Economy & Markets
Eye On The Market podcast
Subscribe today and never miss an episode
Join Michael Cembalest, Chairman of Market and Investment Strategy, as he explores a wide variety of investment topics, including the economy, policy and markets. Your subscription, using your preferred podcast provider, will feature new episodes with the release of each Eye on the Market publication.
Listen on Apple Podcasts
Listen on SpotifyThe Deep End: 2025 Alternative Investments Review
Good afternoon, everybody. This is Michael Cembalest with the December Eye on the Market podcast. This is on our biennial Alternative Investments Review. I started writing this piece about a decade ago. An idea was to focus on returns, risk underwriting, and the outlook for private equity venture hedge funds, real estate and private credit. And it’s that time of the year or two-year cycle again for us to do this.
A couple of quick things. First, I apologize for the Tommy Bahama sweater, but I’m going fishing for Trini—for tarpon in Trinidad next week, and I need to wear the sweater with the fish on it to get me mentally ready and dialed in. And then just a couple of quick slides on another topic that we’ll be talking more about in the, in the, Outlook, which is the debt explosion in the fourth quarter related to all the hyperscalers and data centers. The debt markets were a pretty boring place until the fourth quarter as it related to things, related to data centers. And then all of a sudden, let’s see if I can get this thing to work, there was a, there was an absolute explosion of about $150 billion of borrowing from Meta, Amazon, Google and Oracle. And here we’re including all the debt, including the hotly debated Meta- Blue Ow—Blue Owl structure that, that Meta convinced a bunch of their accountants to allow them to avoid consolidating. But we’re including that here.
So here you can see this explosion at a $150 billion on top of what was really kind of $50 billion on a year-to-day basis. That’s a lot of borrowing. Now to be clear, the net debt to EBITDA ratios for most of the big AI-related companies are still pretty low. Oracle is and maybe IBM are the exceptions. The rest of them are very low, if not negative. Google, Nvidia, AMD, for example, have more cash and cash equivalents than short- and long-term debt. So those ratios are zero. But these numbers are starting to move. And the numbers from Meta and Oracle in particular were significant relative to their amount of cash flow. And unsurprisingly, we’re starting to see the credit default swap markets reflect that. Here you can see the spreads, zooming on Oracle and also Intel with smaller adjustments for the rest of the AI companies.
Anyway, we’ll be talking more about that in the Outlook. So let’s get to this guy. This is a picture of a guy in a pool. And the title of the alternatives piece this year is called “The Deep End.” And the, this is a guy that is up to his neck in and, in investments that he would otherwise like to already have sold. So let’s get into it. This, this was another version of the cove, and for those of you that used to watch this show, you’ll know what I mean.
Okay. So alternative assets are continuing to attract a lot of capital. If you look at private equity and venture assets as a share of world equity market cap, the publicly traded equity markets over the last decade or so, they’ve doubled. So they’ve gone from 6% to around 12%. And the banks have played a pretty significant role in financing a lot of this growth in the private equity and private credit industry. And we have a chart here that shows private equity and private credit loan exposure by the banks as a share of their Tier 1 capital. And you can see, you know, among the top three there, you’ve got Morgan Stanley, Goldman and Wells.
So this is not, the basic synopsis of our analysis this year is on the surface of this pool, it doesn’t look like a lot has changed since two years ago. The top-quartile funds in buyout and venture are still outperforming public markets. The median funds are still kind of tracking public markets. Diversified hedge funds still are outperforming risk-adjusted benchmarks, and private credit funds are delivering higher returns than the regular, broadly syndicated leveraged loan market. That said, a lot of managers are swimming at the deep end of the pool. The buyout and venture portfolios that we analyzed are full of unmonetized companies dating back to the middle of the prior decade. Hedge funds remain, have record levels of asset crowding, concentration and very high bid exposure. And then private credit, unsurprisingly, have loosened underwriting standards after a deluge of committed capital.
The private credit section is the longest section in this year’s paper, in part because of all of the debate and issues and, and cockroach allegations swirling around. So let’s start here. The pool is full. This looks at unmonetized venture and buyout by vintage year as a percentage of all of the value that has been reported to LPs. So let me give you an example. Let’s go to the 2016 vintage year. We’re almost 10 years out from the 2016 vintage year, so you can be forgiven for thinking that those vintage years that the average manager would have sold a lot of the companies that they had bought. For the median buyout fund, they’re still around 30 to 35% of the assets left. And for the median venture fund, this is eye popping, the median venture fund has almost two-thirds of its, its assets that’s invested in still unmonetized from 2016, which is kind of amazing. So that’s what’s been happening in the industry, a stark slowdown in the rate of monetization and therefore distributions. And when I first started working on this and people said to me, well, how do you know that’s different from the past? How do you know that, that that’s not always the pace? And okay, good question.
So I went back in, using five-year intervals going back to the early 90s, and we looked specifically at six-year-old funds to nine-year-old funds. And the current cycle has the highest remaining value as a percentage of total value, which is a, is a jargony way of saying they have more stuff they haven’t sold yet. The trunk of the car as a share of all the value that’s been reported to LPs. And that’s the case in buyout. And it’s also the case in venture current cycle is has the highest remaining value percentages.
So now, so what have these dynamics done? And most of our clients know this. Global private equity distributions have been roughly flat, even though there’s more and more money being invested. Private equity exits have declined in number terms. The average buyout holding periods are rising. All of a sudden, continuation funds and secondary funds are rising as a, as a, as a way of getting some kind of exits. And now the good news is there’s a tentative capital markets recovery taking place, and it could not have come at a better time. And there’s a few different ways to think about what that capital markets recovery looks like. You could look at IPO activity in terms of number of IPOs or deal value. You can look at the average return on the first day of trading, and IPOs has kind of picked up. Secondary activity, secondary placement activity has picked up. And then maybe most importantly, announced M&A volumes by quarter for sponsors has picked up. So this capital markets recovery is underway. It could not have come for a better time for a lot of these managers who are lugging around tons of companies that they would otherwise like to sell.
So all right, let’s, one of the things we do in this paper here is we look at absolute returns, and then we look at relative returns in terms of relative to benchmarks. And private equity and venture, you know, there are these swirling debates amongst academics, and, and also investors and industry analysts in terms of how do you measure performance on an absolute basis? How do you measure performance on a relative basis? And there’s no single answer. The only red line I’ve got, my hard line is pick a methodology, and then you have to use that same one. You can’t hopscotch from metric to metric depending upon, you know, what it does to the data, what you like to look at. So this slide looks at buyout performance in terms of multiples on invested capital and then internal rate of return. For around 20 years, pardon me, through 2019, the median buyout manager multiple on invested capital hovered around 2x with IRR of 15 to 20%. And for most investors, that was perfectly fine.
For the more recent vintage years, the returns are lower, but it takes time to figure out if the recent vintage lower return numbers are just a J curve issue in terms of when money gets put to work, or if it’s a performance issue. So the, the story on absolute returns and buyout look fine for median managers and then in particular for the, for the 75th percentile manager.
Venture is a little bit different story. For several years following the dot-com bust, venture returns barely generated positive absolute returns. I mean it was pretty bad. Then following the financial crisis, things got better because in a scarce capital environment, people were able to make better valuation investments. And then the median MOICs were also around 2x and median IRRs of 13 to 15%. Since 2019, venture performance has been slipping. And I don’t think this is just a J-curve issue. I think a lot of capital was destroyed during the metaverse SPAC investment cycle. And so the poor performance of 20, 2021, 2022 vintages is not just a J-curve issue. They’ve got some, some, bad investments in there that they’re getting around to recognizing.
