August 2026 Market Update: The Return of Capital

Two things happened in August that are worth an owner's attention, and neither is the one that made the headlines.

The first is that corporate earnings got broader. Ten of the eleven S&P 500 sectors grew profits year over year in the second quarter, nine of them by double digits, and blended growth for the index came in at 52% — the strongest quarter since 2021. That is a different market from the one where five companies carried the index.

The second is that the labor market underneath it kept softening. July payrolls fell by 23,000 against a forecast of a gain, May and June were revised down by a combined 103,000, and the trailing twelve-month average is now roughly 34,000 jobs a month.

Strong profits, weakening employment. That tension is the actual state of the public market, and it is not resolved.

But most of our clients hold something the monthly commentaries do not cover, and where the more important shift is underway. In private markets — buyout, venture, credit, real estate, infrastructure and energy — the constraint has moved from what you earn to when you get it back. That is the subject of most of what follows.

The Public Market, Briefly

Equities rose. On a price-return basis the S&P 500 gained 2.6% in August, the Nasdaq 3.9%, the Dow 1.3%, leaving them up 12.3%, 13.5% and 10.7% on the year.

International outperformed. The MSCI EAFE Index returned 1.99% for the month and is up 13.81% year to date; emerging markets did better still, with the MSCI Emerging Markets Index up 3.37% in August and 24.08% on the year. Both are net returns in U.S. dollars, and a weaker dollar is part of why the emerging market figure looks the way it does to a U.S. investor. Volatility closed the month at 14.92, well under its long-run average of roughly 19.5.

Three things underneath that are worth more than the index levels.

Earnings breadth is real. The 52% blended growth figure for the second quarter is the best since 2021, and the composition matters more than the number: only health care declined. Consensus now looks for roughly 14% earnings growth in 2027 on top of it. Whether that is achievable is a fair question — but the argument that this market rests on a handful of companies is harder to make this quarter than it was last year.

Valuation is less stretched than it is usually described. The index trades near 19.6x forward earnings. That is high against a long history, but it is essentially in line with its own five- and ten-year averages of 19.9x and 19.0x. Comparisons to a 16x “historical average” are drawn from a different measure over a different window, and they make the market look more expensive than a like-for-like comparison supports. We would rather say plainly: valuations are full, not extreme, and they tell you very little about the next twelve months.

Bonds have not worked this year. The Bloomberg U.S. Aggregate is roughly flat for the month and modestly negative year-to-date. This is the part most monthly summaries skip. Yields are attractive on new money and genuinely useful for a family funding spending from a portfolio — but the index return reflects the price effect of rates that have not fallen, and 2026 has so far been another year in which core bonds did not do the job of diversifying equity risk.

The Private-Market Throughline: Distributions, Not Returns

Here is the number that explains most of what has happened in private markets over the last three years.

Private equity distributions ran at roughly 6% of assets under management over the twelve months to mid-2025, against a 16% average across 2015–2019 (McKinsey). Bain measures it against net asset value instead and gets to the same place: distributions have been stuck below 15% of beginning NAV for four consecutive years, against a historical norm closer to 20–25%.

However you cut it, capital is coming back at roughly a third to a half of the normal pace, and it has been doing so for four years.

The reason matters more than the number, and it is largely structural. Companies are staying private far longer and getting far larger before they list. The median age of a company at its initial public offering has moved from nine years in the 1990s to twelve to fourteen years today. The count of U.S. listed companies has fallen roughly 40% from its 1997 peak. Private equity holding periods have stretched from a traditional three-to-five years to about seven.

It is the predictable arithmetic of a market where the value creation that used to happen in public markets now happens before the listing.

One place to be careful, because the confident version of this claim is contested. You will read that private equity investors are now net cash-flow negative — writing more into funds than they receive back. The data providers do not agree. McKinsey found that in 2024, sponsor distributions exceeded contributions for the first time since 2015, and Bain puts net cash flow modestly above breakeven. The Institute for Private Capital and MSCI, aggregating differently, still show investors calling capital on net. We are not going to referee that.

What is not in dispute is the timing. A commitment made in 2021 is receiving capital back far behind schedule, and that pushes the whole plan — the liquidity, the reinvestment, the estate work built around it — several years to the right.

Venture is the clearer case. Venture funds ran net cash outflows of roughly $46 billion through the first three quarters of 2025.

Once you see the distribution problem, the rest of the private-market news stops looking like a list of unrelated developments and starts looking like a set of responses to one constraint.

