Q1 Do you agree that greater standardisation in the field of funds’ performance fees is desirable? What should be the goal of standardisation?
SIFA believes there is a call for a minimum level of consumer protection when it comes to performance fee models. The goal should be to set a minimum standard of principles to ensure fair treatment of unit-holders. We believe too much details, or only accepting one model, should be avoided at this point.
A fair performance fee model has the goal to align the interests of the manager and the investor. It is important that such models are available not only to institutional investors but also to retail, although the administration of performance fee models in daily traded retail funds has particular challenges. It is therefore necessary to have some flexibility to use different and suitable models, as long as they fulfil the overarching requirement to be fair and align the interests of the manager and the investor.
Q2 Are there any obstacles to standardisation that could be removed by regulatory action? Please elaborate.
In order for the guidelines to serve their purpose of ensuring supervisory convergence they need to be clarified with regards to the practical implementation of crystallisation periods which differ from subscription/redemption periods. As we have come to understand it only over-performance that is still in place by the end of a crystallisation period (i.e. on the crystallisation date) should be the basis for charging performance fees. This would mean that the performance fee is calculated and accrued in the NAV daily when the fund is over-performing, and then “paid back” as soon as the fund is under-performing.
This raises questions on the equal treatment of unit-holders which have to be addressed and further explained. Is the pay-back of performance fees supposed to go the fund, i.e. to all unit-holders at that point in time, or only to the unit-holders that have been charged performance fees during the crystallisation period? We believe that the latter is not practically feasible in daily traded retail funds. Unit-holders who subscribe to the fund would then have to pay an extra amount corresponding to the reserved accumulated performance fee charged per unit in order to be treated like the existing unit-holders. Furthermore, the issues that arise are not related to single investors but rather to single investments of single investors. The client would need to provide the asset manager with information about which of the previously purchased units the client is now selling. See further our answer to Q4.
The reference period, which is not set in the proposed guidelines, will normally end at the same time as the crystallisation date, allowing the manager to reset the highwatermark (HWM). In jurisdictions that today for consumer protection reasons have a more far-reaching practice when it comes to performance fees it is unclear whether the guidelines are supposed to hinder such practice. We believe that would be counter-productive why we suggest clarifying the guidelines in order not to become an obstacle to higher standards, i.e. an eternal HWM. See further our answer to Q5.
We also believe that it is not absolutely clear whether the guidelines aim to capture all variable fees, or whether there could be other types of variable fees not considered “performance fees”. In order to avoid different interpretations, which per se is an obstacle to standardisation, the notion of “performance fee” would deserve a specification.
Q3 What should be taken into consideration when assessing consistency between the index used to calculate the performance fees and the investment objectives, strategy and policy of the fund? Are there any specific indicators which should be considered (eg: historical volatility, asset allocation composition, etc.) to ensure this consistency? Please provide examples and give reasons for your answer.
We agree that there should be consistency between the fund’s investment objectives, strategy and policy of the fund and the benchmark used for calculating the fee. For example, if a fund takes on market risk (beta) this should be reflected in the choice of benchmark. Different types of benchmarks could be considered consistent with a fund’s objectives. For example, if a fund has a beta of 1,0 versus investment grade bonds, some investors may prefer to pay a 0,1 % management fee + 20 % performance fee from a zero benchmark (HWM applied), rather than paying 0,50 % management fee + 20 % performance fee from an investment grade bond benchmark. Another example could be an equity fund that applies a long-term average equity market return as a benchmark for calculating performance fees (e.g. risk-free rate + 6 % p.a.) instead of a standard equity benchmark. Funds pursuing an absolute return objective would normally calculate its performance fee against a hurdle rate not necessarily representing the funds’ investment universe. It is thus important that the guidelines are principle-based and allow for some flexibility when it comes to the choice of benchmark. Basically, the benchmark should be appropriate for measuring how the fund manager is performing in order to align the interests of the manager and the investor.
Q4 What is the anticipated impact of the introduction of Guideline 3? Do you agree with setting a minimum crystallisation period of one year? Do you think this could help better aligning the interests of fund managers and investors? Please provide examples.
