Publications

Christopher Zimmerman:

Hello, and welcome to the Stradley Ronon Talking Investment Management podcast. We have a special guest today, Erik Sirri. Erik serves as an independent trustee and is the current chair of the board of trustees of the Natixis family of funds, which includes a range of open-end funds, ETFs and interval funds. He also serves on the FINRA Board of Governors and previously served at the SEC, including as director of the SEC's Division of Trading and Markets. We're also joined by Sara Crovitz, a partner here at Stradley Ronon and co-chair of our investment management group. And I'm Chris Zimmerman, a partner at Stradley Ronon. Erik, thank you very much for joining Sara and me today.

Erik Sirri:

My pleasure.

Christopher Zimmerman:

Today we're going to talk about what good board oversight actually looks like in practice, especially in some areas that have become increasingly important and in some cases in increasingly complex areas for fund boards to oversee. And so with that, I'll turn it off to Sara to start us off here.

Sara Crovitz:

Thanks, Chris, and thank you so much, Erik. So before we get into some of the more detailed or more sort of complex issues, Erik, how do you see from just a very general perspective, we often talk about the difference between day-to-day detailed looking at what a management company is doing and oversight. And one way people talk about that is boards should have their nose in and their toes out. How do you think about a director's role vis-a-vis management?

Erik Sirri:

I think for the most part it's generally clear, nose in, fingers out, however it is, that's about right. But I will tell you, from time to time, it's a nuanced matter and you can't always tell. So for example, something that might be very, very clear is you wouldn't tell a portfolio manager to buy a stock or to sell a stock as a trustee. That would clearly get into the role of what the adviser and portfolio manager are doing, but it can become more nuanced when you talk about, say, issues around starting a fund. Some of the questions around then would properly be oversight questions. Where is the business decision to start a fund? That would clearly sit with management.

Sara Crovitz:

And speaking more about the new product launch, what are the kind of considerations that you think the board should be looking at and what is it really that the board should leave just to management? How have you seen that work well?

Erik Sirri:

I think when something like a new product launch works well, there's never a surprise. So that usually starts well before that final board meeting where you start signing paper. I think the best sort of processes occur when management brings the board in early, lets them know months ahead of time, "Hey, we're thinking of starting a particular product along this line of strategy." And then as the key dates move in, give the board progressively more and more information, more detailed information, I think it's especially important to leave time for trustees to ask questions. One mistake I've seen is things that go on where management has a fulsome, complete presentation but underestimates the amount of time that will be needed to answer trustees' questions. And I think you really want trustees to be comfortable before you have them go ahead with something like a new product launch.

Sara Crovitz:

One thing I've sometimes seen is the thought of having two bites at the apple. So you might have one, you'll have a meeting where they're going to, as you say, talk about here's our strategy plans for the coming year. Then you have a meeting where maybe they're like, "Well, here we have the registration statement, we have the basic strategy. Can we get authorization to do the initial filing?" And then the next meeting would be, "Hey, now we're really ready. Let's talk about the agreements we need to get signed. Let's do our 15(c)." Is that something that you've seen?

Erik Sirri:

Very much so. And some products, if you're an equity shop and you start another large cap equity growth fund, that's probably going to be a very measured, comfortable launch. But for instance, if you are running such a shop and for the first time you use some kind of equity derivatives or fixed-income derivatives for some reason, that's going to be a different discussion if the trustees have never had experience with derivatives in a product. When you have something like that, it may even require education. And that's a judgment call I think with management. Management, when it does its job well, it understands trustees' knowledge set, where their points are, where they're likely not to be comfortable, and it manages that accordingly. I think one of the best things that happens is early conversations between management and the board chair or management and the key committee chair offline, away from a board meeting to talk about what's coming so that management gets some insight from the appropriate chairs about how they feel about where their members are at.

Christopher Zimmerman:

I think, Erik, those are just some great points. Good process really matters. The board needs the right information, enough time to consider it and to be able to ask questions and follow up. And I think to one of your points earlier, not every issue requires the same level of attention as well. A routine matter, even for a new fund launch that the board has seen many times may require a very different process than from a novel product or valuation issue or perhaps something that raises a conflict as well. So I think a part of good oversight is not just asking the right questions, but recognizing where something needs more attention than usual as well.

Erik Sirri:

I think that's right, and it's not just new to the board. A new product may bring something new to management. So if the management runs institutional money, and going back to my previous example, if they've been using swaps for 10 years, then the issues that the board needs to confront are largely about their own education. But if management itself is new to the particular instruments, that's a more robust discussion, I feel.

