#8 - Abhishek Lodha: AG Analytics

[Music] You have entered the Muni Matrix. Please welcome your hosts, Matthew Gerstinfeld and Michael Lieberman, co-founders of Muni Change. Welcome to the eighth episode of the Muni Matrix. Today we are joined by Abhishek Loda, Director of Strategy and Innovation at EG Analytics, a new fintech subsidiary of a Shared Guarantee. But before we dive in, first a quick disclaimer. The Muni Matrix is for informational purposes only. Any opinions expressed on this episode should not be relied upon for investment purposes. Your use of the information is at your sole risk. Hello Abhishek. Thank you for joining us today. How's everything going on the West Coast? It's good. We did not get as much rain as we expected here in downtown Los Angeles. I'm feeling better about that. But yeah, there was a little bit of shakiness with the earthquake. But I'll go to my side. Thank you for asking. Wonderful. Well, I'm glad you're okay. And I hope nothing too bad happened by you. So, figured we just get a quick introduction on your side. I know you've been in the market for quite some time now. Very much geared towards technology and actually witnessing positive change throughout the market. Can you walk our audience through your background a bit? Sure. So, hi, I'm Abhishek Loda as some of you may know me. I'm the Director of Strategy and Innovation at a new subsidiary launched by Shared Guarantee called Aegean Analytics. But a lot of my background has been in in munis. Like all of us, right? I don't think I ever grew up thinking that I'd be in munis. But one of my first jobs out of grad school was being a research analyst for an advisory shop in New York. I quickly realized there are a lot of opportunities in technology and somehow pivoted and got between this intersection of muni credit and technology and found myself being a product strategist and product manager for a long time. And now, I'm assured a lot of this focus has turned into a new software company that we're launching to really help empower research analysts and build better products for them. Wonderful. Now, I know you've covered different aspects of the market from a technology perspective. But what exactly inspired you to focus on the intersection of technology and municipal bond credits? That's a great question going deep directly Matt. I think it comes down to my first job as the research analyst and my background in engineering. So I came to the US, found finance very interesting, but came from a mechanical engineering background. And I thought this focus was around building better systems, right? That was always part of that psyche. And when I started as a research analyst at this advisory shop, one of the first problems I came across was, why am I spending so much time collecting data? There was these moments where you wonder you're spending more time collecting data than actually analyzing it. And that really drove this side, like, journey towards, like, making sure we can build better systems and technologies to help alleviate that pain and empower the analyst to be able to cover five or 10,000 credits by, by myself, really. And soon that turned into product management and building the right tools. So municred analysis is one of those last areas I feel where technology is not serving the market as much as a children. There's a lot of room for improvement. So from your understanding, right, various firms are adopting newer forms of technology. And from what we've been hearing, they have to do a lot more with a lot less. So overall, what is the timeline looking for general adoption throughout the market for this particular aspect in which you're focusing on? I think it's an ongoing thing, right? If you think about credit analysis, it's just like how the market is fragmented. So is credit analysis, right? There are different pockets of analysis that happen at different levels across the entire market. The adoption is coming. I think one of the biggest macro level drivers that I've seen is the discussion around technology has become more and more ingrained in our market, right? We've, you know, we've been on panels, we've been at conferences where there are more munitech firms like yourself as well, right? Who are part of part of the condensation who are coming to the table providing better solutions and workflow tools. So there's been a huge cultural shift. I think we are definitely not completely there yet. We will continue improving. That's really a big part of what's driving adoption. The second is easier engineering and programming in general, right? I keep, I keep joking with some of the interns and analysts in that I come across as, you know, start learning Python because it's easy. The cost of data and cloud has gone down tremendously. There's a bigger level playing field and you should be able to build better tools that help you in your job, much faster. And beyond that, I think the other macro economic factor that we all have seen recently is, you know, with algo trading electronic executions separately managed accounts and the SMA business scaling up and and denominations and account minimums reducing. Significantly credit analysis has to pick up, you know, by those standards and there's a lot of room for adoption out there as well. So I want to start talking a little bit more about like historically speaking, how have munisional municipal bonds traditionally been evaluated in terms of credit risk and then what do you view as some of the challenges associated with that approach? That's that's the big one, right? Credit analysis, as you all know, plays a really large function in any asset management team, right? You'd have a portfolio manager, a trader, but you'd have two or maybe three analysts and sometimes there's a bigger team at larger asset managers because the focus has to be by sector by state. So, you know, if you're a candidate, there's so many nuances to that, right? So fundamentally, as an analyst, your focus is number one is issuer or the obligor right focusing on the financial health, the credit health of the entity itself, you focus on the legal structure of the issuance, understanding the covenants