.jpg)
The CU2.0 Podcast
This podcast explores contemporary, critical thinking and issues impacting the nation's credit unions. What do they need to be doing to not just survive but prosper?
The CU2.0 Podcast
CU 2.0 Podcast Episode 401 One Washington Financial and ALM Company Delfi
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Asset liability management - ALM as it’s known - can be fun, good for a credit union’s balance sheet and, dare we say it, sexy.
On the show is Daniel Ahn, CEO of Delfi, a company that - and I quote from the website - “combines deep public and private sector experience with cutting-edge AI expertise—bringing best-in-class technology to banks, credit unions, and other financial institutions.”
The website adds: “Transform your bank or credit union financial decision making and race to the top with our revolutionary platform.
Delfi delivers instant analytics, access to a deep inventory of solutions, and is powered by AI so you can pick the right solution for you.”
This an exciting new, powerful look at ALM in credit unions.
Also on the show is Scott Daukas, a principal at One Washington Financial, a CUSO wholly owned by Washington State Employees Credit Union and an investor in a new Delfi CUSO. But One Washington Financial is more than that with Delfi as you will hear in the show. Indeed it approached Delfi with the suggestion that it form a CUSO - and long conversations followed, not the least of which was explaining what a CUSO is!
One Washington Financial also invested in two other companies that have featured in recent CU 2.0 podcasts - Remynt and Starlight. Why is One Washington Financial making these investments? Daukas explains.
Listen up.
Like what you are hearing? Find out how you can help sponsor this podcast here. Very affordable sponsorship packages are available. Email rjmcgarvey@gmail.com
And like this podcast on whatever service you use to stream it. That matters.
Find out more about CU2.0 and the digital transformation of credit unions here. It's a journey every credit union needs to take. Pronto
Welcome to the CU2.0 podcast.
SPEAKER_00Hi, and welcome to the CU2.0 podcast with big new ideas about credit unions and conversations about innovative technology with credit union and fintech leaders. This podcast is brought to you by Quillo, the real-time loan syndication network for credit unions, and by your host, longtime credit union and financial technology journalist Robert McGarvey. And now the CU 2.0 podcast with Robert McGarvey.
SPEAKER_02Asset liability management, ALM as it's known in the trade, can be fun. It can be good for a credit union balance sheet, and dare we say it, sexy. On the show is Daniel Ahn, CEO of Delphi, a company that, and I quote from their website, combines deep public and private sector experience with cutting-edge AI expertise, bringing best in class technology to banks, credit unions, and other financial institutions. The website ads transform your banker credit union financial decision making and race to the top with our revolutionary platform. Delphi delivers instant analytics, access to a deep inventory of solutions and is powered by AIT and pick the right solution for you. Close quote. Also on the show is Scott Dawkins, a principal at One Washington Financial, a QZO wholly owned by Washington State Employees Credit Union, and an investor in a new Delphi CUSO. But One Washington Financial is more than that with Delphi, as you will hear in the show. Indeed, Dawkins approached Delphi with the suggestion that it form a QZO. And long conversations followed, not the least of which was explaining what a CUSO is. One Washington Financial also has invested in two other companies that have featured in recent CU2.0 podcast, Remint and Starlight. Why is One Washington Financial making these investments? Docus explains. Listen up. I'm turning into a one-person publicist for you.
SPEAKER_01And I appreciate it. You're a great hype man.
SPEAKER_02Well, what you don't know is that there's another one in the pipeline that I'm going to post in a month or two with Gwyneth Borden. Oh, awesome. I love Gwyneth. Whom I discovered looking at CU Insight, and I said, I got to talk to her. And then she said, she said that she was had gotten some seed money from you. And I said, Well, that's that's what distinguishes what you're doing from what similar entities are doing? I mean, why why do I like three of your companies?