Private Equity: The Exit Market Split in Two

Buyout deal value fell in the second quarter — down roughly 24% year over year, with software transactions down far more. That is a slowdown, not a freeze, and entry multiples remain near record levels at close to 12x EBITDA.

The exit picture is more interesting, because it did not simply get worse. It bifurcated — and the first thing to say is that the total shrank. Total private equity exit value fell 46% quarter over quarter, to roughly $102.6 billion. Within that smaller total, initial public offerings went from about 10% of exit value to roughly 31%, corporate sales fell 63%, and sponsor-to-sponsor sales — one private equity firm selling to another — fell 57%, to their lowest quarterly level in at least a decade.

Note what that means: the IPO share rose partly because the denominator collapsed.

The June listing of the largest private company in the United States raised $75 billion in its base offering against reported demand of more than $250 billion. The overallotment was exercised within four days, taking the total to $85.7 billion. For scale, the previous largest initial public offering in history raised $25.6 billion, or $29.4 billion including its overallotment. This deal was roughly three times the prior record. Roughly 30% of it was allocated to retail investors through brokerage platforms — extraordinary for an offering of that size.

The overallotment detail is the one that matters. Underwriters exercise it when demand is genuinely there and the price is holding without support; it is objective evidence that the capital was absorbed rather than merely allocated.

Year to date, U.S. issuers have raised roughly $145.8 billion across 106 offerings — more than the previous full-year record, with four months still to run. And the depth is not one deal: the second quarter was the largest for proceeds since 2021 even excluding that listing, with nine other companies raising a billion dollars or more.

The relevance to the backlog is direct. There are 32,000 unsold private companies and some of them are very large. The constraint on clearing them was never supposed to be whether public markets could write a check big enough. This year settled that question. Two of the largest private companies in the world have now confidentially filed.

One qualification belongs with it: the demonstrated capacity is concentrated in a few themes, principally artificial intelligence and defense. A market that can absorb an $85 billion space and satellite listing has not necessarily proven it will absorb a mid-market industrial roll-up at an acceptable price. The channel is open. It is not open uniformly.

Meanwhile the backlog keeps building. There are roughly 32,000 unsold private-equity-owned companies globally, worth about $3.8 trillion. More than half of buyout inventory has been held longer than four years, against a historical norm closer to 42%. Average holding periods have stretched to about 6.6 years. Of the companies bought in 2021, only about 19% had been sold by 2025 — where roughly 30% would have been sold by year four historically.

The pressure valve is the secondary market, and it is now a real one. First-half 2026 secondary volume set a record — Evercore counts roughly $121 billion, Lazard roughly $124 billion — with high-quality buyout interests pricing around 90% of net asset value.

Three caveats belong with that number, because it gets quoted without them. The ~90% is an average for a selected subset — high-quality buyout interests, in an institutional market — and pricing does not transfer across fund types or quality tiers. The reference net asset value is typically a lagged quarter-end mark, so the effective discount to current value is not the same as the headline. And most individually held interests require the general partner's consent to transfer, carry minimum sizes, and involve intermediation costs. It is a real option worth pricing. It is not a posted bid.

Continuation vehicles — where a manager sells a company to a new fund it also manages — have gone from a workaround to a standing feature. Estimates of their share of exits vary with the measure: roughly 14% of exit value in 2025 on one count, under 10% of total exit value on another. What is less ambiguous is adoption — nearly three-quarters of the largest global private equity firms have now done at least one.

One consequence worth naming: fundraising has become a realized-distributions market rather than a reported-return market. Buyout fundraising fell roughly 16% in 2025 to about $395 billion, closings are taking materially longer than they did three years ago, and capital is concentrating in the largest managers. The funds closing quickly and above target are, with striking consistency, the ones that have sent money back.

Where This Shows Up in How We Build Portfolios

The duration problem described above is one of the reasons we use semi-liquid, open-ended private equity structures alongside traditional drawdown funds for clients where private equity is appropriate. They address a specific set of frictions: capital is deployed at subscription rather than called unpredictably over years, so there is no unfunded commitment to reserve cash against and no J-curve; a single subscription buys diversification across vintages and strategies; the position can actually be sized and rebalanced; and investors receive a 1099 rather than waiting on partnership tax reporting.

On the repurchase cap, which gets written about backwards. These vehicles typically offer to repurchase shares quarterly, limited to around 5% of net assets. That limit is usually described as a restriction on investors. It is better understood as a protection for the ones who stay.