is absolutely critical. So, focus on the service area and its demographics, so you start looking at census data or BLS pure of labor statistics data, understanding whether that community is is growing or deteriorating and then you focus on other aspects of operational performance relative value, non quantitative metrics, news. So, I think like treading that need, as I'm speaking right, you can see there's so many different sources of information that you have to deal with and that really causes what we call swiveling, right? You're jumping from one source to another, just to get the information that you need before you can actually come to an inference or get the time to make a decision. So, broad strokes, I'd say the challenges for credit analysis has been threefolds the way I try to see it every time is number one is data, right? Information generation is key. There are plenty of sources of data today that we know of, right? And one of the one of the things I'm like a broken record, then I say is data is available, that's not the issue. It isn't easily accessible today, right? And that's what we want to solve for it's either stuck in PDF documents or it's in different sources that I mentioned and it requires analysts jumping back and forth between all these sources. So, can we create a better way to centralize this information and create this universal knowledge graph for an analyst? So, that's really the first bucket. The second is analytics, right? So, not all credits are the same. We have different nuances by sector, by state, accounting rules are different. But then there's still overlap in terms of standardizing this analysis. As an analyst, you follow in a muni market, you are probably following multiple folds, sometimes 10 times the number of credits corporate analysts does, right? Just by the function of the number of issues we have in our market, being able to standardize analysis and having those tools to provide that scalable solution is lacking today and there's huge value out there. And the number three category, which is generally overlooked, right, is workflow. So, we all have our own processes. And this is something that you as muni chain focus on a lot as well, right? It's not just about information aggregation, but it's about helping efficient decision making, right? And every asset management firm that I've spoken to has their own decision tree and tree-arging process to go through thousands of credits. And it's in a lot of cases, it's a struggle, right? Because you're doing a lot of these manual tasks outside of a particular system and that leads to a lot of pain. And we all know right at the end of the day munis in as an asset class in most investment firms does not get the same technological attention that other asset classes are given just, you know, either based on the risk return profile of the asset class or the fragmented nature. and thousands of credits, but those are exactly the reasons I believe the technology can play a huge role in making this more efficient. Happy, you brought up the organizational aspects, you know, we've been speaking with all types of firms and what we hear is management is very much focused on not only generating efficiency, but being able to make informed decisions with a lower amount of time involved. Speaking of making solid decisions, I wanted to bring up not only the FDA, but the role of artificial intelligence in the FDA. How does this role in community credit and what are you thinking about in the next couple of years as this gets solved? Yeah, it's both the hot topics of the industry today, right? The Financial Data Transparency Act and what role does AI play in our market? I think there are definitely complimentary, you know, overlaps that we're seeing between the two. How FDA plays out in terms of outcomes is still to be seen as an analyst and a technologist. I think the big focus I've had is digitization of data, right? There's a lot of data stuck in PDFs today. How can we digitize and structure that in an efficient way? And as FDA kind of plays out, a big part of that is, we all know it's not a one size fits all as a solution, right? The timelines are extremely challenging at this point in terms of implementation, but there are ways that technology can help and, you know, data scraping, as you said, so, you know, data scraping is a big part of automation that can help in this market. A lot of folks have seen challenges where because accounting rules are different between different states or by sector, as I mentioned earlier, data scraping technology, right, where you basically take, you know, for lack of better word of Python library, which scrapes a PDF document and creates financial data out of it is not that straightforward, right? And I just can't implement it as is there has to be a little bit of more investment out there and we've seen some firms starting to do that a lot more as a g analytics, it's obviously a big avenue for our interest as well on how FDA plays out in terms of tax on a means and standardization, because once that piece of standardization, like what are the rules of accounting are in place scraping becomes a much easier exercise. The type plays a role for me is is is almost second to your right now the way I think about it right the first one is document scraping in general. We all heard the word OCR or optical character recognition right if you think about PDFs in our space, there are text base PDFs what we call true PDFs right which you can easily control find in today and there are image based PDFs. OCR technologies in general have gotten better to be able to scrape these image based PDFs where a lot of more effort goes into day, than text based PDFs to be able to scrape that data. The second piece once you scrape that data is actually normalizing yet right so you have two issuers. One issuer calls it cash the other calls it cash and investments the third may have two different line items for cash and investments. How do you normalize that right AI can clear a big roll out there but there are other ways to get there as well is like around building better logic in your system to normalize these data sets