The 2026 MiFID II research re-bundling: setting up the new disclosure workflows

This article is for informational purposes only and does not constitute legal advice. Consult a qualified legal professional for advice specific to your situation.

What actually changed, and what did not

MiFID II's original research unbundling regime required investment firms to pay for third-party research separately from execution commissions. The stated aim was to remove a conflict of interest: when research is bundled into commissions, a firm has no direct incentive to control research spend, because the cost is invisible to the client and disconnected from trading volume.

In practice, unbundling produced a side effect regulators had not fully anticipated. Standalone research pricing did not work economically for coverage of small and mid-cap issuers, and sell-side firms scaled back or eliminated research desks covering that segment. The European Commission acknowledged this directly in its own review, noting that unbundling had reduced research coverage of smaller companies at a time the EU was trying to deepen retail and institutional participation in capital markets.

The 2021 Capital Markets Recovery Package took a first, narrow step back: joint payments were permitted, but only for research on issuers with a market capitalisation below EUR 1 billion over the preceding 36 months. Directive (EU) 2024/2811, part of the Listing Act package, removes that threshold entirely. Joint payments for execution and research are now available for research on issuers of any size, at the investment firm's discretion.

What has not changed is the underlying default. Separate payment, either through a Research Payment Account (RPA) funded by a specific research charge or paid directly by the firm out of its own resources, remains fully available and remains the required approach for firms that do not opt into joint payment. Nothing in the Listing Act removes the RPA model or grandfathers firms into bundling. The reform is additive: it restores optionality that unbundling removed, rather than reversing unbundling as a default.

The legal instrument that actually governs the operational detail

The Listing Act amended Article 24(9a) of MiFID II at level 1, but the operational conditions live in the level 2 delegated act. Article 13 of Commission Delegated Directive (EU) 2017/593, the MiFID II Delegated Directive, previously set out the conditions applicable to the unbundling regime only. It needed amendment to accommodate an optional joint payment route without simply reopening the pre-MiFID II conflict of interest problem.

ESMA ran a consultation on the amendments from 28 October 2024, closing 28 January 2025, and delivered its final technical advice to the Commission on 8 April 2025. The Commission adopted the resulting instrument, Commission Delegated Directive (EU) 2026/374, on 20 February 2026, and it was published in the Official Journal on 2 June 2026. Member states had until 5 June 2026 to transpose it into national law, with the new rules applying from 6 June 2026.

This sequencing matters for how firms should read the compliance deadline. The reform is frequently described in the market as "the 2026 changes," which understates how tight the runway actually was between adoption and application. Firms that waited for the Official Journal publication in early June 2026 to begin operational work had roughly four days before the rules applied.

What actually applies to joint payment, and what stays specific to the RPA

The amended framework is easy to misread as imposing one symmetric governance structure on both payment methods. It does not. Reading Article 24(9a) of MiFID II (as amended by the Listing Act, Level 1) alongside Article 13 of the Delegated Directive (as amended by Commission Delegated Directive (EU) 2026/374, Level 2) side by side, the obligations split cleanly into two groups: a small set that applies whichever payment method a firm uses, and a larger, more detailed set that is scoped specifically to firms operating a separate research payment account (RPA).

What applies regardless of payment method

Four obligations sit at Level 1, in the amended first subparagraph of Article 24(9a), and apply whether a firm pays jointly or separately.

The remuneration methodology agreement. An agreement must be in place between the investment firm and the third-party provider of execution services and research, establishing a methodology for remuneration, including how the total cost of research is generally taken into account when setting the firm's overall charges. This applies to the relationship itself, not to a specific payment mechanic, so it holds for joint payment arrangements as much as for RPA-funded ones.

Client disclosure of the payment-method choice. The firm must inform clients whether it pays jointly or separately for execution and research, and make available its policy on payments for execution services and research, including what type of cost information clients can expect depending on the method chosen and, where relevant, how conflicts of interest arising from joint payment are managed.