Then in terms of relative performance, you got to pick your poison. I think a sensible place to start is where most of the industry starts, which is to look at performance relative to the S&P 500. And the good news is, for the most part, from the early 90s to 2020, the median buyout manager, whether we’re looking at the average manager or the median manager, consistently outperformed the benchmark S&P 500. And you know, that’s good. And as usual, the top-quartile managers did much better than the median. And then the fourth, fourth-quartile private equity managers did do poorly since 2020. The buyout performance relative to the equity market has declined. But again, I think this is a J-curve issue. And also the fact that since 2021, the S&P 500 benchmark, we’re comparing private equity buyout returns to, you know, the hyperscalers and the Mag 7 and a bunch of stocks that look nothing like what are in private equity portfolios.
So we, and then on venture, the picture is not as good for investors. You need venture—limited partners need to consistently invest with top-quartile venture managers to outperform the S&P 500. The median venture manager has at best matched public equity markets. And the huge gap between the average and median managers, kind of a clue that there’s a select small group of venture funds that are doing very well, and the rest aren’t, and then fourth-quartile venture managers have been a money pit, like destroying value relative to the equity market consistently since, since the 1990s.
So we, in the piece, which you can read, we get into all the other stuff people debate. What are other benchmarks that could be used? How about a leverage benchmark? How about a small-cap benchmark? We talk about risk-adjusted returns and why that makes no sense in terms of venture and buyout and other illiquid investments, because the, there’s massive autocorrelation of returns. We talk about evergreen funds and secondary funds, dry powder and multiples. So those are all the additional topics that we cover. So let’s move on to the next one, which is hedge funds.
So we, I started doing this around 10 years ago, and it’s an approach that you either accept or you don’t. Most people I talk to do, although maybe they’re just being polite, but this is how I like to look at this question: You can look at a whole bunch of hedge funds, and you can look at their risk and their return. And then you can say, okay, let’s, what would a sensibly diversified institutional-type investor do, whether you’re talking about a, a risk, a plan, a foundation, a large family office, what would a sensible, diversified portfolio look like? And in my opinion, that would be 20 hedge funds. You know, with a diversification mix across that driven macro long, short, you know, and, and a couple of other strategies.
And so what we’re showing here is, there’s actually a really good amount of correlation benefit that comes from doing this exercise, because when you start throwing all those hedge funds together, the returns go down a little bit, but the volatility collapses. And the way this chart works is each blue dot is an individual hedge fund, one of 700 that we looked at. And then the gold dots are hedge fund composites with 20 hedge funds each, again where we force them to have a certain number of each of the style, so you’re not just owning 20 event-driven funds. And you can see how the hedge fund cluster of diverse hedge fund composites really collapses volatility compared to the blue cluster of the individual funds.
So how does, how would you feel if you owned that gold cluster, if all you cared about was absolute return? You’d be disappointed. Because if we just take a 70/30 mix of stocks and bonds using the S&P 500 and a Barclays AGG benchmark, that over the last five years, that would have returned around 8%. And most of the gold dots return less than that.
But look at that black dot on this chart. The volatility of the 70/30 mix was almost 12%, which was higher than almost all of the hedge fund composites. So what we did is, we said, all right, let’s create a bunch of different stock/bond composites with different amounts of stocks and bonds, and then match each hedge fund to a benchmark that has the same volatility. And I started doing this a few years ago with one of the large California ERISA plans. It was a state plan. You know, they got a lot of press because they said, we’re really disappointed in our hedge fund platform, we’ve terminated the managers, we’re shutting it down, the returns were way too low, and we were disappointed. And when I looked at all the constraints that they had put on their managers, they basically gave them a fixed income level of risk, so of course they were delivering fixed income levels of returns. And so ever since then, I’ve been kind of committed to this mission of making sure that we look at hedge fund performance on, on a risk-adjusted basis relative to the risks that are being generated by these managers. And so, as you can see here, the vast majority of hedge fund composites outperformed their volatility-adjusted benchmarks. So I consider that to be good news.
We get into other issues related to selective reporting and survivorship bias. Asset crowding. I mean, the, the, the share of hedge fund stock holdings in the top quintile of data is very high, which again is a jargony way of saying they, they, they own a lot of the big hyperscaler momentum stocks. A hedge fund crowding index is at the highest level that it’s been at for the last 20 years or so. So, you know, there’s, there’s a lot of momentum risk in the hedge fund community.
Okay. So private credit, there’s a lot to talk about. Somebody said something about cockroaches, and it reminded me I went to see a play with my brother-in-law at one point where he was dressed as a bug, but that, that’s, that’s beside the point. What was interesting, what’s interesting to me is that during all of this discussion about private credit, it’s not like private, risky private credit is, is being added on top of other kinds of risk. It’s actually displacing both leveraged loans and high yield bonds. So if you look at the total share of credit as a share of GDP in the U.S., the amount has been roughly the same over the last few years. It’s just that private credit’s been taking share from other stuff.
Now, where are we? Looking backwards, okay, I heard—there’s backwards and then there’s forwards, right? Let’s do them separate. Looking backwards, the performance of private credit versus leveraged loans has been pretty good. And so, again, this is as much art as science, but we put together a composite that looks at three different private credit benchmarks. We combine them into one, and then we compare that to leverage loans. And the composite has been delivering around 300 basis points a year in excess of leveraged loans, which it should in an economic expansion, right? So, given the, given the concentrated nature of private credit portfolios, given the lower credit quality of the companies, at least as the way the rating agencies look at, you should be compensated for that. And during an expansion, 300 basis points sounds about right.
So, and as we look at things right now, there are limited signs of stress in the private credit. Well, there are some signs of stress, but they’re so far limited. The ratio of upgrades to downgrades looks pretty good. Default rates are very low, even when you include selective defaults, which are like defaults that would have happened, but some kind of reorganization plan to avoid it. Nonperforming exposures, a share of total loans, whether we’re looking at middle market loans or the broadly syndicated loans, are still below 1%. And this issue gets a lot of press. The PIK share of gross investment income is barely moving, right? So what does that mean? PIK is payment in kind. In the private credit markets, most of you are all aware that if companies can’t come up with cash, they can, they can pick the payment, which means I’m not paying you just to add it to my balance at the end, and then we’ll assume are all good. And so on an accrual basis, there’s money, but on a cash basis there is a gap. When we look at both publicly traded BDCs and non-traded BDCs, the PIK share of gross investment income hasn’t really moved since 2023 very much, so that tells us that, that there’s not a bit of marked deterioration there.
Okay. But this is, I think this next topic, I think is the most important thing in all of the whole private credit debate discussion, which is a couple of years ago, Moody’s did a really good analysis on what goes on in private credit compared to the broadly syndicated loan market. And what they found two years ago was that the managers in private credit are being much more conservative about underwriting. One way they looked at that was to say how much to be part of either—my solo cup. How much do private credit managers allow EBITDA backs , which is this kind of fictional thing that I can’t even believe exists in the first place, but some of these private credit borrowers get to assume a certain level of synergistic EBITDA or cash flow growth in order to avoid triggering certain covenants. The broadly syndicated loan market allows that sometimes with no cap, much less of that is allowed in private credit.
And then the chart on the right really digs into the stuff that I’m fascinated by, which are the legal terms and conditions under which these borrower deals get done. So, and the broadly syndicated loan market in those blue bars, a lot of things are allowed inside maturities, which is, you know, you all of a sudden you can borrow money at a shorter maturity than an existing loan. If, if there’s unrestricted, unrestricted covenant room, in terms of restrictive payment clauses, you can convert that into raising your debt ceiling. If you sell an asset, you don’t have to use 100% of the proceeds to repay the debt. There’s an automatic debt increase per dollar of new equity committed. And then, you know, I, sometimes there’s no J. Crew IP blocker, and for those of you who are in the legal community you understand what that means, so whereas in private credit, very rarely, you know, only 10 or 15% of the time or any of these things allowed. Now, Moody’s has not updated this analysis in, in the last two years. As it stands, two years ago, this is what it looked like.
