French SCPI funds: what lies behind the fall in exit queues

SCPI: measuring the risk, Part 1 of 6. Documentary research cut off on 13 September 2026. National statistics refer to 30 June 2026.

An investor’s exit request can disappear from a register without the investor receiving any cash. Primopierre’s change of trading arrangements provides a concrete example. Existing redemption requests are cancelled, and holders who still wish to sell must submit a fresh sell order. Until their units are transferred, they continue to own them. (Praemia REIM)

That distinction matters when reading the national figures. ASPIM and IEIF report €1.9 billion of SCPI units awaiting an exit at 30 June 2026, a 31% decline since December. Their 5 August release explains that register resets and prices in the new secondary markets account for the reduction among funds that suspended their variable-capital arrangements. (ASPIM / IEIF)

A smaller reported backlog is therefore not enough to establish a recovery in liquidity. First, how many units actually changed hands?

An exit depends on a buyer

A SCPI, short for société civile de placement immobilier, is a French unlisted property investment vehicle. Investors own units in a portfolio of rental assets rather than buying buildings themselves. These are not exchange-traded REIT shares. Nor does a valid redemption request create an entitlement to immediate payment: France’s market regulator, the AMF, has explained that the wait can be indefinite. This article concerns units held directly, rather than applying their exit rules to holdings inside life-insurance contracts. (AMF, February 2024)

Under the usual variable-capital arrangement, money from new subscriptions can provide the counterpart to outgoing investors’ redemption requests. Requests are processed according to their place in the register and the applicable conditions. A backlog develops when incoming demand is insufficient. (AMF, February 2025)

There can be other routes. A redemption reserve, if established and funded, may support withdrawals within its rules. A holder can also arrange a privately negotiated transfer. Primopierre’s offering document describes these options; neither amounts to an unlimited pool of available cash. (Offering document, pages 16–20)

Income and liquidity answer different questions. A distribution shows that a payment has been made to a holder. It does not establish that somebody would buy the holding. An investor may therefore discover the importance of illiquidity when cash is needed, before any loss has been realised through a sale.

Suspending variable capital changes the route out. The manager organises a secondary market that matches sell orders with buy orders. The investor is now offering units to another investor at a price a buyer may accept, rather than asking for redemption through the previous arrangement. The AMF explicitly distinguishes these two processes. (AMF, February 2025)

Two opposing movements sit behind the national decline

The end-2025 release put the pending stock at approximately €2.8 billion. (ASPIM / IEIF, 2025 results) The next half-year’s reported change breaks down into about −€1.2 billion across eleven funds that suspended variable capital and about +€330 million among funds that retained it. The aggregate decline is approximately €870 million. (ASPIM / IEIF, first half of 2026)

The first number is not €1.2 billion paid back to departing investors. It combines changes to the registers with the re-entry of units at prices observed in secondary trading. The release does not isolate precisely how much comes from quantities being re-registered and how much from prices. (ASPIM / IEIF)

Rounding also matters. Subtracting the headline stocks of €2.8 billion and €1.9 billion yields €900 million, rather than the reported change of roughly €870 million. These are not sufficiently precise inputs for a reconciliation to the last euro. The published figures do not provide an exact reconciliation.

ASPIM itself cautions against treating the total decline as evidence of a general improvement in liquidity. The qualification is in the original release. The risk is that a headline reproduces the percentage without the explanation. (ASPIM / IEIF)

Reconstructing a national backlog on an unchanged basis would require consistent prices and records of subsequent trades, fresh requests and genuine cancellations. The published amounts do not supply that reconciliation.

Unit counts and valuation prices

A backlog expressed in euros depends on both the number of units counted and the price used to value them. A fall in price can reduce the value of the backlog without a single unit leaving it.

This had already happened before the 2026 suspensions. In its first-quarter 2025 release, ASPIM attributed a 4.9% reduction in the value of units awaiting an exit to price cuts, while the number of units remained broadly unchanged. (ASPIM / IEIF, 13 May 2025)

A simple example: lower value, no completed exit

Hypothetical figures, not the accounts of an actual fund. A register contains 1,000 units valued at €100 each, giving a €100,000 backlog. The reference price falls to €60. The same 1,000 units are now worth €60,000.

