Discussion Paper on Key Information Documents for Packaged Retail and Insurance-based Investment Products (PRIIPs) [JC/DP/2014/02]

Transcription

1 Discussion Paper on Key Information Documents for Packaged Retail and Insurance-based Investment Products (PRIIPs) [JC/DP/2014/02] 1.7 Interaction with other EU legislation [p. 15] 1. Do you have any views on how draft RTS for the KID might be integrated in practice with disclosures pursuant to other provisions? The work on draft RTS is an important opportunity to align the PRIIPs Regulation with MiFID II, IMD I and II and UCITS so that disclosure requirements are coherent and consistent across different legislative frameworks. A good example of how this might work in reality is presented in MiFID II, where the Directive envisages the disclosure of costs and charges information in the KID as a sufficient measure of product costs for inclusion in the calculation of an aggregated costs and charges figure covering the whole investment service. The worst possible result would be RTS that contradict or fail to take into account disclosure requirements in other Directives and Regulations, or which lead to different disclosures having to be presented to the client at the point of sale. 3 What are the risks and what could I get in return? [p. 21] 3.3 Definition of risk and reward [p. 23] 2: Do you agree with the description of the consumer s perspective on risk expressed in the Key Questions? We agree with the description of the consumer s perspective on risk expressed in the Key Questions. 3: Do your agree that market, credit and liquidity risk are the main risks for PRIIPs? Do you agree with the definitions the ESA s propose for these? We agree with the ESAs proposal to classify market, credit and liquidity risk as the three main risks for a PRIIP. In terms of market risk, we encourage the ESAs to build on the experience of the UCITS SRRI which has been in place since It is essential to enable off-balance sheet products such as funds to highlight that they provide little or no credit risk as the investor is not exposed to the product provider itself. When considering liquidity risk, the ESAs should also consider the following factors: - the surrender value of a PRIIP during different stages of its life cycle (eg. a life insurance policy s surrender value is most likely negative in the first couple of years);

2 - any time delay between a request to redeem the investment and its effective execution (in some cases it might take several months between requesting an early pay-out for products with a fixed or long-term product horizon and the consumer receiving the proceeds); - we do not agree with the ESAs assessment that not being listed necessarily implies any relevance to a PRIIP s liquidity. From a fund perspective, the terminology used in this Discussion Paper differs considerably from the notion of market and liquidity risk applied for the purpose of risk management and described in the fund prospectus in accordance with the UCITS Directive. It is vital that this does not lead to information disclosed in UCITS prospectuses contradicting statements in the PRIIPs KID, as this would be very confusing for investors. Some of the language in the table on page 23, which uses wording like How much can I win?, is completely inappropriate for an investment product. We would suggest that something like How much am I likely to get back? would be more appropriate. 3.4 Measuring risks [p. 27] 4: Do you have a view on the most appropriate measure(s) or combinations of these to be used to evaluate each type of risk? Do you consider some risk measures not appropriate in the PRIIPs context? Why? Please take into account access to data. We have doubts that market, credit and liquidity risk can be integrated into a meaningful single risk indicator and therefore believe separate risk indicators would be more helpful to investors. Liquidity risk should be based on the specific characteristics of the exit arrangements for the PRIIP. The absence of a secondary market for a PRIIP does not automatically result in a high liquidity risk. Most investment funds offer daily redemptions at a price calculated on the basis of the fund s NAV. Such redemption opportunities should qualify as organised liquidity facilities, but cannot be assessed on the basis of the proposed quantitative criteria, which reflect characteristics of secondary market trading. 3.5 Aggregation of risk [p. 29] 5: How do you think market, credit and liquidity risk could be integrated? If you believe they cannot be integrated, what should be shown on each in the KID? Market, credit and liquidity risks should not be integrated into a single risk indicator. Such an approach would lead to an oversimplified risk measure and provide unhelpful information, disguising (through aggregation) risks which might be especially relevant to any particular investor. If it is decided to use a single risk indicator, the ESAs should address these problems caused by aggregation so as to allow for retail investors to be informed of the underlying differences in the risk characteristics of different products. Also, if there must be a single indicator, equal weighting should 2