My sense is that some of this is still true, but that the underwriting, the underwriting characteristics in private credit have been getting easier and looser, and so this gap that existed two years ago may not be there as much as it was. We can see that with a following four exhibits. So the covenant light share of new loans, private credit rising rapidly relative to the broadly syndicated loan market, private credit deals without maintenance covenants. The bigger the loans, the more frequently they don’t have any maintenance governance at all. The median leverage of BDCs has been going up. And then lastly, S&P did an interesting analysis where they found that the only reason that some borrowers still have a lot of headroom with the covenants is because the, the debt to EBITDA ratio that was used as a covenant threshold was like 8, 10 or 12x. In other words, you haven’t, you haven’t tripped the covenant because they were set on such a toothless basis to begin with. So the private credit standards are changing.
So to wrap up, the performance of private credit managers has been pretty tightly bunched since 2014, again, as one would expect in the absence of a lasting recession. And remember, there was a recession during COVID, but it wasn’t a real recession. And I have a chart in here that explains what I mean by that.
If you look at the prior three real recessions, the S&L recession and the dot-com collapse, and then, and then the global financial crisis, high yield default rates went up a lot. High yield credit spreads went up a lot, but, and they, and they went up and they stayed there for a while. During COVID, they didn’t. So they spiked very temporarily, and then they came back down pretty quickly. So, the point is it’s been difficult for investors to really tell it’s 15 years out from the last real durable recession, which private credit managers have been underwriting better than others. And you know, those are the things that we’re going to see because eventually, when the next recession does hit, you’re going to see a lot more manager dispersion than we’ve seen in the past. We conclude the private credit section with seven questions to ask your private credit manager, just so you can get an understanding of what kind of risks they’re taking.
Okay, so let me wrap this up. We’ve been talking here for a while. Real estate. So the fundamentals are still kind of bad or really bad, depending upon which corner of the real estate industry we’re talking about. But there are signs of life in, in, in commercial real estate, and I’m not surprised by that because we’ve just lived through the third biggest correction. First, we had the U.S. and oil prices, the great financial crisis. And when you look at the history of the declines in real estate returns, this one matches almost the two prior ones. So the more you have a correction and pricing, the more it’s time to kind of say, okay, I know there’s a lot of bad news, but let me start looking for opportunity. And we’re seeing signs in life in terms of investment returns bottoming out. Transaction volumes are picking up. Debt origination is picking up. Distressed sales are picking up. That’s always a sign of market clearing, which is a good thing.
And, and I know that a lot of people are spooked by this chart on this page, which shows the special servicing rates for, for office loans in the CME’s market. Okay, what does that mean? A lot of debt originators don’t really have the capacity to service loans and defaults or modification with their traditional servicing approach, so they get shunted off to these special servicing departments either inside or outside of the bank, and those special servicing rates are skyrocketing. They were 3% a couple of years ago, and now they’re 16%. And, but I would just remind you of a few things, and if you read the 20-year anniversary piece that came out over the summer, you’ll remember this. During the 2008 financial crisis, bank stocks in the S&P bottomed when only 8% of the eventual bank failures had taken place. And the same thing happened during the S&L—during the 1990 savings and loan crisis. High yield spreads peaked, right? The time for investors to buy at the peak, only half of all the eventual corporate defaults in the high yield market had taken place. And, and there’s many, many examples of this. The bottom line is that markets bottom way before the fundamentals stop deteriorating. And I have a feeling that we’re in that kind of environment right now for real estate, where the fundamentals are going to continue to get worse, but asset prices will be bottoming.
And I don’t want to dwell too much on office, as you can see from this chart, around 10 years ago office was king, right? Office was 35 to 40% of the average institutional commercial real estate portfolio. And that number started collapsing in 2015 and is now below 15%. So we shouldn’t spend too much time on office when industrial properties, including data centers and apartments, are capturing way more capital. But I will say the long office nightmare is gradually slowing. There is a slowing construction, underway. There is a slowing delivery pipeline. Sublease availability is starting to come down. There is, and there’s a very high concentration of vacant space in just a small subset of the square footage. So the, the vacancy problems are very concentrated. The rest of the market’s doing a little bit better.
And here you can see—I thought this was amazing—and this was as of the second quarter, it finally looks like demolitions and conversions into residential in the second quarter were higher than new supply delivered. Right. So that’s, that’s really a turning point in the, in the office narrative when office building is converted into other things or torn down are greater than the, the amount of new supply. So I thought that was a notable thing to talk about.
And one last chart that I just wanted to share here because I thought it was interesting. The work from home story hasn’t really changed very much. It’s still about 70%. So this looks at a bunch of different cities. Pardon me, that’s the, the back to work is around 70%. The residual which is work from home is around 30%. And to get to those numbers, we look at a combination of mass transit recoveries compared to 2019, a Stanford survey by, by, Nicholas Bloom and then the Kastle KeyFob swipe data. And so the most part, you know, Austin, the little bit of exception at 80%, most of the rest of the cities, the back-to-work ratio is in the neighborhood of around 70%.
Okay. I think that’s the last, let’s see, the last topic, and I’m not to spend too much time on this is, there was this executive order in August that talked about, let’s, let’s democratize alternative assets, and try to find ways that they can be added into 401(k) plans. And that that was, that was pretty notable. It got my attention. And so I added a section on this question of democratization of alternative assets. And certainly even before anything is done about trying to figure out ways of putting them in 401 (k) plans. Retail fundraising has been rising pretty rapidly. A few years ago it was like $25 billion, and it’s on track for $175 billion this year. So there’s a lot of retail money being made, raised in privately placed that read some BDCs interval funds and different kinds of private equity and infrastructure vehicles.
And there was an interesting paper that came out. I agree with parts of it. I don’t agree with parts of it. Some parts of it, I strongly disagree with. But it’s called “Private Equity, Public Capital and Litigation Risk,” and the general argument of this paper is that if the regulators don’t figure out a way to protect retail investors, the, the, the litigation process will, and that the markets and actual retail litigants through the courts are going to, are going to discipline the industry, even if the regulators don’t.
And their three strongest arguments are related to, seek to, to, unclear, vague language as it relates to things like fiduciary duty, how net asset values are computed, all the different expenses, and whether they’re reported on or not reported on, you know, for, for 1940 ACT vehicles they’re fine because you’re subject to the same protections and requirements as a regular mutual fund. But for alternative asset vehicles sold to accredited investors, they may not be required to show all their fund level expenses and portfolio level expenses, and things like that. So there’s some basis for litigation here. There’s a lot of counter arguments. And well, when I showed the paper to someone who for a living at one of the big white shoe law firms in New York writes these documents to presumably shield the alternative asset managers from retail litigation, he, he, he hit the roof. He actually jumped on a Zoom in a t-shirt because he was so anxious to talk to me about it. But, you know, they, they list their counter arguments and things like that, and you can read through those, why they believe that, that the legal litigation risk presented in this paper are offside.
The only thing I would say is that this whole democratization of alternatives thing, it may be accretive to portfolios in the long run, but if you get a sharp recession and a liquidity crunch at the same time, these documentation firewalls that were supposed to shield the alternative asset managers are going to be tested. And I concluded that section with the review of the ongoing litigation in 401 (k) plans related to excessive fees and, and performance issues, and, and who could forget a lot of the litigation that took place both around the failed auction rate security market around 15 years ago, as well as the SPAC boom and bust. So there are plenty of examples here where people thought they were firewalled. And it turns out that they weren’t as it relates to dealing with retail investors.