SPEAKER_01I will say that, and I say this with humility, obviously, right? But but I I think that we have we since we have spent so much time in the space as practitioners and executives, but also we we obviously want financial returns from these, but we're more we're more focused on impact and on ones that are going to be truly make a difference. We're about to talk here, Daniel. I see on the line about like Delphi. And it's I I think we're just we're finding ones that we really think can make a difference. And that's that hits people, right? So I don't know. That's that's my easy answer for you. I hope I don't know if we're I hope we're not recording it so far. Oh, we are. We are.
SPEAKER_02Do you take bigger risks than some other entities too?
SPEAKER_01We are taking the risk of working with earlier stage fintechs, and so in that sense, yes, because many of the credit unions and the holding companies and the other types of funds in our space don't invest as early as we're investing. So on the same token, though, we're not we're not exclusively investing in early stage. That's really more of our focus. But yeah, I think we're taking some more risk in that regard. And I think that's okay. I think we're we you know, we've done a lot of work with our board on on risk appetite and and on kind of what again, back to impact, what impact do we want to have on the industry? And we really take a look at the full portfolio theory or portfolio management theory, right? Of you know, we're not gonna they're not all gonna work, but we are putting everything we're trying our best to put as much risk mitigate risk mitigations in place to hopefully drive success.
SPEAKER_02Well, one of my tantrums about with credit unions is and it doesn't go over well, is that they're horribly, horribly risk averse. A fintech, you start a fintech, you and I start a fintech. We say, okay, fine, 10% of the things we sign out are gonna go bust. Absolutely. Maybe even 15%. But we can we'll prices say we can live with that. Credit union would say, Whoa, 10% is gonna go bad. This is impossible. But that's I don't expect you to comment on that.
SPEAKER_01But that's no, no, but that that's I will comment on it. I think that's what and that is one of the challenges for why things don't always move as quickly and the why this area hasn't grown in the way that people there's like this disconnect between people wanting to be innovative, credit unions wanting to be innovative, wanting to help design solutions with smart technologists, and then and then the gap is that that they aren't willing to take the risk to do that. It it all works great when the first the first investment you make has a huge exit and that that can fund a lot of um you know future failures. It is a much tougher pill to swallow when the first one goes bad, and to be able to stand and justify, you know, on the portfolio, the grounds of the portfolio, that over time you're gonna make money because you have to have supreme belief in that, and and it's not always the case. So I mean I get it. This it's a it's a higher risk thing for credit unions to do, but I'm just fortunate that the board of the QSO that I work for and the parent credit union are just very focused on impact and on innovation and want us to be doing this type of work.
SPEAKER_02So uh Daniel, your turn to talk. And uh I'm looking at your your company, and first thing I'm looking at is who we are. I'm saying, my heavens, we got three PhDs, right? Boom, boom, boom. And not only they PhDs, no, no, no, Harvard and MIT. I mean, you it's hard to do better than that.
SPEAKER_03The Kremlin on the Charles, I suppose. Yeah, we do have a somewhat unhealthy concentration of PhDs uh on our team, but uh no, uh we we uh also have some great uh uh you know, like our head of ops uh comes from from Deloitte uh with a uh really handling that wealth. So yeah, uh no, really fortunate to have gathered um, I think a really stellar cast um of talent that as Scott was saying, also kind of is supremely confident in the vision as well.
SPEAKER_02Now, very briefly, what is your company selling?
SPEAKER_03Our company is selling the ability to not just analyze a balance sheet or a portfolio, but then help guide what to do about it. I really feel that if there's like one theme that drives Delphi, it's connecting insight to action. Yeah, I've worked at lots of places where they just do a lot of a lot of analysis and insight and insights. I've worked at broker dealers that just try to force action of any kind. But I there's I think the missing link has been how to connect industry-leading insight into your balance sheet that could be delivered by, you know, that normally is delivered by a big quant team, and how to connect that uh into an actual well-structured, well-thought-through action on what should I do with my balance sheet is really what uh is at the heart of Delphi.
SPEAKER_02Are there competitive products out there?