We are strangers to this model since in Sweden most managers apply an eternal HWM, meaning that the HWM generally could not be reset. We do not believe that it is justified to charge long-term investors with performance fees for the same over-performance more than once.
In the absence of a requirement for an eternal HWM we believe that a crystallisation period of one year could benefit the investors to some degree. However, such crystallisation periods are mainly suitable for funds or mandates where performance fees are calculated and charged on the basis of each investor’s individual investments. For retail funds which are traded on a daily basis and often registered with intermediaries such an individual model is not feasible in practice, and performance fees are therefore charged on a collective basis (see our answer to Q2). The use of crystallisation periods for UCITS therefore brings on certain problems. Since fees will be reserved at certain points and then “paid back” to the fund during the crystallisation period it will affect the allocation of fees among investors. For example if a fund reserves performance fees for the first six months of the period and then depreciates in value back to the starting level during the next six months, an investor coming into the fund after six months will benefit from the reserved performance fee being “paid back” into the fund (although the investor has not been part of “paying” any performance fee). Further, the investors who were in the fund the whole time will have to split their reserved performance fee with the new investors. There is a clear risk that some investors will take advantage of this situation at the expense of long-term investors. In this sense the model does not ensure fair treatment among investors. We lack argumentation on this fact, as well as on how investors benefit from the fund manager being able to reset the performance reference period (the HWM).
Q5 Are there any other models or methodologies currently employed that, in your view, should be exempted from this requirement? For example, do you think that the requirement of a minimum crystallisation period of 12 months should also apply to HWM models? Please provide examples on how these models achieve the objectives pursued by Guideline 3.
SIFA strongly believes that models with eternal HWM should be exempt from the minimum crystallization period requirement.
We believe a model with an eternal HWM, ensuring that investors – especially long-term investors – do not have to pay for the same over-performance more than once, is more fair to investors. We understand that at this point it may be difficult to implement such a model in all Member States and that a minimum standard set in the guidelines could be a good starting point. However, it is important that the guidelines do not actually hinder an eternal HWM model, for example in jurisdictions that have such a standard today.
It has been argued that a requirement of eternal HWM would lead to managers taking on excessive risk (to reach a new HWM) or liquidating funds (when the HWM seems difficult to reach). Firstly, we would like to point out that the risk level of a UCITS is limited by law, the fund rules as well as the funds’ prospectus and internal investment guidelines. In addition, there is a legal obligation to act in the best interest of the unit-holders. Secondly, in our experience the method on the contrary incentivises the fund manager not to take excessive risks that might result in losses, since any such losses will then need to be offset before any performance fee can be levied again. Liquidating funds that are underperforming is normally not an option if the manager wants to remain in the market. In Sweden the method has been in practice for a very long time and we have still not seen any evidence of neither excessive risk-taking nor excessive liquidation of funds. Our experience is that such a model creates a good long term alignment between the interests of the investors and the asset managers. On the other hand, resetting the HWM every year, even when a minimum crystallisation period of one year is applied, could incentivise the manager to take excessive risk each year to maximise the expected performance fee.
It is therefore important that the guidelines recognises a performance fee model using an eternal HWM as an alternative to the use of a crystallisation period. The eternal HWM model is not compatible with a crystallisation period. Should the crystallisation period model be introduced as mandatory it would prevail over other models since the application of both a one year crystallisation period and an eternal HWM would not be practically feasible. We are worried this would have a negative effect on retail investors who in our view should have access to the same funds with fair performance fee models that are requested by institutional investors.
In IOSCO:s final report on Good Practice for Fees and Expenses of Collective Investment Schemes (FR09/16) it is argued that, as an alternative to crystallising the performance fee, methods to ensure that cumulative gains are offset in some way by cumulative losses can be considered. Examples of relevant methods include the highwatermark and high-on-high, which require an absolute improvement in investment performance before the performance fee can be paid. Such methods incentivises the CIS operator not to take excessive risks that might result in losses, since any such losses will then need to be offset before any performance fee can be levied again (page 9, point 35). SIFA agrees.
Guideline 3 (point 21) should therefore be modified accordingly.