So that is a brief summary of what is otherwise a long slog through the alternative asset markets that we do every couple of years. And so that’s that. And we will see you again on January 1st when we’re going to launch our 2026 Outlook, most of which is going to, no surprise here, look at what’s going on with hyperscalers and pathways to profitability on $1.4 trillion of hyperscale or cap and spending since GPT was launched in 2022.
Thank you very much for listening, and wish me luck next week. Bye.
(DESCRIPTION)
Title card: JP Morgan, Eye on the market. JP Morgan. Presentation. Title: December 2025, The Deep End. 2025 Alternative Investments Review. Image: A man wearing a dark suit, white shirt, and tie stands waist-deep in a swimming pool. He looks concerned as he looks to the side, with a beach, ocean waves, and a modern house with large windows behind him. A video box on the right. Michael Cembalest has short hair, glasses, and wears a dark zip-up jacket with light blue trim. He sits in front of a virtual modern shelf that holds vases, gold decorative items, books, and geometric objects.
(SPEECH)
Good afternoon, everybody. This is Michael Cembalest with the December Eye on the Market podcast. This is on our biennial alternative investments review. I started writing this piece about a decade ago. The idea was to focus on returns, risk, underwriting, and the outlook for private equity, venture, hedge funds, real estate, and private credit. And it's that time of the year or two-year cycle again for us to do this.
A couple of quick things first. I apologize for the Tommy Bahama sweater, but I'm going fishing for tarpon in Trinidad next week. And I need to wear this sweater with the fish on it to get me mentally ready and dialed in.
And then just a couple of quick slides on another topic that we'll be talking more about in the outlook, which is the debt explosion in the fourth quarter related to all the hyperscalers and data centers. The debt markets were a pretty boring place until the fourth quarter as it related to things related to data centers.
And
(DESCRIPTION)
Slide: Hyperscaler Debt Surge. A bar chart tracks annual change in hyperscaler long-term debt from 2015 to 2025, with modest values through 2023 and a sharp rise in 2024–2025. Labels on the tall 2025 bar mark Amazon Q4, Meta Q4, Google Q4, Oracle Q4, and Q3. Text: Source: Bloomberg, Company sources, JPMAM, 2025.
(SPEECH)
then all of a sudden-- let's see if I can get this thing to work-- there was an absolute explosion of about $150 billion of borrowing from Meta, Amazon, Google, and Oracle. And here we're including all the debt, including the hotly debated Meta Blue Owl structure that Meta convinced a bunch of their accountants to allow them to avoid consolidating. But we're including that here. So here you can see this explosion at $150 billion on top of what was really $50 billion on a year-to-date basis. That's a lot of borrowing.
Now,
(DESCRIPTION)
Slide. A bar chart compares net debt to EBITDA ratios of Direct AI stocks for Q3 2025 and Q4 2025, with most companies clustered near or below the S&P 500 median while a few at the right show ratios rising toward 6x. Text: Zero indicates an excess of cash and marketable securities over short and long term debt, Source: Bloomberg, JPMAM, November 23, 2025.
(SPEECH)
to be clear, the net debt to EBITDA ratios for most of the big AI related companies are still pretty low. Oracle and maybe IBM are the exceptions. The rest of them are very low, if not negative. Google, NVIDIA, AMD, for example, have more cash and cash equivalents than short and long term debt. So those ratios are 0. But these numbers are starting to move. And the numbers for Meta and Oracle in particular were significant relative to their amount of cash flow. And
(DESCRIPTION)
Slide. A line chart compares five-year credit default swap spreads for several AI-related stocks from 2022 to 2026. Oracle and Intel rise sharply near 2026, the Apple Broadcom AMD IBM Dell HP average fluctuates at higher levels, the Alphabet Amazon Microsoft average stays lower and steadier, and Meta follows a mid-range path. Text: Source: Bloomberg, JPMAM, November 24, 2025.
(SPEECH)
unsurprisingly, we're starting to see the credit default swap markets reflect that. Here you can see the spreads zooming on Oracle and also Intel with smaller adjustments for the rest of the AI companies. Anyway, we'll be talking more about that in the outlook.
So let's get to this guy.
(DESCRIPTION)
Slide: The Deep End. Image: A man wearing a dark suit, white shirt, and tie stands waist-deep in a swimming pool. He looks concerned as he looks to the side, with a beach, ocean waves, and a modern house with large windows behind him.
(SPEECH)
This is a picture of a guy in a pool. And the title of the alternatives piece this year is called the deep end. And the-- this is a guy that is up to his neck in investments that he would otherwise like to already have sold. So let's get into it.
(DESCRIPTION)
Image: The same scene as before, but this time with BoJack Horseman giving us his back.
(SPEECH)
Oh, this was another version of the cover. And for those of you that used to watch this show, you'll know what I mean. OK.
So
(DESCRIPTION)
Slide: Synopsis. Two charts appear side by side. The left bar chart shows private equity and venture capital assets under management as a share of MSCI World market capitalization rising steadily from 2014 to 2024. The right bar chart shows bank loan exposure to private credit and private equity intermediaries as a percentage of large-entity Tier 1 bank capital, with FCNCA and MS appearing highest.
(SPEECH)
alternative assets are continuing to attract a lot of capital. If you look at private equity and venture assets as a share of world equity market cap, the publicly tradable equity markets over the last decade or so, they've doubled. So they've gone from 6% to around 12%. And the banks have played a pretty significant role in financing a lot of this growth in the private equity and private credit industry. And we have a chart here that shows private equity and private credit loan exposure by the banks as a share of their Tier 1 capital. And you can see, among the top three there, you've got Morgan Stanley, Goldman, and Wells.
So
(DESCRIPTION)
Slide. Text.
(SPEECH)
the basic synopsis of our analysis this year is on the surface of this pool, it doesn't look like a lot has changed. Since two years ago, the top quartile funds in buyout and venture are still outperforming public markets. The median funds are still kind of tracking public markets. Diversified hedge funds still are outperforming risk adjusted benchmarks. And private credit funds are delivering higher returns than the regular broadly syndicated leveraged loan market. That said, a lot of managers are swimming at the deep end of the pool. The buyout and venture portfolios that we analyzed are full of unmonetized companies dating back to the middle of the prior decade.
Hedge funds have record levels of asset crowding, concentration, and very high beta exposure. And then private credit, unsurprisingly, have loosened underwriting standards after a deluge of committed capital. The private credit section is the longest section in this year's paper, in part because of all of the debate, and issues, and cockroach allegations swirling around it.
So
(DESCRIPTION)
Slide: The pool is full. A line chart compares median remaining value as a percent of fund value plus distributions for venture capital and buyout funds from 2010 to 2024. Both lines rise over time, with venture capital reaching 100% in recent vintages and buyout rising close to that level. Text: Source: Steve Kaplan (U Chicago), MSCI/Burgiss, JPMAM, Q2 2025.
(SPEECH)
let's start here. The pool is full. This looks at unmonetized venture and buyout by vintage year as a percentage of all of the value that has been reported to LPs. So let me give you an example. Let's go to the 2016 vintage year. We're almost 10 years out from the 2016 vintage year. So you can be forgiven for thinking that those vintage years that the average manager would have sold a lot of the companies that they had bought.
For the median buyout fund, they're still around 30% to 35% of the assets left. And for the median venture fund-- this is eye-popping popping. The median venture fund has almost 2/3 of its assets that's invested and still unmonetized from 2016, which is kind of amazing. So that's what's been happening in the industry, a stark slowdown in the rate of monetization and therefore distributions.
And
(DESCRIPTION)
Slide. A line chart compares median remaining value as a percent of total value for buyout funds at different fund ages, with separate lines for vintages 2016–2019 in 2025, 2005–2008 in 2014, 2010–2013 in 2019, 1995–1998 in 2004, 2000–2003 in 2009, and 1990–1993 in 1999. Each line shows how monetization speed varies across fund age and vintage. Text: Source: Steve Kaplan (U Chicago), MSCI/Burgiss, JPMAM, 2025.