The reasoning is not ours; it is the SEC's. In adopting its fund liquidity rules, the Commission observed that a fund forced to sell illiquid holdings on short notice may receive less than carrying value, which would “result in a preference in favor of the redeeming shareholders and a diminution of the NAV per share of shareholders who have not redeemed.” It also noted that a fund meeting redemptions by selling its most liquid assets first “may leave remaining shareholders in a potentially less liquid and riskier fund.” The Financial Stability Board made the same point about private credit in May, warning that stale valuations “may create a first-mover incentive during stress events.”

Worth being precise about the number, because it is widely misunderstood: for registered interval funds, 5% is a regulatory minimum — the rule requires periodic offers of between 5% and 25%, and requires the fund to hold liquid assets equal to 100% of the offer. Most evergreen private equity vehicles are not interval funds; they are tender offer funds, where the limit is contractual and set by the board rather than mandated. The disclosed, contractual limit is the thing to look for, and its absence is disqualifying.

How we actually use the feature is the part that matters for the duration problem. We do not generally submit redemptions during periods of stress — that is when proration is most likely, and it is usually the worst moment to sell anything. We use the liquidity during periods of strong inflows, when repurchases are funded out of new subscriptions rather than asset sales. There is no forced selling, no dilution to remaining holders, and no waiting on an underlying realization. Used that way, the structure returns capital on a schedule the fund controls rather than one the exit market dictates, and it genuinely helps with the distribution shortfall described above.

It is not a substitute for liquidity in the portfolio, and we would not present it as one. The limits are real and they bind. In the first quarter of 2026 one large non-traded credit vehicle received requests equal to 7.9% of shares against a 5% limit; the sponsor upsized the offer to 7% and committed firm and employee capital to cover the remainder. Another manager halted repurchases in one vehicle entirely. As the head of that first strategy put it publicly: investors should never buy these products expecting full liquidity.

And there is a fair criticism to sit with. The International Monetary Fund has pointed out that these tools “have not been tested in a severe runoff scenario,” and that redemption pressure has already “forced certain large private credit fund managers to allow redemptions above the established limits.” A limit that a sponsor exceeds under pressure is arguably a reputational commitment rather than a structural safeguard. That is a reasonable objection, and it is one more reason the manager matters more than the wrapper.

The rest of the trade is straightforward. Cash held for liquidity is a drag on returns, there is a fee layer, and the periodic net asset value at which investors subscribe and redeem is an estimate rather than a market price. Published comparisons show these vehicles tend to lag traditional funds in strong markets and hold up better in weak ones — a smoothing characteristic, not a performance advantage. Money needed on a schedule does not belong in either structure.

Venture Capital: One Market, or Two

Venture dollars are at records. U.S. deal value in the first half of 2026 was roughly $413 billion, exceeding all of 2025 — while deal count ran at roughly half the pace.

The concentration behind that is difficult to overstate. Artificial intelligence took roughly 86% of U.S. venture dollars in the first half. Deals of $100 million or more accounted for about 87.5% of value. Two companies alone — the two largest AI labs — took roughly 43% of global venture funding. In the first quarter, the top five deals were more than 70% of all deal value, the highest concentration ever recorded.

The practical translation for someone holding a venture fund: “venture capital” no longer describes one market. A manager who is not in the AI complex is competing for the remaining 14% of dollars, in a market where deal counts are falling.

On the exit side, the same bifurcation as buyout. Headline exit value looks spectacular, but it is dominated by a single enormous listing, and market capitalization at listing is not cash in an investor's account. The more useful numbers: the median IPO stepped up just 1.1x over the last private round — companies are listing at roughly their last private mark, not above it — and roughly 87% of acquisitions had undisclosed valuations, which usually indicates a markdown.

On performance, one figure is worth carrying. Across fund vintages from 2017 through 2024, every year except one now shows top-decile net returns above 20%. The exception is 2021. And for the 2017–2018 vintages — funds now eight and nine years old — fewer than 20% have returned the capital originally committed.

Private Credit: The Question Is the Marks

Private credit raised roughly $190 billion in the first half of 2026, up more than half from a year earlier. It is doing so into a deteriorating credit environment, and that tension is the story.

Spreads have stopped compressing. After several years of steady tightening, direct lending spreads widened for the first time in three years in 2026 — one lender survey puts first-lien contractual spreads at 489 basis points in March and 510 in June. New deal volume fell roughly a quarter between the first and second quarters, from about $26 billion to $20 billion in U.S. middle-market direct lending, and the broadly syndicated loan market took back the majority of leveraged buyout financing — reversing several years of private credit share gains.