The minimum crystallisation period should not apply to the fulcrum fee model, nor to a model where the HWM may not be reset (eternal HWM), as the characteristics of such models are not compatible with a minimum crystallisation period.
Q6 In your view, should performance fees be charged only when the fund has achieved absolute positive performance? What expected financial impact (e.g. increase or decrease of the manager’s remuneration or increase or decrease of the financial return for investors) would the proposed Guideline 4 have for you/the stakeholder(s) you represent? Are there models or methodologies currently employed where the approach set out in Guideline 4 would not be appropriate?
No. It would lead to only absolute return funds being able to have performance fees. With a relative reference the aim is to outperform the market regardless of the market going down. An investor investing in an equity fund is prepared to pay performance fee if the fund manager generates returns above the market index (benchmark). The investor benefits from this excessive return even if the market has gone down. It would be contrary to Guideline 2 if an equity fund is not able to charge its performance fee against its benchmark.
Q7 If the performance fee model that you currently use provides for performance fees to be payable in times of negative returns, is a prominent warning on this provided to investors in the legal and marketing documents of the fund? If not, should this be provided? Please give examples for your answer and details on how the best interests of investors are safeguarded.
We believe that information about the fact that the performance fee may be levied even in times of negative returns could be clarifying to retail investors. Such information should preferably be provided in the funds’ prospectus. It should also be provided in marketing documents where the fee is mentioned. The latter addition is important since fund marketing could take place in for example short banners where only the name of the fund is mentioned, and there is simply no room to add full information on fees or other aspects of the fund.
Q8 What are your views on setting a performance reference period for the purpose of resetting the HWM? What should be taken into account when setting the performance reference period? Should this period be defined, for example, based on the whole life of the fund (starting from the fund’s inception date), the recommended holding period of the investor or the investment horizon as stated in the prospectus? Please provide examples and reasons for your answer.
SIFA sees no convincing arguments on how an investor could benefit from a reset of the HMW or having a performance reference period that is shorter than eternal, i.e. since inception of the fund or unit-class. An investor should not have to pay multiple times for the same performance. It does not matter how long the recommended holding period is, if the fund manager has not given any (new) value to the investor it should not be able to charge a performance fee.
Q9 Alternatively, would it be possible to envisage predefined time horizons for the purpose of resetting the HWM, such as 3 or 5 years? Please provide examples and details on what you think would be the best practice in order to better align the interests of fund managers and investors.
No, see Q8 answer.
Q10 How long do you think the performance reference period should be for performance fee models based on a benchmark index? What should be taken into account when setting the performance reference period for a performance fee benchmark model? Would it be possible to envisage predefined time horizons for the purpose of resetting the performance fee based on a benchmark, such as 3 or 5 years? Please provide examples and details on what you think would be the best practice in order to better align the interests of fund managers and investors.
Eternal, see Q8 answer.
Q11 Alternatively, do you think the performance reference period should coincide with the minimum crystallisation period or should it be longer/shorter? Please provide examples and reasons for your answer.
See Q8 and Q5 answers.
Q12 What are your views on when the Guidelines should become applicable? How much time would managers require to adapt existing fee mechanisms to comply with the requirements of these Guidelines?
For new funds SIFA believes that it should be linked to the time the local NCAs should report back to ESMA on whether they will comply (i.e. two months after publication).
For existing funds at least 12 months would be necessary. ESMA must also establish how unit-holders should be notified of such a change of the fee model. Notifying all unit-holders in writing is very costly and should, if such notification is deemed necessary, be duly accounted for in the cost-benefit analysis of the guidelines. If ESMA believes that notification is not necessary – for example because the change is deemed to be in the interest of the investor – and that a change of the prospectus and KIID is enough, this should clearly be stated in order to ensure a level-playing-field.
However, where a change of the fund rules, duly authorised by the NCA, is necessary to comply with the guidelines, we cannot see how NCAs could demand such change without legal support.
Q13 Do you consider that the principles set out in the Guidelines should be applied also to AIFs marketed to retail investors in order to ensure equivalent standards in retail investor protection? Please provide reasons.