The annual quality assessment. Article 24(9a), point (c) requires firms to assess, at least annually, the quality, usability, and value of the research they use, and its contribution to better investment decisions. This is the Level 1 basis for Article 13(10) of the Delegated Directive, which is the only paragraph of Article 13 explicitly stated to apply "irrespective of how investment firms pay for execution and research services." It requires the assessment to be based on robust quality criteria and requires remedial action, which can include requesting improved research from the provider, discontinuing use of low-quality research, or switching provider, where the assessment identifies shortcomings.

A lighter cost-disclosure duty. Separately from the detailed RPA reporting described below, Directive 2024/2811 adds a general requirement that, where known to them, investment firms keep a record of the total costs attributable to third-party research provided to them, and make that information available to clients annually on request. This is a real disclosure obligation for joint-payment firms, but it is narrower than the budget-versus-actual, provider-by-provider reporting that applies under the RPA.

What is specific to firms operating a separate research payment account

The detailed governance architecture in Article 13 of the Delegated Directive, paragraphs 1 through 7, is scoped explicitly to investment firms that "operate a separate research payment account." A firm using joint payment is not brought within these paragraphs by the 2026 amendment; they remain the RPA-specific regime, refined rather than extended.

The research budget obligation. Paragraph 1(b) requires firms operating an RPA to set and regularly assess a research budget as an internal administrative measure. This obligation predates the 2026 amendment in substance and continues to apply to the RPA model specifically; it is not a new requirement imposed on joint payment.

Disclosure of budgeted and actual costs, and provider-level transparency on request. Paragraphs 1a and 2 require RPA-operating firms to give clients, before providing services, information on the budgeted research amount and estimated research charge, and, on request, a list of the specific providers paid from the account, the amounts paid to each, benefits received, and how actual spend compares to budget. This provider-by-provider, budget-versus-actual reporting is an RPA feature; it does not apply in the same form to joint payment, which instead carries the lighter "total costs known to the firm" disclosure described above.

The estimated research charge cannot be linked to transaction volume or value. This constraint, in paragraph 2, attaches specifically to the RPA's "estimated research charge." It is not, on the text, a general prohibition reaching joint payment pricing structures. Joint payment's separate safeguard against the equivalent conflict is the Level 1 remuneration methodology agreement described above, which is a different mechanism aimed at the same underlying concern.

Senior management oversight and the audit trail. Paragraph 6 requires RPA-operating firms to maintain senior management oversight of the research budget allocation process, a documented audit trail of provider payments, and a prohibition on using the research budget or RPA to fund internal research. This is an RPA governance requirement; the amended Article 13 does not impose an equivalent formal budget-and-audit-trail structure on joint payment, which has no analogous "research budget" concept in the text.

The practical effect is that the 2026 reform does not equalise the compliance burden between the two models in the way the budget-and-disclosure architecture might suggest at first read. Joint payment carries the Level 1 obligations above, which are real but comparatively light; the RPA carries the fuller, more prescriptive Level 2 regime it always has, now updated to sit alongside the joint payment option rather than being the only compliant route to separate payment.

What this means operationally for firms deciding between the two models

Because the governance burden is not symmetric, the practical decision facing asset managers and other buy-side firms is not simply "which model is administratively lighter." The RPA carries a materially more detailed compliance architecture than joint payment does under the amended framework, and that asymmetry itself is a relevant input to the choice.

For firms with an existing, functioning RPA infrastructure, the case for switching to joint payment turns mainly on whether the RPA's detailed provider-level tracking and budget governance was serving a purpose the firm values, such as granular internal visibility into research spend, rather than on whether joint payment offers equivalent transparency by default, since it does not carry the same built-in reporting layer.

For firms that have found the RPA model administratively burdensome, particularly smaller managers without dedicated research payment operations, joint payment genuinely does reduce the compliance load, not just the payment mechanics, since the detailed Article 13 paragraphs 1 through 7 do not apply. What joint payment still requires is the Level 1 obligations common to both models: the remuneration methodology agreement, client disclosure of the payment method chosen, the annual research quality assessment, and the lighter total-cost record kept where known.