SPEAKER_03Yes and no. I would say there are certainly a lot of incumbent like ALM consultants and advisors. I mean, every bank and credit union obviously has to uh do at least some very basic ALM reporting and and stress testing. And so there's a whole cottage industry of consultants that that work on that space. Uh and then, of course, there are all these broker dealers and corporate credit unions and and stuff like that out there that are also trying to pitch transactions or M ⁇ A activity or whatnot. I think what again uniquely defines us is we're trying to draw a firm line connecting those two dots together. Like you don't really turn to your ALM consultant to help advise you on what you should do with your securities portfolio. Oh, maybe you do, but but it could take like weeks of work to work together with them. And you know, you can go to your broker dealer, and I'm sure they'll give you some basic ALM analysis on your portfolio, but nothing comprehensive and always with an eye uh toward kind of driving that transaction. But we have built a on-demand, near real-time platform that is that again allows you to credit units to analyze your balance sheet in minutes, uh, but then actively think through so what should I do next? What happens if I go ahead and do this MA transaction or do this derivative hedging strategy or reprice my deposit product this way, being able to see that like this is what happens to my uh expected profit, this is what happens to my risk exposure in minutes, we hope is a real game changer uh for these institutions.
SPEAKER_02Is your sense that most credit unions know to ask those kinds of questions that you're asking that you say your your tool will answer?
SPEAKER_03I think yes, if only to feel that they sort of need more kind of structure before they just willy-nilly go ahead and enter into some kind of transaction or some kind of like deal. Like uh, I think a lot of the times there these decisions are made in a somewhat ad hoc kind of dare say vibe-based fashion. But I leave but I think they know that that shouldn't be the case. But then what actually is the question to ask? Like, how should I think about this? How should I frame it in terms of uh what kind of KPIs I should look at and all that stuff? I think that is the missing piece that we're trying to fill here.
SPEAKER_01I would I would also just add from the credit union's perspective, as somebody who oversaw ALM for a long time at my credit union, my past credit union, is that there's information that exists in different systems that help give us answers, but they aren't always brought together in a way that can allow us to make the best op the most optimal decision across the full set of trade-offs that we have to make. And I think that's a huge, a huge thing that Delphi is bringing bringing to the table. We could go into a meeting and talk about interest rate risk and run some scenarios. And then how does that relate to our pricing committee and the and the meeting that we have for pricing and how we're thinking about pricing? And how does that relate to when we're talking about liquidity and which ones we want to sell and which ones we want to buy, those types of things. And so there, I think credit unions have a have a general sense of questions to ask. I think this tool and and the work that Delphi is doing and going to do is going to help them see all sorts of questions they've never thought to ask before and ways in which they will understand this information at a much deeper level. But mostly I think it's just going to help them see it all in one place and make a better overall decision.
SPEAKER_02Well, Scott, many fintechs tell me, without naming names, that credit unions still have a real information silo problem. That there'll be a bucket of pretty good information here, a bucket, there, a bucket, and the buckets don't communicate with each other. So you're fintech and you're supposed to make sense out of all this. It's tricky since the parts don't cohere, they don't snap together. Yeah. Does that make sense to you? Does it does it seem true? Or are these guys just kvetching?
SPEAKER_01No, that is true. I think it's true across a lot of the industry. I would say with respect to this particular topic, credit unions, since they're regulated and have to do this type of work every quarter, the granular, granular level data that's needed to drive ALM is all in one system. But it's usually it's, you know, it's like the you send a file and it's a huge file. But what happens is is that like the output, the the analysis and output is usually done in discrete parts, is how I would describe it. So you'll get a report that tells you if this scenario happens, these are the potential outcomes. Cool. That's important. But can you see that against the other 12 possible scenarios that might exist for uh for interest rates, for example? And that's not always the case. You can't always see these things stacked up against each other in a way that that gives you the best risk-reward trade-off information. So it's a little bit in this case, it's it's about who's doing the analysis and how that analysis gets brought together into a you know a nice unified package.
SPEAKER_02Daniel, what do you need from a credit union for you to be able to work with them?