Default rates depend entirely on who is counting, and the range is wide enough to be its own finding: from about 2.5% on the narrowest definition — payment and covenant defaults on senior secured loans — to a record 6% on the broadest, which includes distressed debt exchanges and maturity extensions. Moody's, working in between, estimates 2025 defaults somewhere between 1.6% and 4.7% depending on definition, and notes that roughly 65% of what it records are distressed restructurings rather than missed payments — which means the rate of borrowers actually failing to pay is lower than any of the headline numbers.

The honest summary is that there is no single private credit default rate, and anyone quoting one without saying which definition they are using is telling you less than it sounds like.

The early-warning indicator is payment-in-kind interest — where a borrower, rather than paying cash, adds the interest to the loan balance. The share of business development company loans paying in kind rose from roughly 6% in 2022 to about 10% in early 2026, according to a Federal Reserve Bank of Boston research paper published August 5. One of its authors, discussing the finding in the press, called it “a sign of stress.” The more telling detail: roughly 56% of loans currently paying in kind were cash-paying when they were originated. That is not a structuring choice; that is a borrower that stopped being able to pay.

Here the sources genuinely disagree, and the disagreement resolves cleanly. Some industry data shows payment-in-kind falling — new loans with such provisions dropped from 25% at the end of 2025 to about 13.5%. Both are true. Lenders are tightening on new loans while the stock of it inside existing portfolios keeps climbing. The stock is what investors own.

And then there is the marking question, which is the one that matters most.

Listed business development companies — vehicles that hold essentially the same assets as their non-traded counterparts — report non-accruals averaging around 2.1%. Non-traded vehicles report around 0.6%, with none above 1%. Those are not different credit portfolios. They are the same kind of credit, marked differently.

Public research this year documented several cases of different lenders marking the same loan to the same borrower at materially different values on the same date — in one instance at 50, 53 and 77 cents on the dollar. Ninety days earlier, the same three lenders had been within four points of each other.

That is not the typical case, and the same research says so: roughly 85% of loans were marked above 97 cents on the dollar, and fewer than 4% below 80. Marks across the market are broadly consistent. Dispersion is concentrated in stressed credits — which is exactly where it matters, and exactly where it is hardest to see from a quarterly statement.

The listed market has already rendered its verdict: listed business development companies traded at an average of roughly 0.81x net asset value as of August 27, having returned about −11% over the trailing twelve months.

The live issue for a family holding these is liquidity terms, not credit. In the first quarter of 2026, repurchase requests at large non-traded credit vehicles averaged roughly 12% of net assets, against limits typically set at 5% per quarter. About 53% of the roughly $15 billion requested was honored — several sponsors paid out above their stated limits.

As discussed above, the limit doing its job is the design rather than the failure — it is what prevents a vehicle from liquidating assets into a weak market and handing the proceeds to whoever asked first. But the practical consequence for an investor is the same either way: a full exit now takes multiple quarters, and requests rise exactly when selling is least attractive. That is a conversation worth having while it is still hypothetical, and the answer is usually to plan liquidity elsewhere rather than to plan on redeeming here.

Manager Selection Now Matters Far More Than It Did

For most of the last decade, direct lending was close to a beta trade. Spreads were wide, defaults were minimal, and the dispersion between a good manager and a mediocre one was narrow enough that the asset class carried the return. That is no longer the case. Roughly one in five borrowers in one large sample has fixed-charge coverage below 1.0x — meaning the business does not generate enough to cover its fixed obligations. More than half of loans currently paying interest in kind were paying cash when they were written. Amendment activity is running above 875 in a trailing year, with a quarter of those involving sponsor equity infusions.

Every one of those is a manager-level outcome, not a market-level one. Who underwrote the credit, what covenant package they insisted on, whether the sponsor behind the borrower has the capital and the willingness to support it, and how honestly the manager marks what it owns — those questions now separate outcomes far more than the spread at origination does.

And on the opportunity side, a distinction worth making carefully. There has been real dislocation, but it is narrower than the coverage suggests. Public credit itself is not cheap. High yield spreads sat near 265 basis points at the start of September — roughly 100 to 180 basis points inside their five-, ten- and fifteen-year averages. Nobody is being paid unusually well to take corporate credit risk in the public market right now.