3 be given to each of the three risk components in any aggregation, as any other approach would appear to be arbitrary. As regards presentation, we urge the ESAs to take account of consumer testing of the UCITS KIID, which clearly demonstrated consumer preference for a scale-based indicator. 3.6 Performance scenarios [p. 30] General approach and methodology [p. 31] 6: Do you think that performance scenarios should include or be based on probabilistic modelling, or instead show possible outcomes relevant for the payouts feasible under the PRIIP but without any implications as to their likelihood? A clear distinction needs to be made between risk indicators and performance scenarios. For investor comprehension, it is essential not to overlay too many pieces of information, each serving a different purpose, into a single section of the KID. Secondly, there is a need to be clear about what these performance scenarios aim to show. These can be either the cumulative effect of costs on a reduction in yield basis or the effects of market volatility using historical/statistical data, or even a combination of the two. The answer to this question will greatly influence the conclusions drawn in the upcoming technical work. 7: How would you ensure a consistent approach across both firms and products were a modelling approach to be adopted? Due to varying cost structures across different types of PRIIPs, it is essential that the performance scenarios should be portrayed net of cost, in order to allow retail investors to assess the impact of costs on their returns Time frame and holding period [p. 32] 8: What time frames do you think would be appropriate for the performance scenarios? Given the wide range of PRIIPs available on the market, we would suggest to consumer test five time frames that should allow for a consistent approach across products: (1) 1 year (2) 3 years (3) 5 years (4) 10 years (5) Product lifetime/recommended (minimum) holding period These time frames should be consistent with the time frames used in the cost disclosure section of the KID. In terms of presentation, we prefer a graph covering a rolling period, similar to the example on the left top of page 43. 3

4 3.6.3 Other aspects of performance to be considered [p. 33] 9: Do you think that performance scenarios should include absolute figures, monetary amounts or percentages or a combination of these? We suggest following the existing UCITS standards (percentages based on net valuation at the beginning and the end of the period), wherever this is sensible for the PRIIP in question. 10: Are you aware of any practical issues that might arise with performance scenarios presented net of costs? Performance scenarios must be presented net of all costs. Not doing so would make it impossible for retail investors to compare expected returns. Investors should be able to see information regarding the PRIIP s prospects without being forced to make their own calculations by referring to other sections of the KID. Moreover, gross scenarios could lead investors to underestimate the effect of costs on performance. Past performance (though far from perfect) is the best available indicator of how a product behaves under certain market conditions. Product costs depend on many factors which cannot all be established in advance, eg. fund volume, performance, trading activities etc., so we support the use of historical data for the purpose of performance presentation. 11: Do you have any preferences in terms of the number or range of scenarios presented? Please explain. No, this is best established during consumer testing, but this section should always carry a warning (similar to the UCITS KIID) to inform investors that these scenarios are based on historic data and/or assumptions about the future and that they cannot be a certain guide to future performance. 3.7 Options for presentation [p. 35] 12: Do you have any views, positive or negative, on the different examples for presentation of a summary risk indicator? Please outline advantages and disadvantages, and provide any other examples that you are aware of that you think would be useful. The summary risk indicator as shown on the top of page 37 of the Discussion Paper will be recognised by many consumers, as it stems from the UCITS KIID, which has been in operation for two years. It is a tested method to present risks to consumers, who are generally positive about this form of presentation. We are in favour of a multi-risk scale where all three risk areas (market, credit and liquidity) are shown by an indicator. If it is decided to integrate the respective individual risks into an overall risk indicator, then all risk elements should be equally weighted for the purposes of the aggregation. In line with the UCITS SRRI, which was consumer tested, we recommend that the indicator be on a scale 4