SPEAKER_03As Scott alluded to, that the first part is indeed the balance sheet data. And thankfully, there is an existing process of gathering uh the balance sheet data from their cores or something like that uh to run uh the platform. After that, what we quote unquote need is ideas. Uh we need to hear from them what ideas they're actually contemplating. Again, it this is almost where the imagination's the limit here. Like, are they thinking about again rebalancing their securities portfolio? Are they thinking about originating new loans? Are they active in loan participations? Are they thinking about directive hedging strategies? Are they thinking about getting acquired? Like all, you know, any of these things actually can be analyzed and and uh vetted by our platform. So what we need from them is really for them to be ambitious and proactive enough to actually start putting these ideas into the platform and realizing that they can get robust assessments of the impact and the consequences of those potential strategies in near real time and internalize that into the decision-making processes that they have.
SPEAKER_02Now, what size credit unions do you do you see yourself working with? I mean, there are T tiny ones that uh some shoebox credit unions, as they used to be called, that literally have their books in a shoebox and their and their money in the shoebox. And then there are multi-billion dollar credit unions. What what size are you looking for?
SPEAKER_03Right. I mean, I'd like to think that credit unions of any size can see value from Delphi. Even the smallest credit unions at least have to do some basic regulatorily mandated stress tests and things like that. And, you know, uh like there's nothing to lose and and and and everything to gain to at least be able to run those things in minutes rather than waiting days to weeks to do this at a much cheaper price point. Uh but I would say that it is some larger credit unions call it 500 million plus that have complexities on their balance sheet that uh have a balance sheet that's complex enough that it it probably otherwise would need a larger team of quant analysts and and and trade structures and stuff to properly think through how to optimize that balance sheet, but the institution does not have the the resources to you know hire expensive PhD quants like myself, basically. So yeah, that's does that answer your question, Robert?
SPEAKER_02Yeah, and how how much does a credit union have to do when it's working with you? And how much do you do?
SPEAKER_03There's always like an initial onboarding sort of process where we kind of review their data. We have, I think, increasingly improved agentic ways of ingesting their data, but it still requires a human eye, uh an experienced eye to look through um the data and make sure nothing is crazy or odd. But yeah, I mean we we can get them up and running within you know 24 hours or so. And then we spend a couple sessions kind of training them up on the platform. But if I say so myself, I think the feedback has generally been that it's a very well designed, very intuitive platform. So it there's not a really a steep learning curve on this at all. And then it's really just you get out what you put in in terms of again, like you uh are you using this to start properly evaluating strategies every time next time you go to your broker or consider doing some kind of transaction. Uh it's worth you know spending those extra couple of minutes to punching, punching it into the platform, running it, and seeing, all right, well, this has in you know increased my mean expected net interest income over the next 24 months by 5%, but it has also increased the volatility of my exposure by 3%, like like things like that is uh the the bandwidth requirement from for from our uh clients.
SPEAKER_02How many banks are clients of yours?
SPEAKER_03So we have about 20 banks that are on the platform uh with uh QSO now set up. We are welcoming our first cohort of credit union users with WCCU and MapCU, and actually also some of these uh Philene um Phi Lab participants at the very head of the queue.
SPEAKER_02How big is your biggest bank client?
SPEAKER_03Our largest is actually a $70 billion uh bank that's actually based in the Middle East, not in the US.
SPEAKER_02Um same same questions that that a US bank would have? I assume so.
SPEAKER_03There's some interchronocies. They you know obviously have Sharia compliant kind of loan structures, so we've had to build some custom stuff for them on that front. And uh they have arguably some FX exposure, so that's also on our roadmap to do, uh expanded to start looking at FX risk exposure as well. But uh their currency is largely pegged to the dollar, so it's not as not as complicated as as some currencies could be. Um but but we're racing for more clients that will have floating uh floating FX currency exposure that we're gonna have to model for them carefully.
SPEAKER_02What's the what's the fee structure?