What has repriced is the listed equity of the vehicles that hold private credit. Business development companies trade at a median of about 0.75x net asset value, with the sector average at 0.81x — discounts comparable to 2008–09 and to early 2020. Part of that is structural: an index-inclusion rule change in 2014 pushed institutional owners out of the sector, and capital has since migrated toward non-traded equivalents, thinning the natural buyer base.

That is a genuinely interesting place to look, and it requires more care than usual, not less. A 25% discount to stated net asset value is either an opportunity or an accurate forecast. The bull case is arithmetic: a discount that wide implies portfolio default rates well beyond anything the middle market has produced historically. The bear case is that stated net asset value is the number in question — that marks lag, that recoveries are running nearer 50 cents than the 70 commonly assumed, that software concentration is understated, and that dividend coverage is thin at a meaningful share of rated vehicles.

Both cases cannot be tested from outside. The discount is too wide to explain with observable credit metrics; the bear case is that the observable metrics are the unreliable part. What follows from that is not that the opportunity is fake — it is that what you buy and at what price carries far more of the outcome here than the decision to be in the sector at all. That is a security-selection problem, and it should be treated as one.

Real Estate: The Risk Has Inverted

Private real estate is the one sleeve where the news genuinely improved, and where the location of the risk has moved.

Values have turned. The main open-end core index returned about 1.5% gross of fees — 1.3% net — in the second quarter, with appreciation positive and accelerating across the first and second quarters. Trailing four-quarter returns are the best since 2022. On a REIT-implied basis, values were up roughly 4% year over year as of a June reading and sit about 14% below the 2022 peak.

But the recovery is wildly uneven by sector. Retail has essentially fully recovered — strip centers are within 2% of their 2022 peak. Industrial is about 11% below. Apartments are roughly 19% below, self-storage 22%. Office remains about a third below its peak, even though it is now appreciating as fast as anything else.

Fundamentals support it. Office vacancy fell at the fastest rate since 2015, with a construction pipeline at its lowest since records began in 1990 and nine consecutive quarters of positive absorption. Industrial demand outpaced new supply for the first time since 2022. Apartment vacancy fell to 4.3%, below its long-run average.

The stress moved into the debt stack. Commercial mortgage-backed securities delinquencies rose 51 basis points in July to roughly 7.9% — with multifamily at 7.7%, a multi-year high, despite those improving fundamentals, and industrial at just 1.1%. Roughly $875 billion of commercial and multifamily mortgage debt matures in 2026, about 17% of the total outstanding. Loans being written now carry rates around 6.24% against roughly 4.76% on the debt they are replacing. On our own arithmetic, and holding income and value constant, that 148 basis point step-up reduces the loan a given property can support by something on the order of a fifth.

Multifamily is the clearest example of the inversion: the best operating fundamentals in years and the worst delinquency rate in years, at the same time. That is not a demand problem. It is 2021–22 floating-rate debt meeting 2026 interest rates.

Infrastructure and Energy: Electrons, Not Molecules

These two sleeves have inverted their usual relationship, and it is worth understanding why.

Infrastructure fundraising is at a multi-year low — roughly $41 billion in the first half of 2026, against about $72 billion in the same period of 2024 — while deployment runs at records. This is not a capital overhang. It is the opposite: money is going out faster than it is coming in.

The reason looks like the same distribution problem visible in buyout. The infrastructure funds that closed above target in 2026 — including the year's largest energy infrastructure fund, which closed more than 50% above its target — did so, by their own managers' accounts, on the strength of capital actually returned rather than reported marks.

The demand side of the infrastructure story is not in question, which is exactly why it is not where the return is. On JLL's count, North American data center absorption in the first half roughly doubled year over year, vacancy sits near 1%, and about 95% of capacity under construction is already committed. CBRE, measuring a different universe, has absorption up 12%, vacancy at 1.4% and preleasing above 80% — a less dramatic version of the same shortage. The four largest hyperscalers guided to something on the order of $720–745 billion of capital spending in 2026, up from roughly $570 billion in 2025.

That demand is fully known and, increasingly, funded with debt and equity issuance rather than operating cash flow. One of those companies took its debt from roughly $16 billion to roughly $100 billion in a year and raised $80 billion of equity; another suspended share buybacks to fund it.

The constraints are where the pricing power sits, and they are physical.