Christopher Zimmerman:

Going back to another point is that the line between oversight and management, that line can be also important where an individual trustee brings significant expertise to an issue, whether investments, in law, accounting, tech or perhaps some other area. And that expertise can be tremendously valuable, but the board should use that expertise to enhance its oversight but not step into management's role. And so if the board starts making operational choices itself, it can blur that line and sometimes potentially compromise its later oversight.

Erik Sirri:

That's absolutely right. And we've all seen as we look to get trustees on our boards, situations where you want expertise. A portfolio manager, ex-portfolio manager can be valuable. Someone who served as counsel to fund boards to have them on your board could be valuable, but you certainly don't want an extra lawyer on your board per se second-guessing fund counsel, nor do you want that portfolio manager second-guessing the strategy of the PM. Healthy questions, yes, but not the other. But I would say expertise is something the board has a right to. The board does have the ability to bring in outside experts. I think that's actually mentioned in either the rule or the statute. You would be able to correct me, but we can bring in experts as we see fit. Sometimes they can be as simple as Broadridge to bring in information about performance, but you can get more nuanced experts as well.

I could ask or the board could ask for say an audit firm to come in and reexamine our allocations of cost when it comes to 15(c) profitability. That I think is entirely reasonable. It's the board's judgment when to use those experts, but I think they certainly have the right to do so. And I don't think that implicates the question of management versus oversight. That's clearly, or it should clearly be, an oversight role.

Christopher Zimmerman:

All right, so I think we would like to pivot to some valuation and liquidity discussion at this point. Erik, one of the statutory responsibilities of fund boards relates to fair valuation. Of course, Rule 2a-5 permits a board to designate the performance of fair value determinations to a valuation designee, which is typically the adviser, but the board retains an important oversight role. Now, oftentimes liquidity goes hand in hand with valuation, particularly when we're talking about private assets. There may not be an observable market price, transactions may be infrequent, and valuation may depend more heavily on assumptions and judgment. So from an oversight perspective, what are the questions that you think boards should be asking? And related to that, how deeply should the board be getting into individual fair-valuation determinations from your oversight perspective?

Erik Sirri:

Maybe let's talk about valuation first since that was what you mentioned last. So valuation, as you mentioned, it's a statutory responsibility. It is, when it comes to certain fixed-income expert instruments, it becomes a technical exercise. It really does. You're pulling together a lot of data sources. You may be, if you're fair valuing securities, you may be using algorithms or formulas that are not commonplace that trustees may not know about just in their day-to-day business experience. At that point, I think you look at process, you try to understand the process that management has in place to value securities. You assess how robust that process is. And going back to one of the most basic things you do, you look for conflicts. There can be so many conflicts in a valuation process, the most simple being the portfolio manager, other things equal, might rather see that security valued more highly than otherwise. That doesn't mean they're going to do that, but you at least have to be sensitive to that.

Christopher Zimmerman:

I think those are some really important considerations, especially as funds increasingly invest in private assets. We've just seen recently a number of recent regulatory developments that really underscore the importance of valuation, including SEC staff guidance on private assets, proposed rulemaking in a couple of different areas, and changes to accounting standards from FASB. I'll just note that two of those proposals, the SEC recent performance fee proposal, which is intended to expand retail investors' access into private markets, as well as the proposed amendments to Rule 17a-7, which just came out recently, they deal with different issues, but both are good reminders that I think valuation judgments can matter well beyond calculating a fund's NAV. And I think that goes back to our broader oversight point. The board should understand the valuation process as well as the governance around that as well. So excellent points, Erik. Thank you.

Erik Sirri:

I would say thinking a little more deeply about the question you asked, it would be very, very surprising if a board got involved in the valuation of a particular security absent some exigent circumstance. But I will tell you, on our board, we will from time to time ask management to bring some typical securities and their fair evaluation process around those. So we may work through, the audit committee may work through one or two securities in the course of an audit committee meeting just to let the committee understand the techniques and the methods that are being used. It ensures that the trustees understand the depth, understand the process of what management is going to, and should something happen down the road, they have a basis to ask the questions. I think being knowledgeable about this, that's part of good oversight. You brought up a question I think earlier that essentially asks, how deep does your knowledge have to be? That's a tough question. I think just deep enough.

Sara Crovitz:

And Erik, how do you feel like the Rule 2a-5 implementation, it's been a few years now, do you feel like that the roles have shifted in a more appropriate way given the reporting and the way that the SEC basically permitted boards to designate management to help with valuation and recognizing that the board really isn't appropriately in the day-to-day valuation? Has that changed or do you feel like that reporting was really what you had anyway?