(SPEECH)
when I first started working on this and people said to me, well, how do you know that's different from the past? How do you know that's not always the pace? And OK, good question. So I went back using five-year intervals going back to the early '90s. And we looked specifically at six-year-old funds to nine-year-old funds.
And the current cycle has the highest remaining value as a percentage of total value, which is a jargony way of saying they have more stuff they haven't sold yet in the trunk of the car as a share of all the value that's been reported to LPs. And that's the case in buyout. And it's also the case in venture. Current cycle has the highest remaining value percentages.
So
(DESCRIPTION)
Slide: Consequences. Text.
(SPEECH)
what have these dynamics done? And most of our clients know this. Global private equity distributions have been roughly flat even though there's more and more money being invested. Private equity exits have declined in number terms. The average buyout holding periods are rising. All of a sudden, continuation funds and secondary funds are rising as a way of getting some kind of exits.
And
(DESCRIPTION)
Slide: Tentative capital markets recovery could not come at a better time. A set of four charts appears. One chart shows US IPO activity on a three-month rolling average with deal value and offering count rising sharply around 2020 then dropping and slowly climbing again. Another chart shows US average IPO return on the first trading day from 1995 to 2025 with bars fluctuating around a median line. A third chart tracks US secondaries activity on a three-month rolling average with deal value and offering count following patterns similar to IPOs. The fourth chart displays total announced M&A volumes by quarter from 2018 to 2025 with strategic and sponsor volumes stacked. Text cites sources from Bloomberg, JPMAM, GS Global Investment Research, and Morgan Stanley Research.
(SPEECH)
now the good news is there's a tentative capital markets recovery taking place. And it could not have come at a better time. And there's a few different ways to think about what that capital markets recovery looks like. You could look at IPO activity in terms of number of IPOs or deal value. You can look at the average return on the first day of trading in IPOs has kind of picked up.
Secondary activity, secondary placement activity's picked up. And then maybe most importantly, announced M&A volumes by quarter for sponsors has picked up. So this capital markets recovery is underway and could not have come for a better time for a lot of these managers who are lugging around tons of companies that they would otherwise like to sell.
(DESCRIPTION)
Slide: Buyout absolute returns. Two line charts sit side by side. The left chart shows US buyout performance by vintage year from 1999 to 2022 with lines for the 75th percentile, median, and 25th percentile, all peaking around 2014–2016 before declining toward recent vintages. The right chart shows North America median buyout IRR by vintage year over the same period with the three percentile lines rising and falling in a similar pattern. Text cites Steve Kaplan (U Chicago), MSCI/Burgiss, JPMAM, Q2 2025.
(SPEECH)
So all right. One of the things we do in this paper each year is we look at absolute returns and then we look at relative returns in terms of relative to benchmarks. And private equity and venture, there are these swirling debates amongst academics and also investors and industry analysts in terms of how do you measure performance on an absolute basis.
How do you measure performance on a relative basis? And there's no single answer. The only red line I've got, my hard line is pick a methodology and then you have to use that same one. You can't hopscotch from metric to metric depending upon what it does to the data, what you like to look at.
So this slide looks at buyout performance in terms of multiples on invested capital and then internal rate of return. For around 20 years, pardon me, through 2019, the median buyout manager multiple on invested capital hovered around 2x with IRRs of 15% to 20%. And for most investors, that was perfectly fine.
For the more recent vintage years, the returns are lower, but it takes time to figure out if the recent vintage lower return numbers are just a j-curve issue in terms of when money gets put to work or if it's a performance issue. So the story on absolute returns and buyout looked fine for median managers and, in particular, for the 75th percentile manager.
Venture
(DESCRIPTION)
Slide: Venture Absolute Returns. Two line charts appear. The left chart shows US venture capital performance by vintage year from 1999 to 2022 with lines for the 75th percentile, median, and 25th percentile, rising sharply around 2010–2014 and then declining toward recent vintages. The right chart shows North America median venture capital IRR by vintage year with the same three percentile lines, which peak in the early 2010s and fall afterward. Text cites Steve Kaplan (U Chicago), MSCI/Burgiss, JPMAM, Q2 2025.
(SPEECH)
is a little bit different story. For several years following the dotcom bust, venture returns barely generated positive absolute returns. I mean, it was pretty bad. Then following the financial crisis, things got better because, in a scarce capital environment, people were able to make better valuation investments. And then the median MOICs were also around 2x. And median IRRs were 13% to 15%.
Since 2019, venture performance has been slipping. And I don't think this is just a J-curve issue. I think a lot of capital was destroyed during the Metaverse SPAC investment cycle. And so the poor performance of 2020, 2021, 2022 vintages is not just a J-curve issue. They've got some bad investments in there that they're getting around to recognizing.
Then
(DESCRIPTION)
Slide: Buyout relative performance. Two line charts appear side by side. The left chart shows US buyout performance as a Public Market Equivalent ratio versus the S&P 500 from 1992 to 2020, with average and median lines rising through the mid-2000s and trending lower after 2016. The right chart shows the same ratio with 75th percentile, median, and 25th percentile lines, with the top quartile consistently outperforming and the lower quartile underperforming. Text cites Steve Kaplan (U Chicago), MSCI/Burgiss, JPMAM, Q2 2025.
(SPEECH)
in terms of relative performance, you got to pick your poison. I think a sensible place to start is where most of the industry starts, which is to look at performance relative to the S&P 500. And the good news is, for the most part, from the early '90s to 2020, the median buyout manager, whether we're looking at the average manager or the median manager, consistently outperformed the benchmark S&P 500. And that's good. And as usual, the top quartile managers did much better than median. And then the fourth quartile private equity managers did poorly.
Since 2020, the buyout performance relative to the equity market has declined. But again, I think this is a J-curve issue. And also the fact that, since 2021, the S&P 500 benchmark, we're comparing private equity buyout returns to the hyperscalers and the Mag 7, and a bunch of stocks that look nothing like what are in private equity portfolios.
And
(DESCRIPTION)
Slide: Venture relative performance. Two line charts appear. The left chart shows US venture capital performance as a Public Market Equivalent ratio versus the S&P 500 from 1999 to 2020, with average and median lines rising and falling together and the average staying noticeably higher for many vintages. The right chart shows the same ratio with 75th percentile, median, and 25th percentile lines, with the top quartile performing far above the median and the bottom quartile remaining well below. Text cites Steve Kaplan (U Chicago), MSCI/Burgiss, JPMAM, Q2 2025.
(SPEECH)
then on venture, the picture's not as good for investors. You needed venture. Limited partners need to consistently invest with top quartile venture managers to outperform the S&P 500. The median venture manager has, at best, matched public equity markets. And this huge gap between the average and median manager is kind of a clue that there's a select small group of venture funds that are doing very well and the rest aren't. And then fourth quartile venture managers have been a money pit, like destroying value relative to the equity market consistently since the 1990s.
So
(DESCRIPTION)
Slide: Other topics. A bullet point list.
(SPEECH)
in the piece, which you can read, we get into all the other stuff people debate. What are other benchmarks that could be used? How about a leveraged benchmark? How about a small cap benchmark? We talk about risk adjusted returns and why that makes no sense in terms of venture, and buyout, and other illiquid investments because there's massive autocorrelation of returns. We talk about evergreen funds, and secondary funds, dry powder, and multiples. So those are all the additional topics that we cover.
So let's move on to the next one, which is hedge funds.
(DESCRIPTION)
Slide: Hedge funds. A scatter plot compares annualized excess return versus T-bills from September 2020 to August 2025 with annualized excess return volatility. Blue dots represent individual hedge funds and brown dots represent hedge fund composites of 20 funds each. Most points cluster between 0% and 20% volatility with a wide spread of excess returns above and below zero. Text: Source: JPMAM, HFR, August 2025.
