Agefi Luxembourg - septembre 2026

AGEFI Luxembourg 28 Septembre 2026 Fonds &Marchés By Tim VANVAERENBERGH, CEO Shelter IM I n the first half of 2026, 68% of ac­ tively managed global equity funds in the European UCITS universe failed to beat their benchmark, mean­ ing only 32% added value after fees. On its own, that number is unremarkable; active managers have fought an uphill battle against passive strategies for years. What stands out in the latest Barometer from Shelter Investment Man­ agement is what lies under­ neath it: a regional breakdown showing how uneven that battle is, and a longer time series that puts the fig­ ure in perspective. Shelter has tracked this underperformance rate every year since 2015, and the range has been wide: from 48% in 2017, the one year active man­ agers as a group came out ahead, to a peak of 88% in 2024. Against that backdrop, 68% for the first half of 2026 sits on the milder end of the historical range. Global active management is still losing more often than winning, but 2026 has not been an unusually bad year for it. The real story this quarter is regional. Split by region, the picture looks very different depending on where a manager hunts for alpha. In Europe, 77% of active funds trailed their index over the first six months of the year. In the United States, 75%underperformed, close behind. In emergingmarkets, the pat­ tern reverses entirely: only 39% of active funds lagged the benchmark, meaning 61% beat it. The gap is toowide to dismiss as noise; it points to a structural dif­ ference in how active manage­ ment functions across developed and developing markets. Efficiency cuts both ways In Europe and the US, deep liquidi­ ty, dense analyst coverage andnear­ instant information flow leave little room for a manager to exploit mispricing the rest of the market has missed. Every edge is quickly arbitraged away, and after fees the median active manager in these markets has struggled for years to keep pace with a simple tracker. The first half of 2026 was no exception. Emerging markets operate under different conditions: cov­ erage is thinner, information less evenly distribut­ ed, and pricing inefficiencies persist longer before the market corrects them, precisely the environ­ ment in which stock picking can still pay off. The Q2 2026 numbers show it, with a clear major­ ity of active managers in emerging markets adding value over their benchmark in the first half of the year. The AI and chip divide A second factor helps explain why global active funds struggled so broadly this year: the extreme concentration of returns around artificial intelli­ gence and semiconductors. This was not con­ fined to largecap technology indices, where the effect might be expected; AI and chip names also drove returns in corners of the market where investors would not normally look for them, including small caps and value strategies. The practical result is that 2026 split active man­ agers into two camps: those who held AI and semiconductor exposure, and those who did not. That single allocation decision, more than man­ ager skill or process, explains much of the dis­ persion in returns this year. Afundwith strong fundamentalsbased selection but limitedAI exposure had little chance of keep­ ing up with its benchmark, however sound its process. Not active versus passive, but where active belongs Taken together, theQ2 2026 data reframes a debate that too often gets stuck at the wrong question. While the data of two quarters is not sufficient to draw major conclusions, we see the trends over the past semester are structural over the long term. The issue is not whether active or passive man­ agement is superior in the abstract, but where active management can realistically be expected to add value, and where it cannot. In efficient, wellcovered markets such as European and US large caps, a lowcost tracker remains a reasonable default; the odds are structurally stacked against active managers there, AI concentration or not. In less efficient markets, where information asym­ metry is greater and coverage thinner, a spe­ cialised activemanager has a genuine opportunity to outperform, as this quarterʹs emerging market numbers confirm. This is the logic behind a CoreSatellite approach to portfolio construction: a broad, costefficient Core built around passive exposure where mar­ kets are efficient, complemented by targeted Satellites where active management has a real chance of making a difference. Shelterʹs Q2 2026 results are a reminder that this is not a theoretical distinction. It is measurable, quarter by quarter, in the performance gap between developed and emerging market active funds. For allocators, the takeaway is not to abandon activemanagement, nor to embrace it uncritically, but to be selective about where it is deployed. Choosing the right battleground, rather