the very start in 2015. And when you look at the muni market, there are what, like 40, 40, 50,000 trades a day. The vast, vast majority of which are for certainly less than a million dollars of notional. And I think it's like 90% or for less than a hundred thousand dollars of notional. When you look at the marginal cost of providing a quote to a five bond RFQ, if you have a human doing it with no workflow automation, the math doesn't work out, you can't do it. The only way you can do it is if bid as spreads are extremely wide because you can charge for it, right?

(28:34):

As firms like ours have incorporated, it's not just us. There are of course, other firms approaching it similarly have incorporated various forms of technology. Some which are just sort of, I would call workflow enhancements, some of which are definitely what people would call machine learning type tools. But as these tools get incorporated into sell-side trading platforms, that dramatically improves the cost structure of providing liquidity, which in turn gets passed through in the form of tighter bit as spreads lower transaction costs. And again, I think that that's a really good thing. And what you end up seeing is that as it becomes easier to trade, you get more trading, which you've definitely seen last year was a record year in terms of number of trades in the muni market. I could wax poetic for hours here about all of the different ways that we incorporate technology, and I'm happy to follow up on any of that if it's helpful, Lynn. But at a high level, again, it really is, it's an efficiency enhancer that produces better outcomes for all participants in the market.

Dan Silva (29:51):

I can speak to, our focus is on the primary pretty much exclusively. And so our focus is really on making sure folks can model up all their options with respect to what's possible. And historically, this is an extremely time consuming process, right? If you want to evaluate what your transaction might look like from a primary issuance perspective, which every issuer and every banker and advisor wants to do that, they want to have a good sense of what are my different options? And you've all probably witnessed the kind of scenario analysis paralysis that happens when you have all these different options in front of you. Technology and specifically the type of technology that adage is interested in and delivers, helps evaluate those scenarios a lot more readily and a lot more easily. Again, trying to lower that barrier to entry, to being able to understand what your options actually are when you're approaching a primary market issuance from all the way in the ideation phase to when you're literally about to go and price the transaction and have it offered to investors. I think there's a lot of value in being able to do that quickly, understanding what that transaction looks like, understanding the relative risks and benefits of one structure versus the other. And so that's kind of what our focus has been.

Lynne Funk (31:20):

What about pricing in general? Pricing is a big issue in this market as most folks know. How has technology improved pricing?

Matthew Schrager (31:30):

I could take a first stab at this. Again, it's definitely a core part of our business. I mean, look, if you want to come up with a good price for a New York City bond, probably a lot of people can do that, right? But as William I think mentioned there are what, 50,000 or more issuers in the market, many most of which are nowhere near as liquid as a New York City. And it's really difficult for a person with limited time. Again, put yourself in the seat of a typical sell side trader. There are this year, year to date, I think we at td, we've seen north of 40,000 bid wanteds a day every day, sometimes higher, sometimes on average per day. If you're a person, if you're a typical sell side trader, you are simply not going to be able to come up with a price for anywhere near even a small proportion of all of those. And so technology is not in, pricing is not a nice to have for that kind of a situation. It's a must have.

(32:46):

Now, how does this end up working? We could do a dissertation on this. I won't bore everybody with the details, but certainly we apply and it's not just us or other firms like us that apply techniques that would fall under the ai slash ml slash ml is very important umbrella. But really to demystify it a little bit, it's not, some of the specifics are a little bit complicated, but the gist of it is really automating what people do and what does a person do if they're faced with a credit that they've never seen before? Well, they're probably going to look and okay, this particular CUSIP hasn't traded in three months, so that's not that useful. Has anything like it traded recently? And the answer to that question is almost always, yes, there's 40, 50,000 trades a day and if you look back over a long enough period of time, there's a lot of data to learn from.

(33:37):

And what computers and machine learning are really good at doing is synthesizing all of those millions and millions of data points to produce a reasonable result for basically any type of a bomb that you could ask for. Which again, and I would go so far as arguing, the only reason that you are able to have 40,000 RQs that number's way up, by the way, over the last five or six years, the only reason you're able to have that many RQs is because this kind of technology exists that allows the sell side to actually provide good liquidity to that many inquiries. So anyway, I'll pause there, but hopefully that's a good starting point.

Tyler Traudt (34:14):

Can I hijack your question a little bit? You

Matthew Schrager (34:16):

Can go for it. Everybody loves a good

Tyler Traudt (34:18):

Hijack. I'm going to get pretty creative in how I'm going to do that. Let me take it away and lemme talk about pricing, but I'm going to take that, I'm going to turn it into financial outcomes for organizations. How about that? We know if you are a treasurer inside of a airport, higher education institution, city, county transit system, whatever it might be, you have a team of people you're recruiting. It's a challenging time right now to recruit and retain your people. There are more and more modern cloud technology solutions that are going to be coming to your organization or organizations like yours. They're going to use it. They're going to go to a different organization, they're going to talk to a peer in the space, they're going to experience working at this organization. We got to use this cloud-based solution to do our job and automate a lot of the work that we were otherwise doing an Excel spreadsheet with paper files, the people are going to want those jobs.