(SPEECH)
So I started doing this around 10 years ago. And it's an approach that you either accept or you don't. Most people I talk to do, although maybe they're just being polite. This is how I like to look at this question.
You can look at a whole bunch of hedge funds. And you can look at their risk and their return. And then you can say, OK, what would a sensibly diversified institutional type investor do, whether you're talking about an ERISA plan, endowment, foundation, a large family office. What would a sensible, diversified portfolio look like? And in my opinion, that would be 20 hedge funds with diversification mix across event-driven, macro, long-short, and a couple other strategies.
And so what we're showing here is there's actually a really good amount of correlation benefit that comes from doing this exercise because when you start throwing all those hedge funds together, the returns go down a little bit, but the volatility collapses. And the way this chart works is each blue dot is an individual hedge fund, one of 700 that we looked at. And then the gold dots are hedge fund composites with 20 hedge funds each, again where we force them to have a certain number of each of the styles so you're not just owning 20 event-driven funds. And you can see how the hedge fund cluster of hedge fund composites really collapses volatility compared to the blue cluster of the individual funds.
So
(DESCRIPTION)
Slide. A scatter plot displays hedge fund composites of 20 funds as brown dots charted by annualized excess return versus T-bills and annualized excess return volatility from September 2020 to August 2025. A single black square marks the 70% S&P 500, 30% Barclays Agg 1–5 benchmark point. Text: Source: JPMAM, HFR, August 2025.
(SPEECH)
how would you feel if you owned that gold cluster? If all you cared about was absolute return, you'd be disappointed because if we just take a 70/30 mix of stocks and bonds using the S&P 500 and a Barclays ag benchmark that, over the last five years, that would have returned around 8% And most of the gold dots returned less than that.
But look at that black dot on this chart. The volatility of the 70/30 mix was almost 12%, which was higher than almost all of the hedge fund composites. So
(DESCRIPTION)
A scatter plot shows hedge fund composites of 20 funds as brown dots charted by annualized excess return versus T-bills and annualized excess return volatility from September 2020 to August 2025. Black squares labeled 30/70, 40/60, 50/50, 60/40, 70/30, and 80/20 represent S&P 500 and Bloomberg US Aggregate 1–5 benchmarks. Text: Selective reporting and survivorship bias. Text: Source: JPMAM, HFR, August 2025.
(SPEECH)
what we did is we said, all right, let's create a bunch of different stock/bond composites with different amounts of stocks and bonds and then match each hedge fund to a benchmark that has the same volatility.
And I started doing this a few years ago when one of the large California ERISA plans-- it was a state plan-- they got a lot of press because they said, we're really disappointed in our hedge fund platform. We've terminated the managers. We're shutting it down. The returns were way too low, and we were disappointed. And when I looked at all the constraints that they had put on their managers, they basically gave them a fixed income level of risk. So of course, they were delivering fixed income levels of returns.
And so ever since then, I've been kind of committed to this mission of making sure that we look at hedge fund performance on a risk-adjusted basis relative to the risks that are being generated by these managers. And so as you can see here, the vast majority of hedge fund composites outperformed their volatility-adjusted benchmarks. So I consider that to be good news.
We
(DESCRIPTION)
Slide: Crowding. Four charts appear. The top left chart shows the weight of the Magnificent 7 stocks as a percent of the hedge fund long portfolio compared with their weight in the Russell 3000 and in the US equity portfolio, with the hedge fund weight rising sharply through 2024–2025. The top right chart shows hedge fund portfolio density, with the weight of the top 10 positions in the median hedge fund long portfolio trending upward toward about 70 percent. The bottom left chart shows a hedge fund crowding index, with crowding increasing over time. The bottom right chart shows the share of hedge fund stock holdings in the top quintile of beta, rising noticeably above a dashed random-draw line. Text: Sources include GS Global Investment Research, Q3 2025, and Empirical Research, November 5, 2025.
(SPEECH)
get into other issues related to selective reporting and survivorship bias, asset crowding. I mean the share of hedge fund stock holdings in the top quintile of beta is very high, which, again, is a jargony way of saying they own a lot of the big hyperscaler momentum stocks. The hedge fund crowding index is at the highest level that it's been at for the last 20 years or so. So there's a lot of momentum risk in the hedge fund community.
So
(DESCRIPTION)
Slide: Private credit, a lot to talk about. A stacked area chart shows US risky credit market shares from 2000 to 2025 as a percent of US GDP, divided into bank loans to small and medium enterprises, high yield bonds, leveraged loans, and private credit. The total share rises steadily over time, with private credit expanding the most in recent years. Text: Source: Bridgewater, October 23, 2025.
(SPEECH)
private credit. There's a lot to talk about. Somebody said something about cockroaches. And it reminded me I went to see a play with my brother-in-law at one point where he was dressed as a bug. But that's beside the point.
What's interesting to me is that, during all of this discussion about private credit, it's not like risky private credit is being added on top of other kinds of risk. It's actually displacing both leveraged loans and high yield bonds. So if you look at the total share of credit as a share of GDP in the US, the amount's been roughly the same over the last few years. It's just that private credit's been taking share from other stuff.
Now,
(DESCRIPTION)
Slide: Private Credit Performance. A grouped bar chart compares annual total returns from 2014 to 2025 for a private credit composite and the JPM Leveraged Loan Index, with the private credit bars shown in blue and the leveraged loan bars in gold. Returns vary year to year, with private credit often showing higher peaks such as in 2021, and an Ann. set of bars appears at the end summarizing annualized performance. Text: Source: Bloomberg, Cliffwater, Preqin, PitchBook, JPMAM, 2025.
(SPEECH)
where are we? Looking backwards. OK, there's backwards and forwards. Let's do them separately. Looking backwards, the performance of private credit versus leveraged loans has been pretty good. And so again, this is as much art as science. But we put together a composite that looks at three different private credit benchmarks. We combine them into one. And then we compare that to leveraged loans. And the composite has been delivering around 300 basis points a year in excess of leveraged loans, which it should in an economic expansion.
So given the concentrated nature of private credit portfolios, given the lower credit quality of the companies, at least as the way the rating agencies look at it, you should be compensated for that. And during an expansion, 300 basis points sounds about right.
And
(DESCRIPTION)
Slide: Private credit, limited signs of stress. Four charts appear. The top left chart shows the ratio of downgrades to upgrades for private credit and broadly syndicated loans from 2015 to 2025, with private credit spiking sharply around 2023 before declining. The top right chart shows default rates with and without selective defaults for private credit, leveraged loans, and broadly syndicated loans from 2017 to 2024, with all categories rising around 2020 and again in 2023. The bottom left chart shows nonperforming loans as a share of total loans for private credit, broadly syndicated loans, and middle market loans, with levels fluctuating between 0.4 percent and 1.4 percent. The bottom right chart shows PIK interest share of gross investment income for publicly traded BDCs and non traded perpetual BDCs in 2023, 2024, and Q2 2025. Text cites S&P Global Ratings and JPMAM.
(SPEECH)
as we look at things right now, there are limited signs of stress in the private credit world. There are some signs of stress, but they're so far limited. The ratio of upgrades to downgrades looks pretty good. Default rates are very low, even when you include selective defaults, which are defaults that would have happened but got some kind of reorganization plan to avoid it.
Non-performing exposures, a share of total loans, whether we're looking at middle market loans or the broadly syndicated loans, are still below 1%. And this issue gets a lot of press. The PIK share of the gross investment income is barely moving.