Yes. Retail investors should be equally protected whether they invest in UCITS or AIFs. Having said that, we are hesitant to whether there is legal ground for ESMA to establish guidelines for retail AIFs based on the AIFMD.
Q14 Do you agree with the above-mentioned reasoning in relation to the possible costs and benefits as regards the consistency between the performance fees model and the fund’s investment objective? What other types of costs or benefits would you consider in this context? Please provide quantitative figures, where available.
In general we agree. It is important that the guidelines are principle-based and allow for some flexibility when it comes to the choice of index/benchmark. Normally several benchmarks could be considered appropriate. Fund managers should for example not be forced to change the chosen benchmark several times because there is one considered “more appropriate”, or be forced to choose a specific benchmark that could be very expensive.
Q15 In relation to Guideline 2, do you think that models of performance fee without a hurdle rate, or with a hurdle rate not linked to the investment objective (but clearly stated in the offering documents), should be permissible? For example, do you think that equity funds with a performance fee linked to EONIA, or a performance fee which is accrued as long as there are positive returns, should be allowed? Please give examples and reasons for your answer.
Decisive should be the consistency between the investment objectives, strategy and policy of the fund and the benchmark used. You could argue that an equity fund could use for example EONIA + equity risk premium or a fixed hurdle rate.
Q16 What additional costs and benefits would compliance with the proposed Guideline bring to you/the stakeholder(s) you represent? Please provide quantitative figures, where available.
Costs for switching from an eternal HWM model without a crystallisation period requirement to a model with a resettable HWM but a crystallisation period requirement will be substantial. Not only in terms of new IT systems, new documentation etc. but also in relation to the investors. Investors who have invested in the funds partly because of the eternal HWM fee model and are satisfied with the outcome would find it hard to accept such a change of fee model. Since the crystallisation period method cannot be considered to be a better method from an investor protection point of view, but rather another method, it would be disproportional to require such a switch. From a unit-holder’s perspective, a switch from eternal HWMs to a model with one year crystallisation period and yearly resettable HWMs would substantially increase the amount of performance fees that unit-holders are charged.
If ESMA believes that is it not necessary to notify all unit-holders of the changes in order to comply with the Guidelines it should be stated specifically. If, on the contrary, such notification is deemed necessary, the costs should be duly accounted for in the cost-benefit analysis of the guidelines. Such costs would include sending letters to all unit-holders. If a change of the funds duly authorised fund rules is required there would be quite heavy costs including for a new authorisation with the NCA as well as a notification with all unit-holders.
It would be beneficial to investors if an indication of the expected outcome of the performance fee model was made available to them.
Q17 What is the anticipated impact from the introduction of this proposed Guideline? Are there models or methodologies currently employed where this Guideline would not be appropriate? If so, please provide examples of these and details of how the best interests of investors are safeguarded.
In general it is positive that there are some basic rules for performance fees. However, from our perspective, having an eternal HWM/performance reference period, it is important that this model is exempted from the crystallisation requirements (se Q5) in order to make this Guideline appropriate. We believe that an eternal performance reference period is in the best interest of the investor and it would be counterproductive switching to a model where HWM could be reset. Since eternal HWM will not in practice be combined with a crystallisation requirement, this would be the outcome should the method not be exempted and seen as an alternative to the crystallisation requirement.
Q18 What additional costs and benefits would compliance with the proposed Guideline bring to the stakeholder(s) you represent? Please provide quantitative figures, where available.
See answer to Q16.
Q19 Which other types of costs or benefits would you consider in the disclosure of the performance fees model? Please provide quantitative figures, where available.
Disclosing relevant information in the legal documentation would not lead to significant additional costs. When it comes to other marketing documents, for the avoidance of doubt and compliance costs, it should be added that this applies to marketing documents where the fee is mentioned. Such documents could be used as an alternative to the legal documents and should therefore contain all necessary information. When it comes to other marketing communications – as long as a fund is mentioned in any communication it would be considered “marketing” – such as a banner, which does not include specific information about the funds features, information on the performance fee model should not be required.
SWEDISH INVESTMENT FUND ASSOCIATION
Helene Wall
General Counsel