The reported value falls by 40%. The number of pending units is unchanged. No trade has taken place. This example isolates the price effect; it is not a reconstruction of ASPIM’s figures.

A register reset introduces another source of uncertainty. Holders who previously requested redemption need not all submit fresh orders at the same time. Some may wait for a better price, revise their plans or arrange a private transfer. Others may still wish to leave without having re-registered. These are possible explanations, not behaviours quantified by the documents examined here.

A register records the instructions submitted under its rules. It does not reveal every holder’s current intentions. Fewer visible orders therefore establish neither that the original requests were fulfilled nor that every original seller still wants exactly the same exit.

Primopierre: from registers to transactions

Primopierre provides a way to test the mechanism against actual records. Its suspension of variable capital was approved on 7 January 2026. Praemia REIM says affected holders must expressly submit their wish to sell into the secondary market. The first matching session took place on 26 March. (Praemia REIM)

At 31 December 2025, the annual report recorded 2,190,372 pending units, including confirmed and unconfirmed requests. Its methodological note explains that unconfirmed requests include incomplete files that have already obtained a place in the register. Reporting had changed in the second quarter of 2025, so a comparison with 2024 would require an adjustment. The annual report labels this row as units awaiting a sale, while its note refers to redemptions. The Q2 2026 bulletin repeats the same December count under pending redemptions, separately from the new sale register. (2025 annual report, page 6, note 4; Q2 bulletin, page 3)

The second-quarter bulletin makes the switch unusually visible. For 30 June 2026, the pending-redemption row contains a dash; a separate row records 146,937 units awaiting a sale, again including confirmed and unconfirmed requests. The dash must not be converted into a zero. The old category is no longer populated on the same basis. (Second-quarter bulletin, page 3)

The published trading history, accessed on 13 September, records 171 units traded in March, 1,756 in April, 2,021 in May and 2,959 in June. That adds up to 6,907 units traded through the organised market in the first half. The final three observations also appear in the quarterly bulletin. The history includes sessions after June; this analysis uses the first half to keep the same observation period. (Trading history; bulletin, page 3)

The half-year trading volume is approximately 0.32% of the 2,190,372 units in the old register. This is a calculation showing relative scale, using the records cited above. It is not the clearance rate of the original redemption queue.

A genuine clearance rate would require identifying units sold by the original applicants, incorporating privately negotiated transfers and avoiding double counting where units traded more than once. The public table does not provide that tracking. Nor can the three numbers simply be subtracted to produce a count of “missing sellers”: a dated stock, a replacement register and a gross trading flow are not a complete reconciliation.

The switch creates a break in comparability. Subtracting the two registered stocks does not measure completed exits. Primopierre is one case, rather than a representative sample of the SCPI market.

The register can change before trading begins

A second manager documents the same sequence. La Française states that LF Grand Paris Patrimoine suspended variable capital from 1 June 2026, cancelling existing requests. Its first matching session took place on 31 July, after the half-year reporting date. (La Française announcement)

That timing matters. The procedural change was already in force at the June snapshot, while trades from the first matching session had not yet happened. July transactions cannot explain liquidity achieved by the end of June.

This provides corroboration of the mechanism at another manager. It does not reconstruct the individual contributions of all eleven funds to the national figure. The publicly accessible material examined for this article did not supply the detail needed for that reconciliation.

A secondary market can be part of the remedy

The change in trading arrangements can serve a practical purpose. An advertised redemption price is of little practical use when no counterparty is available. Matching orders may allow a transaction at a price a buyer is genuinely prepared to pay. Under French law, that interaction between supply and demand determines the execution price. (Monetary and Financial Code, Article L. 214-93, I)

A transfer between investors does not, by itself, force the SCPI to sell a building to fund the exit. The buyer takes over the seller’s units. That follows from the transaction’s structure and distinguishes it from a redemption financed out of the vehicle’s resources. The arrangement can therefore provide useful liquidity, even when the selling price is disappointing.

But a market’s existence does not establish its capacity. A holder unwilling to accept the available price remains invested. Judging improvement requires completed transactions, their volumes and the net proceeds available to sellers, not merely an announcement that an order book has opened.