So what does that mean? PIK Is Payment In Kind. In the private credit markets, most of you are all aware that if companies can't come up with the cash, they can PIK the payment, which means I'm not paying you. Just add it to my balance at the end and we'll assume we're all good. And so on an accrual basis, there's money. But on a cash basis, there isn't. When we look at both publicly traded BDCs and non-traded BDCs, the PIK share of gross investment income hasn't really moved since 2023 very much. So that tells us that there's not been a marked deterioration there.
But
(DESCRIPTION)
Slide: Does this still apply? Two charts appear. The left chart compares caps on EBITDA add-backs across private credit loans under $250MM, private credit loans over $250MM, and syndicated loans, with bars showing how restrictive each category is at 20%, 25%, 30%, 35%, and no cap levels. The right chart compares the percent of deals allowing covenant features such as allowing inside maturities, conversion of unused restricted payments into debt, asset sale repayment of less than 100%, automatic debt increases per dollar of new equity, and no “J Crew” IP blocker, with private credit generally allowing fewer of these features. Text: Source: Moody’s Investor Services, October 2023.
(SPEECH)
this is, I think-- this next topic, I think, is the most important thing in all of the whole private credit debate discussion, which is a couple of years ago, Moody's did a really good analysis on what goes on in private credit compared to the broadly syndicated loan market. And what they found two years ago was that the managers in private credit are being much more conservative about underwriting. One way they looked at that was to say how much to-- pardon me. Let me get my Solo cup.
(DESCRIPTION)
Michael drinks from a red cup.
(SPEECH)
How much do private credit managers allow EBITDA add backs, which is this fictional thing that I can't even believe exists in the first place. But some of these private credit borrowers get to assume a certain level of synergistic EBITDA or cash flow growth in order to avoid triggering certain covenants. The broadly syndicated loan market allows that, sometimes with no cap. Much less of that is allowed in prior credit.
And then the chart on the right really digs into the stuff that I am fascinated by, which are the legal terms and conditions under which these borrower deals get done. In the broadly syndicated loan market in those blue bars, a lot of things are allowed-- inside maturities, which is all of a sudden you can borrow money at a shorter maturity than an existing loan.
If there's unrestricted covenant room in terms of restricted payment clauses, you can convert that into raising your debt ceiling. If you sell an asset, you don't have to use 100% of the proceeds to repay the debt. There's an automatic debt increase per dollar of new equity committed. And then sometimes there's no J Crew IP blocker. And for those of you who are in the legal community, you understand what that means. So whereas in private credit, very rarely only 10% or 15% of the time were any of these things allowed.
Now Moody's has not updated this analysis the last two years. As it stands, two years ago, this is what it looked like. My sense is that some of this is still true but that the underwriting characteristics in private credit have been getting easier and looser. And so this gap that existed two years ago may not be there as much as it was. We can see that with the following four exhibits.
So
(DESCRIPTION)
Slide: Private credit changing standards. Four charts appear. The top left chart shows covenant-lite share of new-issue loans for broadly syndicated loans and private credit over $500 million from 2011 to 2024, with both lines rising and the private credit line climbing sharply in recent years. The top right chart shows the percent of private credit deals without maintenance covenants by deal size, with larger deals showing higher percentages. The bottom left chart shows median BDC leverage as a debt-to-equity ratio with lines for the 75th percentile, median, and 25th percentile from 2008 to 2024. The bottom right chart shows middle market borrowers’ maintenance covenants across lower, core, and upper middle market EBITDA groups with bars for maintenance covenant headroom above 40 percent, multiple maintenance covenants, and maximum leverage threshold. Text cites S&P Global Ratings, JPMAM, and Covenant Review data.
(SPEECH)
the covenant light share of new loans, private credit rising rapidly relative to the broadly syndicated loan market. Private credit deals without maintenance covenants. The bigger the loans, the more frequently they don't have any maintenance covenants at all. The median leverage of BDCs has been going up.
And then lastly, S&P did an interesting analysis where they found that the only reason that some borrowers still have a lot of headroom with the covenants is because the debt to EBITDA ratio that was used as a covenant threshold was like 8, 10, or 12x. In other words, you haven't tripped a covenant because they were set on such a toothless basis to begin with. So the private credit standards are changing.
(DESCRIPTION)
Slide: Private credit wrapping up. A line chart labeled COVID was not a typical recession tracks HY default spreads and HY default rates from 1987 to 2025, with shaded bands marking recession periods and a vertical marker for COVID. Both lines spike during past recessions and rise only modestly during COVID. Text above explains that private credit manager performance has been tightly grouped since 2014 because the COVID recession featured much lower high yield default rates and muted credit spread increases. Text cites BofAML, Moody’s, Barclays, Bloomberg, JPMAM, November 2025.
(SPEECH)
So to wrap up, the performance of private credit managers has been pretty tightly bunched since 2014. Again, as one would expect in the absence of a lasting recession. And remember, there was a recession during COVID, but it wasn't a real recession. And I have a chart in here that explains what I mean by that. If you look at the prior three real recessions, the SNL recession, the dotcom collapse, and then the global financial crisis, high yield default rates went up a lot. High yield credit spreads went up a lot. And they went up and they stayed there for a while. During COVID, they didn't. So they spiked very temporarily. And then they came back down pretty quickly.
So the point is it's been difficult for investors to really tell 15 years out from the last real durable recession, which private credit managers have been underwriting better than others. And those are the things that we're going to see because, eventually, when the next recession does hit, you're going to see a lot more manager dispersion than we've seen in the past.
(DESCRIPTION)
Slide. Text.
(SPEECH)
We conclude the private credit section with seven questions to ask your private credit manager just so you can get an understanding of what kind of risks they're taking.
So
(DESCRIPTION)
Slide: Real Estate. A line chart shows the Open-end Diversified Core Equity real estate index from 1978 to 2025 on a log scale rising over decades with drops during the S&L crisis, the Great Financial Crisis, and a present decline. Shaded bands mark these downturns. Text: Many fundamentals are still terrible, but there are signs of life in CRE including ODCE returns, transaction volumes, debt origination, and distressed sales. Text: Source: National Council of RE Investment Fiduciaries, JPMAM, Q2 2025.
(SPEECH)
let me wrap this up. I've been talking here for a while. Real estate. So the fundamentals are still kind of bad or really bad, depending upon which corner of the real estate industry we're talking about. But there are signs of life in commercial real estate. And I'm not surprised by that because we've just lived through the third biggest correction. First, we had the SNL crisis, the great financial crisis. And when you look at the history of the declines in real estate returns, this one matches almost the two prior ones.
OK. So the more you have a correction in pricing, the more it's time to say, OK, I know there's a lot of bad news, but let me start looking for opportunity. And we're seeing signs in life in terms of investment returns bottoming out. Transaction volumes are picking up. Debt origination is picking up. Distressed sales are picking up. That's always a sign of market clearing, which is a good thing.
And
(DESCRIPTION)
Slide: Asset prices, distress and lagging indicators. A line chart titled Special servicing rates for office and non-office CMBS loans shows office loans (blue line) and ex-office loans (red line) from 2002–2026. Office special-servicing rates rise sharply after 2022, reaching roughly 15%, while ex-office loans rise more modestly toward about 7%. Source: JP Morgan CMBS Research, September 2025.
(SPEECH)
I know that a lot of people are spooked by this chart on this page, which shows the special servicing rates for office loans in the CMBS market. What does that mean? A lot of debt originators don't really have the capacity to service loans in default or modification with their traditional servicing approach, so they get shunted off to these special servicing departments, either inside or outside of the bank. And those special servicing rates are skyrocketing. They were 3% a couple of years ago. And now they're 16%.
But I would just remind you of a few things. And if you read the 20-year anniversary piece that came out over the summer, you'll remember this. During the 2008 financial crisis, bank stocks in the S&P bottomed when only 8% of the eventual bank failures had taken place. And the same thing happened during the 1990 savings and loan crisis. High yield spreads peaked.