August produced the first real evidence that the binding constraint on this trade may be political rather than physical. Texas paused new data center grid interconnections on August 3 pending a state audit of energy and water use. Analysts put roughly 50 gigawatts of load — about a fifth of the U.S. pipeline — at risk of delay, and the Energy Information Administration cut its Texas 2027 load-growth forecast from 14% to 5.6% as a direct result. The Texas legislature reconvenes in January.

On energy specifically, the gap between today's price and the projected one is the whole argument. Brent crude ended August above $93 on supply disruption and traded near $95 in early September. But the Energy Information Administration's August outlook projects Brent averaging roughly $87 for 2026 and $69 in 2027.

That is an agency forecast, not a market price — the futures curve is a separate question and worth checking before anyone concludes what “the market” expects. But it is the U.S. government's own working assumption that a meaningful part of today's price is a disruption premium that does not persist.

Upstream deal activity tells the same story from the other direction. Second-quarter U.S. upstream mergers and acquisitions totaled about $9.1 billion, the third-lowest quarter since 2020 — and the single largest component of it was a federal lease auction, not a corporate transaction. When the biggest seller of oil and gas assets in a quarter is the government, the private market has largely run out of things to buy.

Natural gas is the quieter version of the same theme. Henry Hub is forecast essentially flat near $3.44 this year and $3.31 next, with export volumes rising and gas-fired generation demand climbing.

Those two cash-flow profiles have different characters, and the difference is worth naming rather than ranking. Oil at today's price carries a geopolitical premium that the government's own forecast does not expect to hold. Gas offers a lower, steadier price against rising volume. Neither is a forecast on our part, and neither is a recommendation — but they are not the same kind of exposure, and a portfolio that holds “energy” may hold either.

The Questions This Raises for a Plan

These are questions, not answers. What any of them means for a particular family depends entirely on that family's holdings, timeline, and plan, and none of it is a recommendation to buy, sell, or hold anything.

On liquidity planning. If a spending, tax, or gifting plan was built assuming distributions from 2019–2021 vintage private funds would arrive on a historical schedule, is that assumption still the right one? The gap is not small and it is now four years old. Is the cash plan anchored to what has actually been distributed, or to what a capital account statement says the position is worth?

On the secondary market. Quality buyout interests pricing near 90% of net asset value is worth understanding rather than assuming. What would a specific interest actually fetch, what would it cost to transact, and does the general partner permit the transfer at all?

On structure. For a family adding private equity exposure now rather than holding an existing commitment, the drawdown-versus-evergreen question is a real one with real trade-offs on both sides — deployment speed and tax reporting against capped liquidity, cash drag, and estimated pricing. Which set of trade-offs fits depends on the plan, not on which structure is currently easier to buy.

On non-traded and evergreen vehicles. For anyone holding one: is there a disclosed, contractual repurchase limit at all, and how does proration work? The presence of a limit is a feature, and its absence would be the thing to worry about. The separate question is what a full exit realistically takes at current request levels — which is a term-sheet question, and much easier to ask before it is urgent.

On evaluating managers. In an environment where marks on stressed assets visibly differ between holders of the same asset, how much weight should a reported internal rate of return on unrealized positions carry relative to capital actually returned? And in private credit specifically — where the spread at origination now explains far less of the outcome than underwriting, covenants, and sponsor support do — how much is actually known about the manager rather than the asset class?

On concentration. A venture allocation made before 2023 is now, in practice, heavily exposed to whether that manager has access to a very small number of companies. That may be entirely acceptable. Is it deliberate?

If you hold private assets and want to work through any of this against an actual plan, we're glad to talk it through.

The Bottom Line

Public markets in August gave you broad earnings growth on a softening employment base, at full but not extreme valuations, with bonds still not doing their historical job.

Private markets gave you something more consequential: a fourth consecutive year in which capital came back at roughly a third to a half of the normal pace, and a set of structures — secondaries, continuation vehicles, evergreen funds — that have grown up around it.

The right response to that is not to conclude private markets are broken. Much of the slowdown is duration rather than distress: companies stay private roughly twice as long as they did in the 1990s, holding periods have stretched accordingly, and distributions are arithmetic downstream of that. And the exit channel itself is not the constraint — public markets absorbed the largest listing in history this year, at three times the prior record, with the overallotment taken up inside a week. Realized losses in direct lending are running at about 0.5% in the first half of 2026, roughly half the long-term average. Real estate operating fundamentals are the best since 2022. Most private credit marks are consistent between holders.

Plan around the timing rather than the marks. Four years of distributions at a third to a half of the normal pace is a planning fact, and it is knowable today. Whether any particular mark is right is not.