The national market is also far from uniform. At the end of 2025, ASPIM reported that roughly three quarters of the pending stock was concentrated in fifteen SCPI funds run by seven managers, mainly with office exposure. (ASPIM / IEIF, 2025 results) One stressed fund’s experience cannot be turned into a diagnosis for every vehicle.

Request lifetimes and the statutory warning threshold

The fresh instruction required after a suspension should not be confused with routine renewal. The AMF ombudsman explains that an ordinary redemption request has no expiry date. That is different from a secondary-market sell order. (AMF, February 2025)

There is also a statutory warning procedure. If redemption requests unfulfilled for twelve months account for at least 10% of units, the manager must promptly notify the AMF. A parallel rule applies to sell orders registered for more than twelve months and accounting for at least 10% of units. The manager must then call an extraordinary general meeting within two months of that notification, proposing the disposal of some or all of the assets and any other appropriate measures. (Article L. 214-93, II)

The test therefore involves both unit counts and the age of the requests. It is neither a twelve-month payment guarantee nor an automatic liquidation trigger. The national backlog measured in euros cannot establish whether an individual manager has crossed that threshold or breached its obligations.

Follow the exit through to completion

A more useful monitoring framework would retain a link between the old and new registers. It would follow a cohort, the requests outstanding on a specified date, and record what happened to them. Completed exits, voluntary cancellations, procedural cancellations and fresh registrations could then be distinguished rather than rolled into one apparent improvement.

For each vehicle, four disclosures would make the assessment more robust: the number of units still offered for sale; the age of requests and the rules used to measure it; completed trading volumes, separating the organised market from private transfers; and net selling prices with payment delays. A euro valuation of the backlog would still be useful, provided its pricing basis and the effect of price changes were explicit.

Another tempting shortcut should be avoided: dividing today’s backlog by last month’s trading volume to announce an exit date. That assumes, among other things, that buying demand, prices and execution priority remain consistent with the projection. In an order-matching market, each seller’s price matters. A quotient is not a commitment to execute an order.

A new register changes what the indicator measures. Assessing exit opportunities requires following completed sales and the cash returned to holders. The 31% decline leaves that question open; it cannot establish widespread insolvency either. The second instalment examines the prices at which units actually find buyers.

For further reading, our analysis of semi-liquid private-credit funds and redemption gates examines other exit arrangements. Those funds have different contracts and regulatory rules from SCPI vehicles.

Sources

S01. ASPIM / IEIF : First half of 2026. Published 5 August 2026; data to 30 June.

S02. ASPIM / IEIF : 2025 results. Published 9 February 2026; stock at 31 December 2025.

S03. ASPIM / IEIF : First quarter of 2025. Published 13 May 2025; price effects on the backlog.

S04. AMF : Redemption execution delays. Published 1 February 2024; directly held units.

S05. AMF : Redemption requests and sell orders. Published 3 February 2025; different exit procedures.

S06. Praemia REIM : Primopierre unit market. Dynamic page accessed 13 September 2026; March–June observations used.

S07. Primopierre : 2025 annual report. PDF page 6 and note 4; exact publication date not confirmed.

S08. Primopierre : Second-quarter 2026 bulletin. PDF page 3, definitions on pages 6–7; exact publication date not confirmed.

S09. Primopierre : Offering document. Version updated 1 August 2026, PDF pages 16–20.

S10. La Française : LF Grand Paris Patrimoine. Announcement accessed 13 September 2026; suspension on 1 June, first matching on 31 July.

S11. Légifrance : Article L. 214-93. Version in force since 28 July 2013, accessed 13 September 2026.

Limitations

This investigation draws on ASPIM and IEIF releases, managers’ documents, the AMF ombudsman’s explanations and the Monetary and Financial Code. Industry statistics are attributed to their producers; they have not been reproduced from an exhaustive independent dataset. The Primopierre case reconciles disclosures from the same manager. That is a consistency check, not an independent accounting audit.

Tables were read alongside their notes. Calculations are stated, and missing entries, rounded values, observation dates and definition changes are preserved. No interview or requested response from a manager is presented as having been obtained. Possible explanations for holders not re-registering remain hypotheses. This is an analysis of market arrangements and public data, not an individual recommendation to buy or sell.