French SCPI funds: the price of getting out

SCPI: measuring the risk, part 2 of 6. Documentary research closed on 13 September 2026. Transactions examined: March–August 2026.

On 18 August 2026, 120 units in Élysées Pierre changed hands at a reported €300 net to the seller. Yet HSBC REIM had been publishing a €516.06 withdrawal price since 3 August. The gap is 41.9%. This is a comparison between two exit mechanisms, not a unit price falling by that amount in fifteen days. (HSBC REIM)

For a saver who needs the money, that distinction matters. A withdrawal price sets the amount applicable when a withdrawal is completed under the fund’s rules. It does not promise an immediate buyer. The August auction records orders that actually matched, but only for 120 units. It says nothing about how many additional units could have sold at the same price. France’s financial regulator warns that even a properly submitted withdrawal request can remain outstanding indefinitely. (AMF Ombudsman)

The first part of this investigation examined changes to the registers used to track exit requests. This article follows the transactions instead. A resale price needs a date, a quantity and a clear account of what the buyer pays and the seller receives.

One unit, several very different prices

A SCPI, short for société civile de placement immobilier, is a French unlisted vehicle through which investors own an interest in a portfolio of rental properties. It is not a stock-exchange-listed REIT. Its documents contain several prices and valuations because these answer different questions. (AMF)

The subscription price applies when new units are issued. In the usual variable-capital arrangement, the withdrawal price applies to an exit through the fund’s withdrawal mechanism. A secondary-market sale transfers existing units to another investor, through orders matched under the vehicle’s rules. Élysées Pierre describes both primary and secondary routes: the existence of its periodic auction does not mean its variable-capital mechanism has been suspended. (Élysées Pierre offering document, pages 9 and 13–15)

The realisation value is an appraisal-based measure of net assets: the estimated market value of the properties plus other assets, net of liabilities. The reconstitution value adds the costs of rebuilding the portfolio. Neither is a standing bid for every investor’s units. (Élysées Pierre offering document, page 9)

Nor do the rules governing the subscription price create a floor under resales. Article L. 214-94 of France’s Monetary and Financial Code requires a subscription price more than 10% away from reconstitution value to be justified and notified to the AMF. It is not a promise that another investor will buy at that level. (Légifrance)

These figures can usefully be compared. Treating them as interchangeable, however, produces misleading claims about losses and bargains.

Four trades, with the quantities left in

The following observations come from the managers’ published trading histories. They are selected case studies, not a representative sample of the SCPI market. The funds differ, as do the trading dates. (HSBC REIM, Praemia REIM, La Française: LF Grand Paris Patrimoine, La Française: Crédit Mutuel Pierre 1)

*Nominal euros per unit. “Net seller” follows the manager’s market convention, not an individual investor’s after-tax return. The buyer price includes the charges reported in the trading history. Volumes count units, not people. *

One tempting shortcut is to compare an old purchase price, including entry costs, with a recent net sale price and call the difference the investment’s “performance”. A total-return calculation would also need the distributions received, the dates and the actual costs incurred. This table does not reconstruct any individual holding period.

Another shortcut is to multiply the latest trade price by every outstanding unit and treat the result as money that all investors could collect. A trade shows that a particular quantity found counterparties. It does not establish how a much larger sale would have affected the price.

There is also a gap within each transaction: the buyer’s outlay is not the seller’s proceeds. At the Élysées Pierre auction, those amounts were €327.60 and €300, a €27.60 difference. That difference must not disappear from a comparison. Equally, it cannot be turned into a universal fee schedule: the relevant duties, commissions, calculation bases and rounding rules are vehicle-specific. (HSBC REIM; offering document, page 17)

Primopierre: six auctions from March to August

Primopierre provides a sequence rather than a single observation. Its variable-capital mechanism was suspended on 7 January 2026 and its organised secondary market began on 26 March. Between that first auction and 27 August, the reported net seller price moved from €45 to €36, a decline of 20%. The August auction matched 1,747 units. (Praemia REIM)

Inspect all six observations, including buyer prices

Primopierre : published auctions, March–August 2026

Date	Net seller (€)	Buyer (€)	Units
26 Mar 2026	45.00	49.41	171
30 Apr 2026	40.00	43.92	1,756
28 May 2026	40.00	43.92	2,021
25 Jun 2026	40.00	43.92	2,959
30 Jul 2026	38.71	42.51	1,637
27 Aug 2026	36.00	39.53	1,747

Source: Praemia REIM’s official trading history, consulted on 13 September 2026. These are organised auctions, not an exhaustive record of any privately negotiated transfers.

Together, the six auctions account for 10,291 units. That is approximately 0.054% of the 18,948,080 units outstanding on 30 June. This comparison gives a sense of scale, not a count of investors who have exited: the same unit could trade more than once. It also compares a March–August flow with a June stock. (Trading history; second-quarter bulletin, page 3)

It would be equally misleading to extrapolate a month’s trading volume into a forecast of how many years it would take to clear an exit queue. New orders, revised limits and changing demand make that a false precision.