For firms operating across multiple strategies or client types, a mixed approach is explicitly permitted. Nothing in the amended framework requires a single firm-wide choice. What it does require is that the choice per scope, and its rationale, be documented and defensible, and that firms track which detailed obligations apply where, since the two models are not interchangeable on compliance architecture even though clients can be served under either.

Summary of the key dates

| Milestone | Date | | ---------------------------------------------------------------------------------- | ---------------------------------- | | Capital Markets Recovery Package introduces EUR 1bn threshold joint payment option | 2021 | | Listing Act (Directive (EU) 2024/2811) adopted, removing the threshold | 2024 | | ESMA consultation on amendments to Article 13 of the Delegated Directive | 28 October 2024 to 28 January 2025 | | ESMA final technical advice delivered to the Commission | 8 April 2025 | | Commission Delegated Directive (EU) 2026/374 adopted | 20 February 2026 | | Published in the Official Journal | 2 June 2026 | | Member state transposition deadline | 5 June 2026 | | New rules apply | 6 June 2026 |

For RPA-operating firms, the transparency logging requirement is the part most likely to be underestimated

Of the obligations sitting under the amended Article 13, the one most likely to be treated as a documentation afterthought by firms continuing to operate an RPA is the requirement to make provider-level payment data, budget-versus-actual comparisons, and benefit disclosures available to clients on request. This is not a report that can be generated once a year and filed. It is a standing capability: a client can ask at any point, and the firm needs to be able to produce accurate, current data.

For RPA-operating firms running research payment accounts across a large number of client mandates and providers, this effectively requires a research spend ledger that is maintained continuously and reconciled against the budget on a schedule tight enough that a client request does not surface a data gap. Firms that treat this as a compliance reporting exercise, built at year end, are likely to find themselves unable to respond to an in-year client request with defensible numbers.

Joint payment firms face a lighter version of this problem: the total-cost record they must keep where known, and disclose to clients annually on request, is a simpler figure than the RPA's provider-by-provider, budget-versus-actual breakdown, but the 'where known to them' qualifier means the duty is best-efforts rather than absolute. Firms should nonetheless treat it as an ongoing record rather than something reconstructed only when a client asks, since a gap at the point of request is hard to defend.

What firms should be doing now

For firms that have already made the joint versus separate payment decision, the immediate task is confirming that the obligations actually attaching to their chosen model are operating in practice, not just documented in policy. RPA-operating firms need the budget-setting, provider-level tracking, and audit trail processes under Article 13 paragraphs 1 to 7 running with evidence, not just described. Joint payment firms need the Level 1 obligations, the remuneration methodology agreement, client disclosure of the payment method, the annual quality assessment, and the total-cost record kept where known, actually in place. A policy that describes required controls without evidence that they are running is a gap a national competent authority will identify quickly, whichever model is in use.

For firms that have not yet decided, the decision should account for the fact that the two models carry materially different compliance architectures, not comparable ones. The RPA brings more prescriptive, provider-level reporting; joint payment brings a lighter but still real set of Level 1 obligations. The choice is genuinely about which architecture, not just which payment mechanic, a firm wants to operate.

For firms operating third-party research relationships across multiple providers, the practical bottleneck under either model is likely to be the remuneration methodology agreement, specifically demonstrating that it does not create a best-execution conflict and, for RPA-operating firms, that the estimated research charge is not linked to transaction volume or value. That documentation should exist before a client or a competent authority asks for it, not after.

For the broader MiFIR review context in which this reform sits, including the consolidated tape and market data pricing changes, see the EU consolidated tape: how MiFIR amendments change market data sourcing. For how EU directives and their delegated acts interact more generally, see how EU financial regulation actually works.

Forseti monitors MiFID II delegated acts, ESMA technical advice, and transposition deadlines continuously, anchored to verified official sources including EUR-Lex and ESMA publications. Start for free.