than the right manager alone, may be the more decisive factor in the year ahead. Methodology: the Shelter Barometer screens UCITS equity funds above EUR 100 million in assets with a correlation above 0.70 to their benchmark over the preceding three years, using net­ offee euro returns. The Q2 2026 edition covered 893 global, 319 European, 298 US and 256 emerging market funds. Source: Morningstar Direct. Active managementʹs geography problem: Why emerging markets are beating developed ones By Virginie LEROY and Lissie GOLDBACH, Co­ Founders&Managing Partners, Leroy&Goldbach W hen the ordinary exit routes narrow, private equity does not stop selling—it changes counterparty. In 2025 the global se­ condarymarket reachedUSD240 billion, up 48%year on year, ac­ cording to Jefferies. General partner (GP)led transactions accounted for USD115 bil­ lion, up 53%, and for the first time singleasset continuation vehicles ex­ ceededhalf of all continua­ tion vehicle volume. A continuation fund allows a sponsor to transfer one or more assets from a fund approaching the end of its term into a newvehiclemanaged by the same sponsor. Existing investors choose between liquidity — a cash exit at the transaction price, funded by incoming secondary capital — and rolling their interest into the new vehicle. It is a different transaction in kind: the sponsor is simul­ taneously the seller’s fiduciary, the buyer’s man­ ager and the architect of the price. Nothing about that is improper in itself; everything about it de­ pends on process. Why the regulator is looking In its 2026 priorities for supervising the investment fund sector, publishedon 31March 2026, theCom­ mission de Surveillance du Secteur Financier (CSSF) announced onsite controls of the valuation organisation of investment fund managers (IFMs) together with “thematic samplebased reviews re­ lating to openended private assets funds (includ­ ing continuation funds)”. Neither the reserved alternative investment fund (RAIF) nor the unregulated partnership is itself a supervised product; supervision reaches the trans­ action through the alternative investment fund manager (AIFM) — in practice an authorised one — inwhose records the transaction lives. Its place­ ment is instructive: the regulator locates continua­ tion vehicles within the valuation problem in private assets. For a transactionwhose central diffi­ culty is the price at which a sponsor sells to itself, scrutiny belongs exactly there. The vehicle next to the legacy fund NorthAmerica accounts for roughly 68% of GPled volume against some 26% for Europe, per Lazard. But for vehicles that need to reach European institu­ tional investors, the trend line runs through Luxem­ bourg— the passport under theAlternative Invest­ ment Fund Managers Directive (AIFMD), and a service ecosystem in which, as two Luxembourg practitioners put it lastNovember, “youdo not need to explainwhat a continuation vehicle is here”. Acontinuationvehiclerunsonatransactioncalendar, not a fundlaunch calendar. The workhorse is the Luxembourg limited partnership —most often the speciallimitedpartnership(SCSp),whichhasnolegal personality separate from its partners, although the common limited partnership (SCS), which does, serves the same purpose where a vehicle with legal personality is preferred. Either is deployed as an al­ ternative investment fund outside any Luxembourg product law, or as a RAIF. TheRAIF is not subject toprior approval by a super­ visory authority — its offering document must say so on the cover page (Article 39 of the Lawof 23 July 2016onreservedalternativeinvestmentfunds)—but must be managed by an authorised AIFM (Article 4(1)).Itisreservedtowellinformedinvestors(Article 2) and offers umbrella compartments whose assets and liabilities are ringfenced by lawunless the con­ stitutive documents provide otherwise (Article 49). The nonproduct partnership is lighter still, at the price — today — of no compartment regime. That gap is closing. Bill 8814, depositedwith theChamber of Deputies on 30 July 2026, would amend the 2013 AIFMLaw to allowSCS and SCSp vehicles qualify­ ingasAIFstoestablishringfencedcompartmentson twoconditions:anAIFMauthorisedinLuxembourg or another Member State, and compartments pro­ vided for in the constitutive documents. Investors’ and creditors’rights are then confined, by default, to their compartment’s assets. A sponsor running successive GPled transactions could house each singleasset continuation compart­ mentwithinonepartnershipinsteadofincorporating anewvehicleforeverydeal—onemoreoptioninthe Luxembourg toolbox.Abill isnot a law; a voteisanticipatedbeforetheendof2026. Eligibilityrulesthatdonotfiteveryin­ vestorexpectedtoroll—coinvestors, theteam’scarryvehicle,smallerlim­ ited partners (LPs) — force exclu­ sionstheconsentprocessmustthen explain. The vehicle