To get an early read on a manager, I start with redemption disclosures. That is where structural stress shows up in arithmetic before it shows up in a narrative.
Partners Group gave me a clean worked example this year. In the second quarter of 2026, redemption requests in its Global Value SICAV reached about 9.8% of net asset value against a 5% quarterly liquidity limit, which tripped the gate on 3 June 20261.
The number that made me look harder was not the breach itself. It was the concentration: Partners Group said 79% of first-half 2026 outflows came from just three mature evergreen strategies2.
Broad-based selling looks like a market mood. Outflows bunched in a handful of ageing vehicles look like something in the product design is being tested.
38% · of Global Value SICAV’s Q2 2026 redemption requests were deferred beyond the fund’s 5% quarterly cap (Partners Group)
The mechanics behind that figure: the gate let investors redeem 62% of what they asked for, and held back the other 38%3. What happens to that 38% is the whole story.
Here is the part that most coverage skated over. In the SICAV, deferring a redemption does not cancel it.
A July 2026 shareholder notice, distributed via the platform Endowus, spelled it out: investors could take the withheld portion at the next dealing day, or elect to remain invested45. The gate slows the exit, leaving the decision with the investor.
This was not a single June event either. By a later July 2026 cycle, the fund deferred 22.8% of requests in old share classes and 54.0% in most new share classes678.
So the answer to how many consecutive cycles the SICAV has now gated: at least March, the June quarter, and July, with the withheld share worsening materially for the newer classes. I found no Partners Group schedule laying out every subsequent dealing date with its own payout ratio, so I will not pretend the full series is public – it is not6.
I owe readers the messy bit. An earlier July 2026 notice put the new-share-class deferral at 96.8%, not 54.0%9.
Those two figures do not reconcile, and I found no statement in the record that explains the gap between them. I am flagging it as an open inconsistency rather than choosing the number that reads better.
Three other mature evergreen funds, about $9.7bn combined, were guided to Q2 2026 requests of 3.5% to 5% – sitting right below or on the gate line1011.
That is exactly the cohort I would want to track into the next cycle. But I found no Partners Group filing confirming whether any of those funds crossed the 5% gate afterwards, and no payout-ratio breakdown for that band beyond the request percentage itself. Both questions stay open in the public record.
Now cross the Atlantic, where the plumbing is different. Partners Group’s $16bn Delaware-domiciled evergreen vehicle drew May 2026 tender requests of about 6% of NAV against its own 5% tender-offer threshold121314.
A tender offer does not defer the excess to a future dealing day. It caps what gets accepted in this window and settles the rest by whatever the structure dictates.
The sharpest illustration is a separate US gateway vehicle, GVF. It was expected to pay about 62% of a 30 March 2026 redemption request – while a 29 April 2026 cut-off could receive 0%15.
That is the difference in one line. A tender structure can produce complete non-payment for a given window, extinguishing the claim rather than deferring it.
My thesis has a spectrum: from a discounted debt claim at one end to nothing at the other. I can document the “nothing” end15. I cannot document the “discounted claim” end for these Partners Group vehicles.
I found no Partners Group or Delaware-fund filing that quantifies how the unmet 6% overshoot is economically treated – no confirmed conversion into a discounted note, only the general statement that withdrawals would be capped12. So the payout-shortfall comparison against Luxembourg’s 38% deferral does not resolve to a clean like-for-like number. A tender cap resolves by cutting, not by queuing; that much the record supports.
Partners Group would push back, and the push-back is not weak. Its own June 2026 disclosure framed the 5% quarterly limit as a standard feature across its evergreen vehicles, not an emergency lever16.
The platform data backs the composure. Evergreen funds still made up 26% of new client commitments in H1 2026, and gross client demand rose 33% year-on-year to $16.0bn1718. First-half new-money intake beat expectations for client demand19.
On that reading, this is standard liquidity management working as written.
Two things I set out to establish, I could not. I found no dated sequence showing which signal – the gate breach, the shift in commitment share, or the gap between gross and net demand – moved fastest from first concern to public breach.
And I found no comparable disclosure from other evergreen sponsors laid out the same way, so I cannot claim this three-part pattern is industry-wide rather than a feature of how Partners Group happens to disclose. I am not going to assert a structure I could not source.
The redemption story did not stay in the footnotes. On 1 September 2026 Partners Group’s CEO announced his departure, with H1 net profit down 13% year-on-year to CHF 502 million20.
By that date the shares had lost roughly a third of their value21. The market was pricing something broader than one gated fund.
The number I keep coming back to is not the 9.8% or the 6%. It is the governance difference behind them: in Luxembourg the holder decides what happens to the deferred slice, and in Delaware the structure decides.
Whether a Partners Group investor gets a say in the outcome depends entirely on which side of the Atlantic their vehicle sits.
This is an observation and does not constitute investment advice. Nothing here is a recommendation to buy, sell or hold any security or fund; the decision, as ever, stays with the reader.