5 of seven (ie. 1-7). Since some risks are more relevant than others for certain PRIIP types, showing indicators for all risks ensures a balanced presentation. In order to describe the risk-reward profile meaningfully, it is important to make clear to retail investors the link between risk and reward Abstract presentations of performance [p. 40] 13: Do you have any views, positive or negative, on the different examples for presentation of performance scenarios? Please outline advantages and disadvantages, and provide any other examples that you are aware of that you think would be useful. Although past performance per se is not a feature of the KID (as compared to the UCITS KIID), it should nonetheless form a basis for the performance scenarios. Standardised scenarios should allow for enough flexibility in terms of adapting to changing market conditions (after sufficient time for market players to adapt their KIDs). Furthermore, it has to be made sufficiently clear to the retail investor that these conditions are prescribed and based on assumptions about possible future outcomes, along the lines of the current UCITS KIID warning. Among the examples contained in the discussion paper, we prefer the graph displayed on the top of page 43, left-hand side, provided that it is based on a reliable and robust methodology. Performance presentation should be net of costs, in order to present credible outcomes to investors. In this regard, performance scenarios currently required for structured UCITS, which are also presented net of costs, should be considered as a model for further work by the ESAs Combinations [p. 44] 14: Do you have any views on possible combinations of a summary risk indicator with performance scenarios? We have a preference for the combination option 3B as presented in table 9 on page 44. This means that risks and performance should be depicted separately, but within one section of the PRIIPs KID, as foreseen by the Level 1 Regulation. Given the importance of the information on risk and reward, we recommend using the necessary space to show separate risk indicators for market, credit and liquidity risk, as well as multiple performance scenarios. However, it may be appropriate to display multiple performance scenarios in one visual element, as in the example at the top left of page What are the costs? [p. 48] 4.3 Key Questions [p. 51] 5

6 15: Do you agree with the description of the consumer s perspective on costs expressed in the Key Questions? Yes, broadly we agree that the key questions in table 10 describe the consumer perspective and reflect the purpose of the PRIIPs Regulation, which is to enable retail investors to understand and compare PRIIPs (Article 1). We note that one key question is about knowing the cost, one is about comparing the costs and the other six are about understanding the costs. This demonstrates that understanding is more than simply knowing the overall cost and is significant in that it is indicative of the importance of providing sufficient granularity about the different types of cost of each PRIIP. This is consistent with ESMA s view (ESMA advice p115 pp15) that one of the core objectives of the review of MiFID, strengthening investor protection, means providing clients with a more detailed and granular picture of costs and charges. This underpins our recommendations regarding the specification of the summary indicators. However, we would add one further key question: "Who am I paying?" Question eight is about comparing the costs with other products. Additionally, we think it is important that consumers know who they are paying in order that they can compare the costs with similar products offered by other providers. In respect of offering some form of benchmark for considering costs we do not think this should be done within the PRIIP KID. The UCITS experience is that there a numerous KIIDs produced and many are frequently revised in response to changes in factors specific to each UCITS. Introducing a benchmark, which is subject to changes not related to the PRIIP in question will serve only to increase the frequency of revisions. We think benchmarking or peer comparison is better addressed by data vendors. The Investment Association has done a lot of work on cost and charges disclosure and a copy of our discussion paper, Meaningful disclosure of costs and charges, February 2015, is attached. This paper includes our views on the necessary granularity of cost disclosure to provide information that is useful in helping investors make informed decisions Examples of observed cost structures [p. 53] 16: What are the main challenges you see in achieving a level-playing field in cost disclosures, and how would you address them? We see the diversity of product charging structures as being the main challenge. Achieving a levelplaying field is fundamental to the purpose of the PRIIPs Regulation, which is to enable retail investors to understand and compare PRIIPs (Article 1). The ESA s have highlighted the key difficulties in respect of insurance PRIIPs, structured products and derivatives. It is essential that simpler products or products with different organisational arrangements are not made to look artificially more expensive simply because they have greater transparency of costs. We believe the potential to level the playing field is the greatest benefit that the PRIIP KID can deliver and we would encourage the ESA s to continue to aim for this goal. 6