so. I think in a nutshell AI plays a big role but there are other technologies which it's going to complement to be able to provide a more wholesome solution to structure this data. And the jury still out on what that final outcome with FDTA looks like i've been on one of the working panels with NFM made discussing that topic and it's a bit concerned across the market right not just from an issuer but also from an investor standpoint where you don't get the skinny down version which is it checks the box but then is it really usable and valuable to to an analyst. Do these emerging technologies play a factor in firms who are skeptical of either adopting it or exploring the potential of OCR or artificial intelligence methodologies. I think I think it comes down to this I don't know who I was talking to but you know yeah is a great example right AI is like the ending of sopranos right everyone talks about it everyone's excited but very few people get it. And that's what happens with the new technologies of such right there's this this lack of understanding and education on how they can like fit into your workflow and I think that's that's something that it's getting better as I said like the culture shift has happened there more and more people who want to on board better technology and figure it out especially given that. The market dynamics have changed right I mean let's talk about SMA again your your fees are being squeezed significantly so you need to find better ways to to your point Matt do more with less so there is there's apprehension just due to lack of understanding but I think it's it's getting better in our market but I'd love to I'd love to definitely see more of that happening. The market's definitely improving and I kind of want to stay a little bit on that topic of challenges i'm in your opinion what do you view are the key challenges that municipal practitioners are facing and where do you see technology helping them. Yeah so I'll stick to that same like data analytics and workflow paradigm right I think one of the key challenges as I said is bringing all this data together and and a big piece of that is that data generation piece so we need to we need to start unpacking these PDFs a lot better than we can right and and OCR or data scraping can work into ways the easier the low hanging fruit right now is the financial data piece right bringing that in the challenge out there lies on the normalization. But if we take a step back in terms of credit research there's a lot more other information in PDFs today which is not financial data right so maybe quality of factors or information's you know coming from the management discussion analysis piece or the notes of the financials or understanding news and litigation so there's this whole avenue of unstructured data where there's huge benefits for technology and again easier said than done but a huge avenue just given the the sheer number of documents that an analyst deals with on the corporate side we all see you know gen A I based or or just general machine learning or NLP natural language processing based companies scraping through all of these PDF documents and that hasn't really unpacked in our market yet but there's there's a big avenue out there. And beyond I think beyond the fundamental credit analysis research analysts in our market have to do so much more than you know what's typically part of their you know job right so great example is climate analysis. You know that's become a huge part of the discussion now where you know we're part of the heartic wake here in California extreme heat weather is happening storm surge and that's become more and more in green part of their analysis and and the challenge comes with that is I'm a research analyst but now I'm supposed to be a climate analyst as well right so very similar to how it used to be with pensions back in 2012 2013 right so these added lenses of analysis that keep getting added on which are just adding to the burden of of pain points and continues to increase that swiveling where again technology bringing in all this information in one singular system and being able to standardize or or systemize your analysis has huge benefits. So one thing I continue to ask myself and some of the other folks we're speaking with is if there are specific standards and there are different methodologies to adhere to these standards wouldn't in theory all these firms are operating on their own version of what they think is correct how do you see this playing out down the line. I think it's a great question I don't see that that that changing a lot in terms of you know every every firm is going to have its own way of analyzing and you know managing workflows right I think we're we're definitely coming to terms on certain things right so when we talk about like fundamental credit analysis as an example as well there's definitely agreement on what are the key credit metrics people look at what are the key demographic metrics people look at this there's almost like a standardization to a certain degree that's happening but beyond that there's always going to be a flavor of customization you know bespoke analysis that every firm is going to do so you know I think it's it's interesting like this is the conversation you always have right flexibility versus scalability and how do you find the right balance between the two. From a technological perspective as a g analytics a lot of our focus has been to build the right platforms which are modular in such a way that people can bring in their own analysis but then can in a no code low code way implement them efficiently within the system right and that's the thesis where we're operating under so I think long story short I know I people winded answer out their mat but there's going to be some overlap but I think to a certain degree there is going to be this bespoke nature of analysis that every firm is going to continue doing I imagine you have seen firsthand specific entities who are still manually calming through thousands of pages of documents and trying to understand it versus other firms who are now leveraging modern solutions to be more efficient. So