question, correctlyanswered,deliversthe parties to the conflict question. The samemanager onboth sides A continuation fund places the same sponsor on both sides of a sale — and usually the same AIFM as well, sponsorowned or thirdparty. The AIFMD requires the manager to identify, pre­ vent,manageandmonitor conflictsof interest—ex­ presslyincludingconflictsbetweentwofundsunder commonmanagement —and its Level 2 delegated regulation turns the principle into concrete obliga­ tions: awrittenpolicy, arm’slengthprocedures and, where those do not suffice, disclosure to investors. Fair treatment completes the frame: no investormay obtainpreferential treatment unless it is disclosed in the fund’s rules or instruments of incorporation. AIFMD II left this framework essentially intact but added one provision that speaks directly to this transaction: anAIFMmanaging anAIF at the initia­ tive of a third partymust now evidence to its home regulatorhowitprevents,orotherwisemanagesand discloses, the conflicts arising from that relationship (Article14(2a)). The evidence supervisors expect has a recognisable shape: a conflictminutedbeforeprice discussions begin; a decision trail showing who acted forwhich side; informationparitybetween in­ comingandexisting investors; anda limitedpartner advisorycommittee (LPAC) involvedwhile itsview can still change the transaction. One point specific to the Luxembourg thirdparty AIFMmodelregularlysurprisessponsors:theAIFM cannotsimplysignoffonthetransactionthesponsor has negotiated. Its responsibility for valuation and conflicts stays with it, whatever the service agree­ ment says. The division of labour that survives scrutiny is not “theGPdecides, theAIFMoversees” —it is: theGPproposes, and theAIFMmust be able to show, in its own records, that it decided. Valuation: who signs, and onwhat basis UnderArticle 19of theAIFMD, valuationperformed by theAIFM itself must be functionally independent from portfolio management; the AIFMmay instead appoint anexternal valuer, but theappointment does not discharge it—theAIFMremains responsible for the proper valuation of the assets. Hence a fairness opinion is not by itself a defence: it does not replace theAIFM’s own valuation obligation, and it does not answerthequestionsaninspectorwillask—whode­ termined the value, under which policy, and how it reconciles with the last net asset value and with the price atwhich thirdparty capital committed. The CSSF’s feedback report on its thematic reviewof the valuation framework for less liquid and illiquid assets, published on 4 June 2026, asks every IFM to benchmarkitselfagainstthereport’sobservationsand recommendations and take corrective measures whereneeded. Themessage is clear: the strongest file combines competitive price discovery with an inde­ pendent valuation, a documented reconciliation to prior carrying values, and a record showing that the sponsor’s own economics were identified and weighed rather than assumed away. Consent: the election is the deal For existing investors, the transaction is the election: sellatthetransactionprice,orrollintothenewvehicle. TheLPACcomesfirst—andwhatitcandoisafunc­ tion of the legacy fund’s documents, not of the trans­ action. An LPAC that can withhold approval of an affiliatetransactionisacontrol;onethatismerelycon­ sulted is awitness. TheInstitutionalLimitedPartnersAssociation(ILPA) publishedacontinuationfunddisclosuretemplateon 27 January 2026; its draft updated guidance — con­ sultedonthissummer,finalversionduelaterthisyear —points in a consistent direction: an election period of at least 30businessdays; dataroomaccess equiva­ lent to the buyer’s; no increase inmanagement fee or carried interest for rolling investors; crystallised carry reinvestedinthevehicle.Aboveall,therollmustcon­ tain a genuine status quo option — participation in thenewvehiclewithnochange ineconomic terms.A rollofferedonlyonneweconomicsisnotachoicebe­ tweenstayingandleaving;itisanexitdressedasone. Won in the drafting Bythetimethetransactionisannounced,thegovern­ ing questions have usually been answeredyears ear­ lier, in the legacy fund’s partnership agreement. May thesponsortransactwithanaffiliateatall,andsubject towhoseapproval?IstheLPAC’sroleconsentorcon­ sultation?Whobearsthetransactioncosts—andwho bearsthemifthedealfails?Noneoftheseisanswered by theAIFMD; all of themare answeredbydrafting. The CSSF’s interest is not a threatening one. Luxem­ bourg’sframeworkalreadycontainseverythingade­ fensible continuation fund needs — the vehicles, the conflictsmachinery,thevaluationdiscipline.Whatthe regulator’s attentionchanges is the standardof proof. A continuation fund that survives scrutiny is rarely defendedinthedataroom.Itisdrafted,minutedand valued long before anyone calls it a transaction. Continuation funds in Luxembourg: structuring the transaction the regulator is now looking at