7 Also fundamental to the level-playing field is the ability to compare similar products offered by different providers. To achieve this it is essential that consumers can compare the service costs of manufacturing, operating and managing the products of different providers without distorting these figures by combining them with the volatile costs of implementing the investment strategy in the market. To address the challenges we suggest building on the level 1 requirements which provide two specific and separate elements; summary indicators and total aggregate costs to ensure comparability and show the compound effect of costs on the investment. These two elements should have distinct characteristics. The summary indicators should be simple (but not overly simplistic) and factual in nature and could be based on the structure of the costs section of the UCITS KIID. We note that the cost categories identified by the ESAs (p58-59), notwithstanding the potential difficulties in quantifying these items, can be classified as one-off, recurring (or ongoing), or contingent in nature. We think presenting the summary indicators in this way is essential to helping investors to understand the costs and to answer some of the key questions in table 10 about the uncertainty of costs. The compound effect of costs necessarily rely on assumptions and estimates. Ideally assumptions should be minimised and used only when unavoidable in order to avoid giving false expectations and misleading consumers. In our answer to question 26 we set out what we believe to be the benefits of characterising the level 1 requirements in this way. 17: Do you agree with the outline of the main features of the cost structures for insurance-based investment products, structured products, CfDs and derivatives? Please describe any other costs or charges that should be included. Yes, we agree with the ESAs assessment of the various cost structures. 18: Do you have any views on how implicit costs, for instance costs embedded within the price of a structured product, might be best estimated or calculated? No. 19: Do you agree with the costs and charges to be disclosed to investors as listed in table 12? If not please state your reasons, including describing any other cost or charges that should be included and the method of calculation. Our answer is organised according to the cost categories on table 12. Look-through costs 7

8 It is essential that look-through costs are properly captured and disclosed in order to prevent opportunities to structure products to hide costs. This issue is tackled in the UCITS ongoing charges figure by requiring a synthetic element to be added representing the costs in underlying funds. The approach used in set out in the UCITS guidelines (CESR/10-674) and is triggered by a fund holding a substantial proportion in underlying funds. This term has been interpreted differently in a number of Member States and we would recommend something more robust is developed for the PRIIP KID. In the UK we recommend that UCITS managers consider the potential effect of the underlying funds charges on the overall ongoing charges figure because this captures both the magnitude of the underlying charges and the significance of the underlying fund as part of the portfolio. Portfolio transaction costs The ESAs have produced a list of possible types of portfolio transaction costs in table 12. The ESAs have considered the cost disclosures required of UCITS (p ) and in particular the list of exclusions from the ongoing charges figure in paragraph 5 of CESR s guidelines (CESR/10-674). We comment more generally on this material in our answer to question 1, but in relation to portfolio transaction costs we have made specific comments here. We note that the CESR guidelines exclude transaction costs from the ongoing charges figure and the ESAs take the view that these costs should be included in the PRIIPs KID disclosures. However, we do not think that CESR considered whether all items mentioned in the exclusion were valid transaction costs or just valid exclusions. In consulting (CESR/ p50 question 61) CESR asked respondents whether they agreed with the proposed methodology for identifying which items should be included in the ongoing charges figure. We think that respondents would have considered whether the methodology defined the exclusions appropriately and not whether it defined transaction costs. Therefore, we would encourage the ESAs to explore the nature of the UCITS exclusions that have come to be listed as portfolio transaction costs in table 12, in particular, bid-ask spread and market impact costs. Broker commissions We note that in developing the UCITS KIID, CESR expressed a desire for greater transparency of transaction costs in certain asset types, such as debt securities, but noted that it could not be achieved easily without the involvement of the sell-side (brokers and other execution venues). The ESAs will need to address this in order to extract transactions costs that are embedded in the bid-ask spread. Entry and exit charges paid by the fund The ESAs describe these charges as the amounts deducted from incoming or outgoing investors in order to protect existing and continuing investors from dilution caused by the costs of buying and selling investments. These charges are fundamentally different to other costs and charges because the amount charged to incoming or outgoing investors is paid into the PRIIP for the mutual benefit of existing or continuing investors. The parties operating and managing the PRIIP do not receive or benefit from the charge. Moreover, many investors will experience a negative charge in this respect because usually there will be incoming and outgoing investors at the same time. Where inflows exceed outflows, incoming investors will be charged to protect the existing PRIIP from the dilutive 8

