← All insights

The deal

On 16 December 2025, HarbourVest Global Private Equity Limited (HVPE), a FTSE 250 investment company, announced binding terms to sell a curated portfolio of five HarbourVest fund positions — one co-investment fund and four fund-of-funds vehicles, all focused on Buyout — for net proceeds of $300 million. The company disclosed the price outright: 94% of the combined funds' NAV as at 30 June 2025, a blended 6% discount. The board's own language on timing is the detail that matters most for this comparison: the assets were "exited at a relatively early stage in their life cycle."

Buyer and exact fund vintages weren't disclosed, and "relatively early stage" isn't a precise age. We're not treating this as an exact match to any one age — it's a real, disclosed clearing price for young Buyout fund interests, which is the closest thing to ground truth a public deal like this offers.

How it lines up against the published survey

HVPE's own disclosed price and Jefferies' independently surveyed Buyout dealer-desk average for the same half both land at almost exactly the same number — one real transaction and one blind market survey agreeing to a tenth of a point is a stronger corroboration of the survey data than we could get from the survey alone.

How it lines up against our own models

We can't reproduce HVPE's exact five funds — their identities, vintages, and cashflow histories aren't public. What we can do is price our own representative Buyout snapshot at the young ages "relatively early stage" points to, the same run behind our age-by-age survey comparison, and see where the real deal falls relative to it — both against each model's fair-price point, and against its full P25–P75 Monte-Carlo band.

Each cell is the fair price, with the model's own P25–P75 band in parentheses. HVPE's actual 94% sale price falls inside every single one of these bands, at every age and every model — including Equisect Cohort, whose own fair-price point (62-68%) misses by 26 to 32 points.

What a lower required return implies about this deal

15% is our platform default, not HVPE's actual hurdle — we don't know what return the buyer underwrote this at. So we re-ran the same snapshot at 10% required return, to see how much of the gap above is really just a required-return assumption rather than a modeling disagreement.

At 10%, Equisect Bayesian's age-6 fair price is 93% — within a point of HVPE's actual 94%. Read literally, that suggests HVPE's buyer priced this deal closer to a 10% than a 15% hurdle, which isn't an unreasonable required return for a mature, cash-generative Buyout position — exactly the kind of asset the board said these were. TA, which already ran close at 15%, now overshoots at every age; Equisect Cohort closes most of its gap but is still the cheapest of the three either way.

What this one deal actually shows

The TA baseline's age-5 fair price sits within a point of a real, disclosed $300 million transaction — a genuinely good result for a transparent, auditable curve model, and consistent with how close it ran to the published survey average in the comparison this deal is drawn from. Equisect Bayesian — the one that actually passes our overfitting battery and the one we validate against held-out history — priced the same young-Buyout segment 11 to 22 points cheaper than what a real seller actually got paid. Equisect Cohort, the cleanest survivor of the three on paper, misses by the widest margin of all: 26 to 32 points below HVPE's actual price. On point estimates alone, that reads like a problem. On the full band, it doesn't: every model's P25–P75 range — Equisect Cohort included — comfortably contains the real price HVPE actually got paid. One transaction is one data point, not a verdict on any model; but it's the same direction and a similar magnitude as the gap we already found against the broader published survey, which makes it a second, independent confirmation rather than a coincidence — and a reminder of exactly why we ship a band next to every point estimate instead of just the point estimate.

Why we're publishing the version that doesn't flatter us

It would be easy to only publish the model that lines up well. We ship Equisect Bayesian as the validated price specifically because it passes the statistical test that matters most — surviving out-of-sample, where TA doesn't. That doesn't mean it always lands closest to what a real buyer happened to pay in one disclosed deal, and pretending otherwise would defeat the point of publishing any of this in the first place. A fair price built on a required-return hurdle and one seller's actual clearing price in one disclosed deal are two different questions; both are worth knowing, and knowing where they disagree is more useful than only knowing where they agree.

Research and software, not investment advice. Deal details sourced from HVPE's official 16 December 2025 announcement ("HarbourVest Global Private Equity Limited Announces Asset Sale"); survey figures are third-party published market data (Jefferies Global Secondary Market Review H1 2025; Campbell Lutyens Secondary Market Overview FY 2024), reproduced here for comparison, not verified independently by Equisect. No underlying fund identities, vintages, or ages were disclosed by HVPE — comparisons here are to a representative snapshot, not the specific funds sold.