Published auction histories also need not capture every transfer. For example, Élysées Pierre’s first-half bulletin separately reports 388 units transferred in privately negotiated transactions during January–June 2026. That demonstrates a distinction between channels; it is not an estimate of unreported activity in Primopierre. (Élysées Pierre bulletin, page 2)

How the order book sets the price

A call auction gathers orders for a scheduled matching session. Sellers specify the minimum they will accept; buyers specify their maximum. The auction chooses the price that permits the largest number of units to change hands. Further rules resolve ties between possible prices, while execution priority follows price and then the age of an order. (AMF, article 422-213 and article 422-214; Élysées Pierre offering document, pages 13–15)

The resulting price is neither an average of sellers’ wishes nor a fresh property appraisal. It reflects the orders present at that session. A simple example shows how an auction can discover a price and still leave most offered units unsold.

Auction lab: why 600 units offered can result in only 250 trades

Entirely hypothetical. All prices use the same simplified, net-seller-equivalent basis. Fees are excluded and partial execution is allowed. This is not a live order book or a simulator for a personal sale.

The hypothetical orders

Limit price (€)	Units offered at this minimum	Units sought at this maximum
40	100	100
50	200	170
60	300	80

At €40, all buyers would be willing to pay: they seek 350 units, but sellers offer only 100 at that price. At €50, there are 300 units available and demand for 250. At €60, all 600 offered units become eligible, but buyers seek only 80.

Compatible quantities at each candidate price

Price (€)	Cumulative supply	Cumulative demand	Executable units
40	100	350	100
50	300	250	250
60	600	80	80

The auction clears at €50 for 250 units. The other 350 offered units do not trade. Sellers who were prepared to accept €40 also receive the €50 execution price.

Now cut every buying quantity to one-tenth, leaving the sellers unchanged. Executable volumes become 35 units at €40, 25 at €50 and 8 at €60. The new maximum is 35 units at €40. In this example, nothing about the properties has changed; only the buy orders have.

Finally, suppose no buyer will offer more than €20 while sellers require at least €40. There is no match. The absence of an execution price does not mean a unit is worth zero.

Mechanism: AMF article 422-213 and article 422-214. Hypothetical orders and calculations by l0g. These examples do not require the tie-breaking rules.

A displayed order book is not a guaranteed future trade either. Orders can change before the next matching session. In the Élysées Pierre material consulted for this article, the buyer and seller books even carried different update dates. Combining them into a supposedly executable real-time quote would introduce an assumption the source does not establish. (HSBC REIM; La Française’s market explanation)

Property risk, debt and liquidity in the discount

Here, a discount means a transaction price below a specified reference value. Primopierre’s €36 net seller price on 27 August was 49.1% below its €70.76 realisation value on 30 June, calculated as 36 / 70.76 − 1. That compares a unit transaction with an earlier appraisal-based net asset value. It does not show that every building has lost 49.1%. (Primopierre trading history; bulletin, page 3)

Several explanations are compatible with such a gap. A buyer may demand compensation for a difficult future exit. The buyer may also expect weaker rent, more capital expenditure, financing constraints or further valuation cuts. What changes hands is an interest in a vehicle with liabilities and operating costs, not a slice of property that can immediately be sold at its published appraisal. Rental, financing and liquidity risks are described in the fund’s documentation. (Primopierre offering document)

The collected evidence does not identify how much of the discount comes from each factor. Describing the whole gap as a “liquidity discount” would turn an observed price into an explanation that is more precise than the evidence allows. Neither the sellers’ personal circumstances nor the buyers’ reasoning is public.

The opposite claim : that this is merely an order-book phenomenon with no deterioration in the assets : also fails. LF Grand Paris Patrimoine reports a 14.0% like-for-like decline in appraised property values between December 2025 and June 2026. Its realisation value per unit fell 24.9% to €130.31 over that period. The bulletin explains the amplification through liabilities and other net balance-sheet items. Those are the manager’s figures and interpretation, not an independent reappraisal by l0g. (Bulletin, pages 3–5)

A stripped-down, hypothetical balance sheet illustrates the principle. Properties worth 100, funded with debt of 40, leave 60 for investors. If the properties fall to 80 while debt stays at 40, only 40 remains. Property values have fallen 20%, but net assets have fallen by one-third. A unit’s underlying value can therefore decline faster than the buildings’ value, even before resale conditions enter the calculation.

What a resale changes for the fund

The secondary market can still serve a useful purpose. An investor replacing another as shareholder does not, in itself, require the fund to sell a building. The money paid to the seller is not a new subscription adding that amount to the collective property portfolio. The offering documents distinguish these transfer mechanisms. (Élysées Pierre; Primopierre)

It would therefore be wrong to treat each discounted resale as an equivalent cash outflow from the SCPI itself. The same distinction explains the manager’s argument for an auction: investors may obtain an exit without forcing immediate property disposals. La Française presents the arrangement in terms of preserving the portfolio. It guarantees neither a buyer for every seller nor protection against forced disposals for other reasons. (La Française; LF Grand Paris Patrimoine bulletin)

For a buyer, the logic cuts both ways. Purchasing below realisation value creates no entitlement to be repaid at that value tomorrow. What happens next depends on rent, appraisals and market conditions; the discount may persist. An observed discount is not enough to establish a bargain.