(35:12):

They're going to want to work in those organizations. And everybody here has heard that we've heard it. I feel so blessed that we work inside of an organization that invests in its people and its tools to allow me to spend less time with my spreadsheets, pulling down formatting, sending information to my boss in more time, working on what really, really matters. So I think if you gave your team better tools, if you invested as a treasury team, your organization, your people are going to have more success. They're going to spend less time doing operational work, they're going to spend more time doing strategic work. And if I were the manager and I were setting the KPIs for my team, I would say, let's optimize risk. Let's lower costs, let's improve financial disclosure. Let's improve efficiencies. Let's build better systems internally so we can withstand turnover.

(35:59):

These would be the metrics that I would be using to rate my team's performance. But if they're stuck all day long doing it as they have had to do it for a really long time, number one, they're going to get bored and they're going to be frustrated and they're going to quit. They're going to go find a different job most likely, or they won't even come and work for you if you're not using those tools. But if you do give them that time back, they'll stop working on the data and they'll start working with the data. You'll have more success, they'll have more success, and your organization is going to have more success. And I can't tell you exactly how I get back to you like perfectly improve pricing, but there's no way it goes the other direction if the organization's materially improved. So the question now is if we gave all of our teams more time to work on more important things, would they do a better job? I believe absolutely. Yes. That's the truth. And so it should show up ultimately in the financial performance of these organization, which includes pricing.

William Kim (36:53):

That was good.

Jennifer Fredericks (36:55):

Well done.

William Kim (36:57):

I think for us, when we look at pricing, we really look at it as how when you're putting together a primary scale, when you're looking at a transaction, how can you instantly pull down all the secondary trading for your credit? So not just on an issuer specific level, but I just want to look at my water and sewer credit and look at all the secondary trades right away on top of that, issuers and investors really thinking credit spread. So we offer everything on a credit spread basis, but the main addition or value add that we have is we also have this option pricing, right? So secondary trading, a bond was issued with a 10 year park call five years ago. Now it's a five-year park call. So you can't really use that credit spread to inform your 10 year pricing. So what we do is we strip out the option value and then re-add it back in so that you can look at what kind of a synthetic 10 year call would be so you get better comps.

(37:51):

And when you're looking at the primary pricing, if you're an issuer and you're going to your banker and they give you the 5% scale with your spreads, but in your budget it would be better to have a 4% coupon for less interest. How do you determine what the correct or most attractive relative credit spread is? For a 4%? You need option pricing for that. And I think that was traditionally a realm for PhDs or advanced quants in the market, but it's something we provide for all our issuer clients as well as bank clients, pricing deals.

Jennifer Fredericks (38:26):

And the same, I mean, salt's got a scale view or a scale, right? Or a secondary ticker. And it just allows the whole conversation to happen, happen at your desk, and you don't need an extra quant group of whoever's going to run all of that. You have it right in front of you. You can run it 15 different ways and really have just access to more data faster.

Lynne Funk (38:50):

It's kind of a good segue because we've all sort of touched on this, but to move to the connection between the buy side and issuers, I would say, how have this evolution of more data, of better data, it helped the buy side with credit analysts, credit analysis, all of these things, how has it tied the issuer to the end purchaser in a more streamlined way?

Jennifer Fredericks (39:20):

I think it's just really given more levels. It's given real levels. It's given more transparency to what is really out there, what's happening, what's available. And also right now, predictive pricing is really kind of a, it's there, but right now it's kind of like a sanity check. Am I in the right ballpark? Am I in the right place soon? It won't be that. And it'll give you even more transparency into what's really happening and where things should be.

Lynne Funk (39:52):

Well, I kind of want to talk about next, just more broadly, how do you see technology aiding growth of the industry, the municipal industry? We can touch on headcount in a couple of minutes. It's not necessarily that, but where does this help with growth in Muni in the municipal industry?

Jennifer Fredericks (40:15):

I think just, I teach on the side. I'm an adjunct professor, and I think when AI first came out, machine learning and everybody was terrified that we're going to spend all of our time now making sure people are not using this to just create

(40:31):

However we're going to create it. But what we've found is that it allows for better discussion. It allows, it makes you teach people to ask better questions. And I think in the same way time that you're able to eliminate something that was a time suck, you're able to have growth because you've got more time to make phone calls, more time to reach out to clients, more time to double down on what they're doing, what they have been doing, where they could be going. And I think anytime that you create time where there was no time, you create growth if you use it effectively.

Matthew Schrager (41:11):

Sure. Because we're talking about the buy side here. I think a really interesting case where these topics come together, I'm not sure if you're going to ask about this later, but is the growth of the SMA product? Yes. I think it's a really good example of how technology leads to exactly what you said, which is growth of the asset class. So if you think about what an SMA offering is, it's not a mutual fund or you've got a giant pool of money that's managed on behalf of all the investors. It is tens of thousands or across the whole industry, hundreds of thousands of individual accounts, each of which has some sort of a bond ladder, basically tens or even sometimes hundreds of individual CUSIPs and each of those tens of thousands of accounts.