generally speaking, tell me, how can practitioners actually prepare themselves for the shift of technological solutions that are underway? How does it play a factor in hiring great talent and growing your firm in the long run when now everyone is on a race to adopt these technologies? That's I think that's the most important question we as an industry can ask ourselves right I think technology should not be thought of as as solutions in like looking for problems, but rather we as an entity need to focus on how can we do things better and faster right and I mean you must have heard this quote which is repeated in many like you know tech conversations at least that I've been part of is every company is a technology company now right I think that's the most important thing everyone needs to focus on first right whatever our byproduct or outcome maybe we're a tech company no matter what right we're in a world of software today so first think you know like let's have a mind shift or change that mindset towards tech is not a back office operation anymore right it's a core business enabler now based on that you start thinking about making tech as part of your strategy right if I want to get from you know X to Y I need to keep having that conversation how can I set the processes how can I have the human resources and how can I have the technological resources to achieve that better faster and at scale so it's you know I remember one of my other startups I've worked at and this VC comes in and he said this beautiful thing it's like technology is man machines and biases right humans and technology need to work together to be able to deliver a really good value proposition in any industry for that matter so I think that mind shift needs to happen a little more and this idea on technology is not a replace for but an enabler right so what it allows what technology and automation really allows you to do is redeploy human efforts towards more important business critical goals right let's let's take their time away from the mechanical things are the manual things that they're doing and that's really been part of how we try to think as an organization and a startup as well right if I can do things if if I'm a tech company and I need to produce products for other company I need to be tech enabled internally as well so that's the first piece right from a cultural standpoint the second is implementation right ideas of cheap execution is the name of the game and think of technology as an incremental solution it's not going to be an all or nothing game right you want to focus on you know building a small use case test the proof of concept and then build it out so that way you're not over investing upfront without knowing the outcomes and you're able to iterate on the processes as you build it and like we all hear what agile and software development a lot right I think thinking in agile ways to you know conceptualize and implement technology small and then scale it up is absolutely important because it takes away that overwhelming nature of what technology means and you know this this running joke I you know my wife asked me what the hell is a muni market she has little idea about what I do she's in the healthcare industry but I keep explaining our munis is like organic chemistry it's full of exceptions right so technologies are not going to solve everything immediately but if you can get there 80% of the way it helps you redeploy your time towards other important things so that's going to be my last thing and in terms of hiring talent I think you're on the right track I think that the skill set is changing to a certain degree new analyst in terms of mentees that I talk to I try to make them more you know I mean Python is the new excel right let's think of it that way or any other programming language you need to be able to deal with big data the way you dealt with excel 10 years ago and I'm still a big proponent of excel but it starts to have certain limiting factors with you know when it comes to big data and actually Microsoft just announced they're going to have Python running within excel today or yesterday so that's going to be fun but I think improving those skills set up bringing in technological and quantitative resources and you seeing that at some firms right like firms that are ahead on the technology game are having traders or credit analysts will hand in hand with quants or data scientists and building like a more consistent team which is both you know human effort and technical effort focused at the same time are there any activities that are still taking place in the market that just surprised you and you're just like why are we still doing it this way. I keep saying control left right it's like I still remember when I joined munis and I was at one of my first muni conferences and and someone said look at the technological innovation we've had we can now control find and PDFs and and that just blows my mind. I think you know I think we're still doing a lot of this manual labor and I know I'm a broken record when I say this but there are better and and ways to get there and I think more and more people are realizing it it's just helping them find the right tools so I would say the hand spreading numbers is something that definitely. You know gets my attention a lot and and I keep saying like I want to I want to be an analyst not a data Wrangler and and that's the focus I want to have as an analyst so. Yeah that's that's I think going to be the big pet peeve of mine. So I think that's the key to that is that it's been absolutely incredible to have you on today your insights your focus on watching the market evolve and playing a very active role in the future and evolution. View a g analytics take a look let us know what you think and Abishek we look forward to seeing new things roll out on your site thanks for joining us. I'm really excited here and Matthew and Michael thank you for the opportunity this was great appreciate it. Thanks for listening to this episode of the muni matrix by muni chain to be a guest or recommend a topic please contact the muni matrix at muni chain dot com. Stay tuned for another episode.