SPEAKER_03So we actually have a interesting dual structure. We can allow for just a standard annual subscription, but we and at least on that front, we we don't insist upon like three-year or five-year lockup periods or things that we think the product speaks for itself. Uh but at least initially for our initial uh pilot users, we are also providing the sort of on-demand uh pricing. So it will just be like just based on how many runs they do and how complex the balance sheet is. We also search as charge uh based on that. Uh we actually just to give you a like a sense, uh, it costs maybe about $150 to run a full on you know 300,000 strong kind of balance balance sheet and run a full kind of multi-carlo simulations uh with thousands of simulations on that. That full run costs about $150 to $200 to run. And again, it generates in maybe like 15 minutes or so. So instead of again having to pay whatever $75,000, $100,000 per year um to get like four reports over the course of four quarters, uh, you can now sort of just generate as many as you want. You can generate multiple of these uh in a single afternoon, costing a couple hundred bucks apiece.
SPEAKER_02So so a credit union can get in touch with you, say, I want you to do what you just say. It's 150 bucks, right? And and that would be the extent of the obligation on their part.
SPEAKER_03Yes, I mean obligation. Uh there, they're they can, you know, again, just have access to the platform um and run as many as as they want. Uh uh, we know we really would encourage them to really use the what if strategic strategic decision tool and the AI kind of hedging recommendation tool to its fullest. But yes, I mean, all they really need to do is just send us the same data that they already send uh to their existing ALM provider and then uh yeah, spend uh a couple of sessions with us. Uh kind of walking them through how to use the app. If they don't even want to do that and want to have their hands completely held, then we have a partnership with a consulting firm called CFO Consulting Partners, and they provide additional kind of support of their kind of power users of the Delphi platform. So they really know how to generate exactly what the client sort of may want. But again, this is sort of done on demand. So uh they can manage uh the costs and uh de-risk uh this as much as they want.
SPEAKER_01Robert, so the one of the things I was gonna add to that is that you know all the credit unions and the and all of the bank clients that that they'll have, you know they have to they have to produce these regulatory required reports on a quarterly basis. And so the tool can do all of that, right? And so from that standpoint, like I would see credit unions determining that the platform works the way that they think it works and that they can produce their regulatory requirements and and and and all of that. And I think they can do that in a quicker and more cost-effective way. But the real value then is in those what-ifs and in the ability to stack those what-ifs against each other and make good trade-off decisions. And I think once credit unions get their hands on it and start to like you know, see the power of that, I think that's that's really where they're going to gravitate towards. And so the the model, the pricing model is designed truly to like get them in the platform and get them using it, because that's where they're gonna see like this is a different, this is a different way of thinking about approaching balance sheet strategy for than what they currently have today.
SPEAKER_02So so Daniel, the question I keep hearing from credit unions, not they're not asking me to answer it, but it's something they're mulling. Should we what should we do regarding our car automobile loan portfolio? Should we be more aggressive on this? Should we be retrenching this? What should we do? Is that the kind of question that you can help with?
SPEAKER_03Kind of yes and no. What uh they can do is basically run sort of simulations like, all right, so if I let's say I am more aggressive on this and I do decide to expand um uh even more aggressively into this market, but at you know, but I have a menu of different potential pricing points, but and those pricing points have different, you know, therefore likely levels of default, right? Uh they're able to then put each of those sort of simulations and like each of those projections, run simulations and see exactly what the consequence is and what and how best to maximize their their performance for um the given amount of risk, uh risk that they have. So it's it's really at the heart of of Delphi is the ability to kind of digitally replicate both their existing balance sheet and hypothetical sort of balance sheets, um, uh which maybe which would involve some kind of change, right, from their existing balance sheets. Uh and we're able to do that in 100% fidelity uh again in minutes. So it's a tool that I think takes a lot of mathematical drudgery out of simulating these things so that the CFO and and the ALCO and the leadership can really think through strategically, all right. So then what are the the trade-offs between going more aggressively or less aggressively in auto loans or CRE or whatever else? Whatever the challenges uh they may be uh strategic challenges they may be facing.
SPEAKER_02Do you have a cheat sheet for clients that shows them some possible what-if scenarios they could run?