The risk lies in the terms of an achievable exit

These four cases do not describe the entire industry. ASPIM and IEIF report €2.2 billion in net SCPI inflows in the first half of 2026. They also record €489 million in matched withdrawals and unit trades combined. That second measure is broader than call-auction turnover. (ASPIM/IEIF, release of 5 August)

Those flows provide a useful counterweight to the selected cases, not an answer to every investor seeking to leave. Money subscribing to one SCPI does not automatically finance an exit from another. At fund level, the relevant evidence is the price reported as executed, the quantity traded, the remaining orders and the changes in income and asset values.

The investigation establishes substantial gaps between some published reference values and transactions in limited quantities. It does not establish the liquidation price of every unit, or a single cause for every discount. The warning is concrete enough without that extrapolation: a published asset value does not tell an investor how much cash can be recovered, for how many units, or on what date.

Part three will examine distributions: a quoted yield, the income actually received and the change in that income are different measures.

Sources

Inspect the 16 primary references

1. HSBC REIM : Élysées Pierre: trading and August notice. Trading history and August 2026 notice. Withdrawal price effective 3 August; auction dated 18 August. Exact page publication date unconfirmed.
2. HSBC REIM : Élysées Pierre offering document. Offering document, pages 9, 13–15 and 17: price definitions, order matching and costs. Version inspected includes prices effective 3 August 2026; publication date unconfirmed.
3. HSBC REIM : Élysées Pierre H1 2026 bulletin. First-half 2026 bulletin, pages 2–3. Accounting period January–June; also includes a later price change. Publication date unconfirmed.
4. Praemia REIM : Primopierre trading history. Official history of six auctions, 26 March–27 August 2026. A page update is not a new trade; no independent settlement confirmation.
5. Praemia REIM : Primopierre Q2 2026 bulletin. Second-quarter 2026 bulletin, page 3. Outstanding units and net asset reference as at 30 June. Publication date unconfirmed.
6. Praemia REIM : Primopierre offering document, August 2026. Offering document, version dated 1 August 2026. Transfer mechanisms and charges. Version date is not a verified upload date.
7. La Française : LF Grand Paris Patrimoine trading history. Official LF Grand Paris Patrimoine trading history: auction of 31 July 2026. Other order-book entries carry separate dates.
8. La Française : Crédit Mutuel Pierre 1 trading history. Official trading history, read directly: auction of 31 July 2026.
9. La Française : LF Grand Paris Patrimoine Q2 2026 bulletin. LF Grand Paris Patrimoine bulletin, 30 June 2026, no. 26-03, pages 3–5. Appraisal and net asset changes versus December 2025. Manager-reported figures; aggregate chart values are rounded.
10. AMF : Understanding SCPI documentation. AMF explanation of SCPI documents. Used for definitions of realisation and reconstitution values, not older yield conventions. Primary indexed text.
11. AMF : General Regulation, article 422-213. Price maximising executable units, tie-breaks and no-match cases. Rule cross-checked against the Élysées Pierre offering document, pages 14–15.
12. AMF : General Regulation, article 422-214. A single execution price and price/time priority. Rule cross-checked against the Élysées Pierre offering document, page 15.
13. AMF Ombudsman : SCPI withdrawal delays. AMF Ombudsman, published 1 February 2024: a valid withdrawal request can remain outstanding indefinitely. Direct holdings, not life-insurance terms.
14. Légifrance : Monetary and Financial Code, article L. 214-94. Monetary and Financial Code, article L. 214-94, version effective 28 July 2013. Subscription pricing rules, not a guaranteed resale floor.
15. ASPIM/IEIF : Retail property funds, H1 2026. ASPIM/IEIF, published 5 August 2026, covering January–June 2026. Industry aggregate; matched withdrawals and unit trades are a combined measure.
16. La Française : SCPI call-auction markets. La Française explanation of call-auction markets. Manager-produced material; its portfolio-protection argument is not an independent test of effectiveness.

All sources were consulted on 13 September 2026. Fund managers are primary sources for their own vehicles; their publications do not independently establish every economic explanation discussed above.

Limitations

This article covers units held directly. Liquidity under a life-insurance contract requires a separate examination of its terms. Prices and quantities are reported by the fund managers. We have not inspected bank settlement confirmations, investors’ motivations or an exhaustive register of privately negotiated transfers. No request for comment is presented as having been sent.

Accounting periods, appraisal dates, price-effective dates and auction dates are distinguished throughout. Calculated gaps are identified and are not total returns. The auction lab and simplified balance sheet are entirely hypothetical. This documentary analysis is not a recommendation to buy or sell.