How will asset managers fund investment research? Will Commission Sharing Agreements (CSAs) survive? 

We have more clarity on these questions following publication by the European Commission of the long-awaited MiFID II Delegated Directive. Under MiFID II, managers must either pay for investment research themselves or by a separately identifiable charge. Contrary to ESMA’s final advice and a victory to industry lobbying, the Commission’s text allows funding for research to be collected alongside transaction commission. Therefore, subject to national implementation measures, we believe that CSAs can survive in some form, albeit with renegotiation and onerous upfront and ongoing transparency obligations on managers. 


  • Decide whether to pay for research out of the manager’s own funds (and perhaps increase the management fee to cover the additional costs), or utilise a research payment account (RPA).
  • If using an RPA, ensure the RPA meets the required eligibility conditions, and consider how clients will fund the RPA (whether separately from other charges or together with commission).
  • For RPAs, implement a policy for valuing research, allocating its value between portfolios and assessing a research budget using quality criteria, an audit trail and senior management oversight.
  • Review client reporting and information requirements, including in terms of business and prospectuses.
  • Renegotiate CSAs where appropriate to ensure compliance with RPA set-up.


We charted the detailed background to the debate around inducements and the payment for research in an earlier paper: MiFID II, Payments for Research: Unbundling the Confusion

In brief, under MiFID II, managers must not receive investment research from a third party unless it falls within narrow exceptions (such as certain generic research). 


The provisions in MiFID II apply to MiFID portfolio managers. However, they are also relevant to the provision of MiFID services by alternative investment fund managers (AIFMs) and UCITS managers (for example, portfolio management and investment advisory activities). In addition, when implementing MiFID II provisions in the UK, it is likely that the Financial Conduct Authority (FCA) will gold-plate its rules where possible, so that there is a levelplaying field for all asset managers.


To avoid the receipt of research being considered an unacceptable inducement, a manager must decide whether to:

  • pay for research out of its own resources (whether funded itself or by increasing the annual management charge if this is a realistic option); or
  • pay for research from a separate research payment account, which meets detailed conditions (considered below).


Some managers have already publicly declared that they will pay hard for research, however, it is too early to say whether this is a trend.


An RPA is an account controlled by the manager but which is funded by a specific research charge to each client. The manager pays for previously budgeted investment research from funds in the RPA. If a manager wants to pay for research using an RPA, the account must satisfy the following conditions:

  • The RPA is funded by a specific charge to the client. The manager sets this charge based on a research budget which it reasonably assesses will meet its need for third party research in providing services to its client(s). The research budget and client research charge must not be linked to the volume and/or value of transactions executed on behalf of the clients. Also, the total research charges received must not exceed the research budget. The manager must periodically assess the research budget.
  • The manager must regularly assess the quality of research purchased with funds from the RPA. It should base this assessment of the research on robust quality criteria and its ability to contribute to better investment decisions. To do this, a manager must draw up, and provide to its clients, a written policy. The written policy must address the extent to which the research benefits the client’s portfolio and the approach the manager takes to allocate costs fairly to the various clients’ portfolios and between funds where a number of portfolio managers are benefitting from the research. This requirement may limit, for example, a manager imposing a flat regular fee on each customer
  • A manager must not use the research budget or RPA to fund its own research. There must be appropriate controls and senior management oversight over the allocation of the research budget to ensure it is used in the best interests of the manager’s clients. Such controls must include a clear audit trail of payments made to research providers, and how the amounts paid are determined with reference to the quality criteria. UK managers already should have in place systems and controls to avoid conflicts of interest, particularly following a specific attestation request from the Financial Services Authority on the issue in November 2012. However, these new requirements are much more prescriptive, therefore managers need to review their existing controls to ensure they are MiFID II compliant. 
  • Meet the disclosure points referred to in ‘What information does a manager have to give the client?’ below.


The total research charges received must not exceed the research budget. Therefore, managers should meet any shortfall out of their own funds as passing on this charge to clients is unlikely to be considered reasonable by the client or the regulator. Managers will need systems and controls in place to ensure they do not charge clients in excess of the agreed research charge and budget. 

The manager must rebate to the client(s) a surplus balance leftover at the end of a year, or it must offset the surplus against the research budget/charge for the following period.


The manager is responsible for the RPA. However, it may delegate the administration of the RPA to a third party so long as the delegate makes transactions in the name of the manager without undue delay and in accordance with its instructions.


The Commission’s text envisages that it remains possible to collect the research charge alongside a transaction commission. However, where a manager does this, the fee must indicate a separately identifiable research charge and still use an RPA and comply with all the conditions which apply to payment through an RPA (discussed above). Therefore, while this provision appears to permit the continued usage of CSAs, the manager must identify the research charge in advance and its use will be subject to stringent quality assessment criteria, transparency and reporting obligations. Additionally, to be compatible with the RPA requirements, existing CSAs will require amendment and/ or renegotiation in a number of respects.


The manager must agree the research charge as budgeted with the client, in its terms of business or investment management agreement. It should also agree the frequency of deduction of the charge over the year. However, in a departure from ESMA advice, the client’s agreement is not required for increases in the research budget. Instead, managers must give clients clear information about intended increases beforehand. Presumably this gives clients the opportunity to move their business should they be dissatisfied and a move is practicable.


Managers must consider a number of new notification obligations including:

  • In the terms of business or investment management agreement, the research charge, as budgeted, and the frequency of deduction over a year. A manager must also give the client a copy of its written policy on investment research, and clear information about intended increases to the research budget beforehand.
  • Before providing a service, an estimated research charge for that client and information about the budgeted amount for research. In the context of an AIFM or UCITS manager, this is likely to require additional prospectus disclosures. 
  • Annual information on the total cost the client has incurred for third party research.
  • Managers must be able to produce (on request) for the client or the regulator, a summary of certain information about the funds into and payments out of the RPA, providers paid, benefits received and how the RPA spend compares to the budget. 


Subject to further European approvals the requirements will apply from the date of implementation of MiFID II – now likely to be 3 January 2018, assuming the well-publicised delays materialise. 


It is a victory for the industry that it remains possible to collect a research charge from a client together with an execution fee. However, as the new European rules currently stand, unbundling takes place in all but the administrative payment process and managers choosing the RPA path have the dilemma of how to attribute the value of the research and evidence this assessment, as well as meet the increased transparency requirements. While the text is not yet in final approved form, it is unlikely to change substantially. Therefore managers should watch the market and begin to formulate their preferred direction on how they intend to pay for (and fund) research in the MiFID II world.