SPEAKER_03Yes, probably the the most obvious kind of things that we actually already prepare for them. So uh first thing I should say is that um if any credit union uh connects with us, the first thing we do for their demo is we prepare a custom simulation for them based off of their call report data. We have the ability to ingest any institution's call report data. And actually, we recently added that feature so that any user can pull off anyone else's call report data as well. But, anyways, we can generate a approximate balance sheet off of their latest call report data and run a simulation of the balance sheet on that. So with the asterisk, that it is approximate data, so it's never going to be exactly the same as what you know would be a report based on their actual true granular item level data. We're then able to like show things like well, you know, if you decided to hedge your securities portfolio, this is what the AI hedging co-pilot recommendation system would have suggested as the swap you should go out and ask your swap dealer to do if that's what you wanted to hedge. We have a partnership right now with a loan brokerage called CCT, where we are computing the impact of entering into any one of the loan participations and uh estimating the goodness of fits of those loan opportunities on the balance sheet of the credit union that's explored. So those are two kind of examples of ready-to-go kind of actions that they may consider doing. But obviously, you know, there isn't a one-size-fits-all cookie cutter to everything. Every balance sheet is unique, every strategic landscape faced by a credit union is unique. So really the tool is designed so that whatever strategy they are considering, they're easily able to upload that and evaluate the risk versus return trade-offs.
SPEAKER_02If a credit union analyzes a competitor's call report data, what kind of information would they glean from that?
SPEAKER_03Perhaps the most interesting thing that we actually generate from call reports is estimates of things like deposit betas. And we do it in this kind of forward way. So as you probably know, in call report data, they report what's on their balance sheet, at least aggregates of what's on their balance sheet, but also kind of backward performance, right? Like what was their net interest income or things like that in the previous quarter. We don't think that's a terribly useful guide for what's going to happen in the future. So what we do is we take what's is on their what they say is on their balance sheet, and we project out what's going to happen uh to their balance sheet uh going forward under different interest rate environments and stuff. So we in our peer group analysis feature on the platform, uh clients are able to put in whatever um gears uh they want to set up in their peer group and then evaluate where their performance, where their interest rate risk exposure, where their deposit beta, where their margins stack up going forward relative to where their peers are projected to do going forward?
SPEAKER_02Well, and this will seem a little off track, but I've always seen FICO score. The FICO score is a really interesting analysis of my past credit behavior, but I see it as borderline useless on my future credit behavior. Does that same question come up regarding the data that you're analyzing?
SPEAKER_03100%. I think uh it is very dangerous to just look at past behavior, to look at backward-looking sort of analysis, uh, to sort of give guidance on what you expect uh future behavior, whether of your depositors, whether of uh your the performance of the balance sheet as a whole, whether what the market uh might do sort of going forward. I mean, that's why and Dufy we do do the standard regulatorily sort of mandated peril rate shocks, you know, plus 400 basis points to minus 400 basis points. Yes, we do do all that. But then what we actually do is we take what the forward curve is currently pricing in to be the most likely future path of interest rates, and we also take the observed market options prices because in built into those options prices, you can back out what the market is pricing in as the probability, the likelihood that interest rates will deviate away from that forward curve. This is what's known as the option-implied probability distribution for interest rates. So we pull that from options prices and the forward prices every day, and we project out a simulation across the entire cone of likely possibilities according to the market of where interest rates will go, then generate basically a value at risk style bell curve distribution of the performance of the balance sheet. And then you can overlay those plus 400, minus 400 things on top of it, and you'll quickly see that these plus 400, minus 400 are like nigh nigh meaningless when trying to evaluate what actually is the most likely path for uh what your balance sheet uh might do kind of going going forward.
SPEAKER_02Now, what convinced you that you wanted to have a QZO? And to me, uh when I first heard about QZOs, this was some years ago, I thought this was the most esoteric, silly thing I'd ever heard of. But I've I've become I'm a big believer in Q Zos for credit unions. But it took it took a while for me to come across to that. So when you first heard about this, did you say, oh, I gotta do this now, or did you say, What? What the heck is this?