Right. The time for investors to buy at the peak, only half of all the eventual corporate defaults in the high yield market had taken place. And there's many, many examples of this. The bottom line is that markets bottom way before the fundamentals stop deteriorating. And I have a feeling that we're in that kind of environment right now for real estate where the fundamentals are going to continue to get worse but asset prices will be bottoming.
And
(DESCRIPTION)
Slide: Office. A line chart titled MSCI commercial real estate property type allocations shows Industrial rising toward the highest share, Apartments trending upward, and Office declining sharply by 2025. Retail slopes downward while Medical Lab Life, Self storage, and Other remain small. Text: Source: MSCI, JPMAM, Q3 2025.
(SPEECH)
I don't want to dwell too much on office. As you can see from this chart, around 10 years ago, office was king. Right. Office was 35% to 40% of the average institutional commercial real estate portfolio. And that number started collapsing in 2015 and is now below 15%. So we shouldn't spend too much time on office when industrial properties, including data centers and apartments, are capturing way more capital.
But I will say the long office nightmare is gradually slowing. There's a slowing construction underway. There's a slowing delivery pipeline. Sublease availability is starting to come down. And there's a very high concentration of vacant space in just a small subset of the square footage. So the vacancy problems are very concentrated. The rest of the market is doing a little bit better.
And
(DESCRIPTION)
Slide. A bar chart compares demolitions, conversions, and new supply deliveries from 2018 to 2025 with new supply deliveries forming the tallest bars each year and a projected drop in 2025. Smaller bars show modest conversions and low demolitions until a projected rise in 2025. Text: Source: CBRE Research, CBRE Econometric Advisors, Q2 2025.
(SPEECH)
here you can see-- I thought this was amazing. And this was as of the second quarter. It finally looks like demolitions and conversions into residential in the second quarter were higher than new supply delivered. So that's really a turning point in the office narrative when office buildings converted into other things or torn down are greater than the amount of new supply. So I thought that was a notable thing to talk about.
(DESCRIPTION)
Slide: WFH. A scatter plot compares Work From Home indicators across major US cities using green dots for mass transit recovery vs 2019, red dots for a Stanford survey, yellow dots for Kastle KeyFob data, and black triangles for the average. The points cluster mostly between 50 percent and 80 percent with some outliers near 100 percent. Text: Work From Home.
(SPEECH)
And one last chart that I just wanted to share here, because I thought it was interesting. The work from home story hasn't really changed very much. It's still about 70%. So this looks at a bunch of different cities. Pardon me. The back to work is around 70%. The residual, which is work from home, is around 30%. And to get to those numbers, we look at a combination of mass transit recoveries compared to 2019, a Stanford survey by Nicholas Bloom, and then the Kastle KeyFob swipe data. And for the most part, Austin a little bit of exception at 80%, most of the rest of the cities the back to work ratio is in the neighborhood of around 70%. OK.
I think that's the last-- let's see. The last topic-- and I'm not going to spend too much time on this-- is there
(DESCRIPTION)
Slide: Retail. A stacked bar chart titled Retail fundraising for alternative investments by year shows rising totals from 2020 through an annualized 2025, with the largest portions coming from Non traded and Private Placement REITs and Non traded and Private Placement BDCs. Smaller stacked segments include Interval funds, Private equity, Infrastructure, and Other. Text: Retail includes publicly registered non traded REITs publicly registered non traded BDCs interval funds non traded preferred stock of traded REITs Delaware statutory trusts opportunity zone funds and other private placement offerings. Text: Source: Robert A Stanger & Co Alts Wire JPMAM 2025 Note: 2025 figures are annualized.
(SPEECH)
was this executive order in August that talked about let's democratize alternative assets and try to find ways that they can be added into 401(k) plans. And that was pretty notable. Got my attention.
And so I added a section on this question of democratization of alternative assets. And certainly even before anything is done about trying to figure out ways of putting them in 401(k) plans, retail fundraising has been rising pretty rapidly. A few years ago it was like $25 billion, and it's on track for 175 billion this year. So there's a lot of retail money being made, raised in privately placed REITs and BDCs, interval funds, and different kinds of private equity and infrastructure vehicles.
And
(DESCRIPTION)
Slide: "Private Equity, Public Capital and Litigation Risk": Areas of potential litigation. Text.
(SPEECH)
there was an interesting paper that came out. I agree with parts of it. I don't agree with parts of it. Some parts of it I strongly disagree with. But it's called private equity, public capital, and litigation risk. And the general argument of this paper is that if the regulators don't figure out a way to protect retail investors, the litigation process will and that the markets and actual retail litigants through the courts are going to discipline the industry even if the regulators don't. And their three strongest arguments are related to unclear, vague language as it relates to things like fiduciary duty, how net asset values are computed, all the different expenses, and whether they're reported on or not reported on.
For 40 Act vehicles, they're fine because you're subject to the same protections and requirements as a regular mutual fund. But for alternative asset vehicles sold to accredited investors, they may not be required to show all their fund level expenses and portfolio level expenses and things like that. So there is some basis for litigation here.
But there's a lot of counterarguments. And when I showed the paper to someone who, for a living at one of the big white shoe law firms in New York, writes these documents to presumably shield the alternative asset managers from retail litigation, he hit the roof. He actually jumped on a Zoom in a t-shirt because he was so anxious to talk to me about it.
But they list their counterarguments and things like that. And you can read through those, why they believe that the legal litigation risk presented in this paper were offsides. The
(DESCRIPTION)
Slide. Text: Democratization of alternatives may be accretive to retail portfolios in the long run, but if a sharp recession and liquidity crisis coincide, the documentation firewalls designed to shield alternative asset fund managers will likely be tested. Four panels depict litigation and market stress indicators. The top left panel shows rising 401k excessive fee litigation settlements with bars for the number of settlements and a line for total settlement amounts. The top right panel shows yearly counts of 401k excessive fee and performance lawsuits. The bottom left panel shows the percentage of failed auction rate security auctions in 2008 rising sharply over time. The bottom right panel shows bar charts of core federal filings across categories including Cyber-security Crypto Cannabis COVID-19 SPAC 2023 banking turbulence. Text: Sources: Encore Fiduciary 2024 Journal of Financial Economics 2019 Gibson Dunn September 2023.
(SPEECH)
only thing I would say is this whole democratization of alternatives thing, it may be accretive to portfolios in the long run. But if you get a sharp recession and a liquidity crunch at the same time, these documentation firewalls that were supposed to shield the alternative asset managers are going to be tested. And I concluded that section with the review of ongoing litigation in 401(k) plans related to excessive fees and performance issues, and then who can forget a lot of the litigation that took place both around the failed auction rate security market around 15 years ago as well as the SPAC boom and bust.
So there are plenty of examples here where people thought they were firewalled, and it turns out that they weren't as it relates to dealing with retail investors.
So that is a brief summary of what is otherwise a long slog through the alternative asset markets that we do every couple of years. And so that's that. And we will see you again on January 1 when we're going to launch our 2026 outlook, most of which is going to, no surprise here, look at what's going on with hyperscalers and pathways to profitability on $1.4 trillion of hyperscaler capital spending since GPT was launched in 2022.
Thank
(DESCRIPTION)
Title card: JP Morgan, Eye on the market. JP Morgan. Presentation. Title: December 2025, The Deep End. 2025 Alternative Investments Review. Image: A man wearing a dark suit, white shirt, and tie stands waist-deep in a swimming pool. He looks concerned as he looks to the side, with a beach, ocean waves, and a modern house with large windows behind him.
(SPEECH)
you very much for listening. And wish me luck next week. Bye.
(DESCRIPTION)
Logo: J.P. Morgan.