(42:14):

And you start thinking about that, it becomes very obvious that it is impossible to run that business without significant amounts of automation and technology. And if you go back in time 10, 15 years ago, that asset class was much smaller, right? It's grown astronomically over the last couple of 10, 15 years. And I think you get into this interesting sort of symbiotic relationship here actually between the effects of technology on the buy side and the sell side. Because again, if you think about that type of an offering, if the cost of trading smaller pieces of bonds is very large, which used to be the case, well, that means you can't really have very many accounts that have lots of smaller pieces in them because it's too inefficient, right?

(43:12):

As I mentioned earlier, transaction costs on the 95% of trades in the market that are for smaller sizes have dramatically decreased. You've seen these types of accounts that require that kind of liquidity explode. And not only because they've attracted assets in the way that they operated 10 years ago, but also because of things like they can now, because liquidity on the smaller sizes is better, they can reduce account minimums. And that literally brings in new investors who could never have accessed the asset class before. It used to require a million dollar minimum investment to get into an SMA. Nowadays, there are people that do it for a hundred grand. That's a great story. That really is the asset class growing and it would not be possible without the efficiency that this kind of technology brings.

William Kim (44:11):

I think technology is the great equalizer, right? It allows the best ideas and kind of best practices to impact the entire market. And it's kind of the core of conferences like this, right? People come here, the best ideas get disseminated, and that's put into practice by all the members here. And so technology gives that scale to every person, not just the 90 or plus esteemed issuers in attendance today, but to the entire market. And so that builds up the expertise level and kind of the tools that are available to everyone instantaneously as opposed to the slow process of training up people. And I know you'll talk about this later about headcount, but it does fill in the gaps in terms of workforce and churn at the issuer level, at the bank level. You're seeing substantial historical levels of change in all sectors of the market there.

Tyler Traudt (45:11):

If we're talking about how technology is going to change the municipal market, make it stronger. I'll just remind, and everybody here knows this, but when I was working in the industry, I always took this for granted. The people in this room are helping the treasury function inside of the organizations that we as a country need to deliver the most foundational and important services to every single person, my family, my kids, your families. Everybody needs water, everyone needs safety, everyone needs care, healthcare, senior care, all those different types of services. And ultimately, the organizations that we're trying to serve here collectively, the organizations that we work for, their customers are buying those things that they need. And so if we can improve outcomes, just get us all the way to, if we can improve outcomes, we're going to deliver more or better services that are really critically important to the quality of life of every single person in this room.

(46:09):

And you all know this, you walk down the street and you see this and you see that and you're like, oh, public finance absolutely had an impact there. It's just really, really true and it's really important. And so I think very simplistically, if we can provide better access to information, if we can reduce the amount of manual work, teams collectively will optimize for better outcomes and better outcomes will get passed on to the communities. And they won't know it. They won't know that we were in this room talking about it. They'll probably still complain. But these folks here can absolutely feel good that the work they've done is really meaningful. And that's important. They'll definitely still complain.

Lynne Funk (46:54):

So essentially, I think maybe if we could maybe ask if there's any questions in the audience. No.

Tyler Traudt (47:04):

Does anybody wants us to read the FDTA? We could still do that.

Lynne Funk (47:07):

We have a whole panel

Tyler Traudt (47:08):

On that. We can miss the cocktail reception and just read the FDTA.

Lynne Funk (47:11):

We have a whole panel on that on Wednesday. I'm moderating it. You should come. I won't read it, I promise. But let's just find, let's maybe finish off here with the headcount question. You all touched on it in a certain way, but I think if you would sum it up, that is sometimes the fear and from various aspects from the bankers to traders, like, oh, is electronic trading going to take over my, there's no need for that person anymore. What do you say to those folks?

Jennifer Fredericks (47:41):

Muni finance has always been a collective beautiful sector of art and science. We're giving you the science. If we can get you 90 yards down the field, you still need that 10% magic of whatever your approach, your vibe, your decades of experience, your new fresh eyes, whatever, that it's still needed. I think it's the beauty and speed of access to more, better, faster, deeper, broader, and more transparent.

Dan Silva (48:17):

I think this is interesting that we still talk about is technology going to take my job when it's a question? Technology has always evolved and jobs have always existed. And we don't have 25% unemployment right now, and we have a hell of a lot more technology today than we did 50 or a hundred years ago. And I think that's also going to be the case going forward. Will jobs be different? Probably, but they probably will still be jobs. So I don't think the smart, talented people in the room need to be worried about whether or not the jobs are going to exist in the future. I just think they're going to be maybe a little bit different and hopefully more rewarding and less monotonous. I think that has been the trend.

Tyler Traudt (48:58):

Just speaking, just speaking for the borrowers, for the treasurers, walk into a treasury office and show me somebody that's not overwhelmed, they don't exist. There's so much work to be done. They need automation. They need something to close the gap. And so certainly for the folks that we're trying to serve, I think we're putting people out of work. I think there's too much work and not enough people.

Matthew Schrager (49:25):

The nature of the work will change, but the jobs will still exist.

Lynne Funk (49:31):

Alright, well thank you so much to our panel. It was wonderful. I'm sure this discussion will continue as it evolves. Thank you. Thanks Lynne.

Tyler Traudt (49:39):

We're also super biased.