SPEAKER_03I mean, all credit most go to Scott here for being a great mentor, um, a great friend, and a great guide to how to, you know, what is a CUSO, how to navigate sort of the uh the credit union kind of landscape, uh, uh, and really how a QSO uh can be helpful as a vehicle uh to you know both attract strategic investment, but also to really send a signal on how um we uh want to uh make a mark and and uh change credit union processes, uh uh uh improve uh credit union processes uh in the space. Uh and uh it was yeah, so kind of when I first heard this, I was like, that's that's interesting. I never heard of this this kind of thing before, but uh once kind of Scott walked me through it and I also uh did more kind of reading and research, and I've spoken to other uh founders who've also formed AQSOs. Uh I thought this was uh a great, um, a great way for, you know, I I know you were talking with Scott about uh kind of risk aversion in among credit unions. Uh this has been an adaptive way uh for credit unions uh to uh get exposure to kind of technological innovation um occurring in the sector.
SPEAKER_02I heard you're right. And this is this I find this fascinating. Scott is the person who sold you. You didn't sell him on investing.
SPEAKER_01I mean, it's I I love this. This is this is great. Hey. So honestly, we were brought we were brought together by a wonderful guy named Tom Penton, who who does a lot of work with with fintechs. And it and you know, I I'll admit being a QSO is not the right approach or strategy for every for every fintech. And so it was really actually this exploratory conversation and allowing me to explain what I think are the benefits and and maybe even help Daniel think through and talk about whether or not it was right for for Delphi. But I had one added advantage, which is I well, maybe I consider it's two added advantages. I saw Delphi, I saw Daniel pitch at the Finovate New York in 2024 and voted for him and his company to win best in show from the audience there. So I was already familiar with this company and this product. And my and the second piece to that is that most people who are sitting in a fintech pitch event don't have a strong ALM and balance sheet management background, but I I actually do because it's of the work that I used to do prior. So I understand this problem quite well that they're solving, and like felt very compelled to to want to be a part of it if Daniel wanted us to be a part of it. And Daniel, it's probably fair to say it was not like, oh yeah, let's just do this. Uh that that conversation, even just to talk about what a QSO is and why that strategically might make sense, was a several month and several meeting conversation, ultimately transpiring in us partnering with MAPS Credit Union to make the first initial investment into Delphi QSO. And we are hopefully going to close the round in about 45 days, maybe even less than that, depends. We've got a lot of other interested parties that want to want to be a part, a part of the Delphi QSO ownership group.
SPEAKER_03This was again um a very much a learning journey um by by myself, but the more I I saw it and the more we I spoke with a real uh kindred spirit. Again, we we love uh leaders like Scott who are also familiar with the challenges of asset liability management, you know, going at Finovate Fall. There are a lot of other you know great startups out there that were using the Chat GPT or LLMs to automate this, you know, like like or or improve that, but no one was tackling this massive headache of a problem that is uh asset liability, you know, asset liability management and balance sheet optimization within the office of the CFO, right? So people like Scott who have that background and know the problem, and then people like like uh maps maps uh uh CU as well, who say they were looking for a solution like this for a long time. We take uh what they you know, we take what they say very uh to heart and really want to find the best vehicle to connect with those kind of thought leadership, uh thinking CFOs who are really trying to think through what should I, you know, what what tools do I need to better optimize my balance sheet.
SPEAKER_02You're making what I had thought was kind of tedious, boring stuff actually sound sexy and useful.