Erik Sirri:

It's sort of interesting. The reporting that we had anyway was replaced by a different kind of reporting, and I think it's useful. Every so often in my mind, I harp back to that old chart that we used to have that I found useful. But that said, I think it's a very structured process. It's something we understand, and I would mention something we haven't talked about up to this point. We're not alone when we do this work. We have not only experts like valuation experts who are brought in, but also we have a CCO and we rely on our CCO for all sorts of things. I think anytime you talk about board governance and the process around board governance, it's hard to imagine that wouldn't go hand in hand with a tight relationship with your CCO.

Sara Crovitz:

So let's switch gears again. One topic that it seems that directors and the rest of the professional world can't stop talking about is generative AI. Erik, how are you thinking about the boards that you're working with, thinking about the board's role in understanding how advisers and service providers use gen AI?

Erik Sirri:

I think a particular challenge there is the speed with which that is changing. So unlike valuation or liquidity where we have frameworks, we have a rule and we know what to ask, here, whatever the adviser or service providers are doing today, they're almost certainly not doing the same thing a year later and maybe not even a quarter later. So I think that brings in some interesting questions and some things the board has to think about. So going back to the basics, hopefully you would have some questions about AI in your 15(c) process. So you might go back to your service providers, management, the adviser, others, and ask them questions about their use of AI that can be very targeted, but you want to have some specific language, I think, where that language gives a clear response where you understand what they're saying. And then I think it brings up an interesting question of, can you be comfortable with that response sitting on it for 12 months and then refreshing in the next cycle of contract renewal, or should you get updates?

I think there's a colorable case that you can make for getting some updates along the way, asking management for some assurances that there hasn't been a substantial or material change to the way your service providers have talked about what they've been doing with AI.

Sara Crovitz:

Yeah, it feels like in 2025, everyone was sort of starting out, we're using gen AI for marketing or back office or efficiency, but it does feel a little bit more this year you're really hearing management talk about no, we're really using gen AI in the investment management process, whether that's in research or analysis. And so it does feel like that brings more pressure on the board to consider how that is being done and whether it's being done appropriately.

Erik Sirri:

It really does. And I think you often hear assurances that say things like, "Yes, we use AI to do some function, but we have a human check it as well." And at some point to get the full leverage out of AI, a human's not going to check something at some point. It just won't be doable. Now, maybe that's already happening today when the investment manager has an AI tool summarize 50 research reports on the same company. That wouldn't be surprising to me, but we'll have to understand, as a trustee, I would want to understand where management is in terms of loosening the reins on AI and to the point where things are going operational without a human checking the output. I think that's going to be a focused discussion with the board.

Sara Crovitz:

Yeah, no, that leads me to another consideration for use of gen AI, which is of course directors' own use of gen AI. So I think that we are certainly hearing here from our clients, many directors really do want to have the opportunity to use, there's some gen AI sort of tools out there that I think are being presented to be helpful for boards in summarizing material or doing other things. I think that the lawyers honestly have been a little reluctant. I think we have some concerns around privilege and issues like that, but clearly those tools could be useful at some point maybe. And so I think for us and working with our clients, I think it's a conversation, I don't know how strongly your boards have talked about that, but it's certainly something that I hear in the boardroom from a lot of my trustee clients.

Erik Sirri:

Well, certainly, I mean, it's easy to see the draw, isn't it? To say, "I have hundreds if not thousands of pages of board material here. I've been getting that every quarter for some period of time. Wouldn't it be nice if I could ask a question of this amalgam of data to help me do my job better?" It's pretty easy to see how as a trustee, if you've written prompts and if you're broadly familiar with AI, where that would be useful. But as you pointed out, Sara, counsel the independent trustees have counseled us to be very, very careful about that for just the reasons you said. And I think, I can only speak for myself, but our boards have steered clear of that. Not the least of which is most of us don't have tools that keep that material within the confines of our meeting space. If I were to use a public tool that I have that maybe I paid $20 a month for, I don't have the security protections on that, that the adviser that management's going to have and their desktop AIs that they're using.

Sara Crovitz:

No, that's absolutely right. But even where, for instance, your adviser might have an internal tool, having directors want to use that tool or have the information fed through the adviser's tool and then you send it to the directors and then you've sent it to the directors and there are issues there. This is, as you say, an area that's advancing very rapidly. I imagine at some point there will be a service provider out there that solves this issue and finds a way for directors to use it inside their own sandbox. So we'll keep an eye on that.