SPEAKER_03Um like a platform right now. I come from kind of the, you know, like my my background, I I mostly worked at at big you know, Wall Street banks and broker dealers and hedge funds. And there, those the role, the job of the CFO is essentially kind of parallel to the job of a portfolio manager um at a hedge fund, right? They're the ones who are trying to decide what assets to put on, not so much liabilities, but there's obviously a major liability question uh for asset managers as well. And that that is the that is the the sexiest job, right? Um so uh but I also felt um, I mean this is what kind of motivated me in part to to start Delphi was uh so you know, at these big broker dealers, we were working with corporate treasuries, but also with community banks and credit unions and again pitching them on things. Uh and there was a let's put it, there was an ace, there was an informational asymmetry, partially because as kind of Scott said, like the all the information they need about their balance sheet is sort of siloed and scattered across different places, but also because uh yeah, like you know, honestly, I think sometimes they were a little intimidated by um the very kind of math-heavy sort of like things that uh broker dealers would bring in. And I said, that's that's not right. Uh these uh institutions should not be gatekeep away from effectively using capital markets, effectively using derivatives, and effectively using uh uh you know all the tools that are available for big asset managers and big hedge funds. You know, these small institutions should also be able to use them just as effectively. The only reason why they weren't was because, you know, again, like uh uh it's only a hedge fund that otherwise could afford armies of PhD quants and traders and structures. Well, guess what? Now we are in a world where you don't have to hire those things. This can all be programmatically put into um uh uh into machine uh you know learning machines on the cloud and make this accessible for anyone and everyone.
SPEAKER_02You you got your PhD from Harvard in economics. In your class, where did the people who received a PhD go to get employment?
SPEAKER_03Uh of course, the uh what's the word standard track? I guess the most plain vanilla track is to become an academic. And actually, that was a path that I was considering very, very carefully, but decided instead uh to try and get some real world uh kind of experience. Uh quite a lot of us went to Wall Street and yeah, became, you know, went to banks or hedge funds, um, things like that. Uh still many of us there. And uh perhaps in my year, it was notable for uh quite a number have started joining tech firms, Amazon, Google, Meta, that kind of stuff. And uh there's actually, I don't know, people know this, but there's there's quite a lot of economists and and financial economists working at places like Amazon or DoorDash or Uber, and they're carefully running experiments, testing what uh the the impact of you know surge pricing. One of my friends is uh was the chief economist at Uber and was doing a lot of work on designing the right kind of search pricing that will most effectively meet uh supply and supply and demand uh in that kind of space. So yeah, lots of economics and a lot of data science uh that has to be done in that space, too. Yeah, the others go go to policy uh to government or to think tanks. I myself actually had a had a stint um in policy at the Fed and at the State Department and at the IMF, uh where it kind of uh went went to a bunch of different uh uh avenues they might many of your peers went into the working in the real world as opposed to academia or government for that matter, which I find interesting.
SPEAKER_02This this is interesting phenomenon where hiring.
SPEAKER_03Yeah, it makes more sense when you hear uh so Hal Varian is was a longtime kind of economist at Joab. He's a very famous uh professor at Berkeley, but also um uh was a uh economist, chief economist or something like that at Google. And he likes to say that economists and econometricians are the original machine learning experts and original uh data science experts before the term kind of became sexy. Because actually, uh yeah, like they're they're the same thing machine learning, data science, uh econometrics, statistics, they're all very close, have a very close relationship uh with each other. And so uh in fact, my my brother and co-founder at Delphi, who has his PhD from Harvard Business School, but specialized in in data science and machine learning, was working on on neural networks and uh uh stuff um before it kind of swept the world uh with uh uh the the onset of Chat GPT. So uh I think there's been a huge demand for people with uh these uh ability to manipulate and draw insights from ever larger and more complex data sets and translating that again into action and optimization of some kind.
SPEAKER_02This is a whole new way of looking at what some people used to call a dismal science. Before we go, think hard about how you can help support this podcast so we can do more interviews with more thoughtful leaders in the credit union world. What we're trying to figure out here in these podcasts is what's next for credit unions. What can they do to really, really, really make a difference in the financial scene? Can't all be mega banks, can it? It's my hope it won't all be mega banks. It'll always be a place for credit unions. That's what we're discussing here. So figure out how you can help, get in touch with me. This is RJMegarvey at gmail.com, Robert McGarvey again, and that's rjmeggarvey at gmail.com. Get in touch, we'll figure out a way that you can help. We need your support, we want your support, we thank you for your support. The C U 2.0 Podcast.