Erik Sirri:

Yeah, I think we're all going to be interested in where the advice from our counsel comes down on that point.

Christopher Zimmerman:

Going back for a minute to adviser use of AI and board reporting, Erik, I think boards need to understand the significant use cases, the risk and controls, but doesn't necessarily need to understand the underlying technology at the same level as people implementing it, as you mentioned before. So other than frequency, do you have any other views as to how much information that you want to see? I know you mentioned 15(c) materials already and then some sort of frequent, maybe a semi-annual or a quarterly update on the use of AI, but anything in terms of substance that would help you oversee advisers' use of AI?

Erik Sirri:

We have asked for particular use cases to be brought in front of us. So even from our auditor, we've asked for use cases. So as they prepare their materials that come to the audit committee, we've kept a pretty constant drumbeat on them. How is that going? What is your AI process? Where is it being used? With management more directly, I think we've asked them targeted questions about where is it in particular in the investment process. I think that's something that it's very easy to understand how it could fit in. And a step a board could take is when PMs come in, they could request of management that there be at least a slide that talks about that PM's current use of AI, not in great detail, but just to keep the board up to speed on what's going on. But yeah, I think understanding use cases, getting assurances from management, making sure you understand what's in your 15(c), and keeping those 15(c) questions refreshed to reflect the current times that you're in, I think these are all things that boards are going to have to do with AI.

Christopher Zimmerman:

Yeah, and I think materiality is certainly an important part of that as well. I think one question boards can ask is, "If AI gets it wrong, what happens?" And in the investment decision-making process, that could have potential consequences. So to that end, I think if there's greater potential consequences, I think boards are going to want to understand the use case and controls around that as well.

Erik Sirri:

Well, that's an interesting question. If I'm an active portfolio manager and I think China's the place to be and I get it wrong, no harm, no foul, that's not good for the shareholders. Returns aren't as high as they'd like, but the matter would sort of end there with the trustees' oversight of that PM. But if an algorithm did something incorrect, let's say if an AI algorithm processed its inputs incorrectly and chose a set of stocks or other securities it shouldn't have chosen had the algorithm worked correctly, now you may have something that's actually a mistake and that starts to connote other things. I mean, you tell me, if your algorithm is poor, does that start to bring up AXA or those kind of issues to the staff were they to find out? I'm not really sure. I just think it's interesting. The active portfolio manager who makes a mistake in his judgment, his or her judgment, not an issue, but a poorly coded AI algorithm or approach, I'm not so sure. I'd be curious to know what you think about that.

Christopher Zimmerman:

I think there certainly can be model risk in terms of implementing a model, and I think that's really, really good and interesting questions. And this is why we see disclosure about model risk in the fund's prospectus as well, but it's a really interesting question.

Sara Crovitz:

I was going to say interesting too, because you could have a gen AI model where you put in, you're asking it to spit out its investment recommendations and it does one thing the first time you ask it and actually it does something different the second time you ask it, even if it has the same assumptions and criteria because it's learning and it's iterating. So I don't even know how you would consider that. Is that a mistake? Is that just the way an active manager might say, "Gosh, I got some additional market color, and now I'm changing my views"?

Erik Sirri:

I don't know. That's why we have counsel help us with those problems. No, it's good. Chris, you asked about materiality. The other thing that I thought of is when it comes to disclosure, to the extent parts of our offering documents are written with AI, well, there's strict liability there. So you could have a relatively minor mistake that isn't material in the sense of the way an ordinary person would read plain English, but within the context of an offering document, if it happened to be in a fee table or something with performance, even if it was minor, that could be significant for the fund and for trustees.

Christopher Zimmerman:

Yeah, I think that's a great point. I think we've all experienced hallucinations with respect to AI, and I think everyone is and should be properly concerned with any material misstatements in a fund's prospectus. And those are big considerations that I think fund groups are going to have to deal with when they are thinking about implementing AI in that process to the extent they are.

Sara Crovitz:

So this has been a really great, fascinating conversation. Erik, we just want to thank you again for making time to chat with Chris and me. For those listeners out there, we hope to see many of you at the IDC conference in Scottsdale. Thanks a lot for listening.

Erik Sirri:

My pleasure. Thank you.

End:

From obtaining industry-first exemptive and no-action relief to assisting in the development of novel products, we have helped our clients become and remain industry leaders. We touch trillions worth of assets as counsel to more than half of the top 20 U.S. fund firms, serve as counsel to independent directors to more than 100 fund groups, and represent more than 40 fund sponsors. Learn more about us at stradley.com.