9 effect of buying more investments. However, outgoing investors will receive a higher proceeds than if they were selling the underlying investment of the PRIIP directly. These charges are a part of the client dealing mechanisms of the PRIIP and should be disclosed as such, not as costs and charges. Transaction Tax and Stamp Duty We agree that these taxes are related to portfolio transactions. However, we think it is inconsistent to treat these taxes as costs and charges whereas other types of tax are excluded and this distorts the relative costs of investing in different markets. Other taxes include irrecoverable tax withheld on dividends paid to a PRIIP and tax on profits passed on to insurance products (where the insurer may suffer tax on income and/or gains on investments backing the product and passes these tax charges on to their policyholders as a deduction from their investment value). In principle, tax should either be included or excluded but the DP proposes an approach whereby transaction tax is included and other tax is excluded. Bid-Ask spread We note that ESMA acknowledges that in some markets transaction costs are embedded in the bidask spread which makes these costs difficult to quantify but this should not prevent clients from receiving full disclosure in respect of these embedded costs (p119 pp24). This suggests that ESMA does not regard the bid-ask spread as a transaction cost, but that a part of the spread represents embedded transaction costs equivalent to the broker commissions observed in equity markets. The ESAs identify that spread depends largely on the liquidity of a particular security (p59) and propose that spread is a potential measure of liquidity risk (p114). In discussing liquidity risk (p26), the ESAs identify the link between liquidity and market price. We agree that liquidity and market price are inextricably linked. In the case of real estate, for example, determination of the fair market price relies on an assumption of an appropriate marketing period. Owners of real estate (including individuals that own their own home) understand that they will most likely realise a lower price if they want to sell within a week than if they market the property for three to six months. However, it is doubtful that sellers seeking a quick sale would regard the discount to the price as a transaction cost along with their agent s fees. As with real estate, other asset classes experience the trade-off between liquidity and market price to a greater or lesser degree. The ESAs also observe (p25) that credit risk of the underlying assets of a PRIIP may end up being reflected in market risk. This would be the case where, for example, the credit spread on a corporate bond widens due to a deterioration in the credit rating of the bond and, as a result, the market price falls. The interrelationship between the market price and the liquidity and credit profiles of an asset mean that the bid-ask spread, being the difference between the market prices for buying and selling the asset, is inherently a function of underlying market risk. Therefore, notwithstanding that they may be some transaction costs embedded within it, the bid-ask spread should be excluded from the cost and charges disclosures required by MiFID II. It follows that it also should be excluded from PRIIPs costs and charges disclosures. Market impact costs 9

10 The ESAs describe market impact as being the cost incurred because the transaction itself changed the price of the asset. ESMA notes that underlying market risk relates only to movements in the value of capital invested caused directly by movements in the value of underlying assets (ESMA advice p119 pp25). Market impact depends on the size of the transaction, its urgency and the market s ability to absorb the transaction in the timeframe required, in other words, it is a function of the market in general and not the counterparties involved in the transaction. Just as the performance of an investment product is a function of the skill of the product s manager in selecting the right investments at the right time, so the effect of market impact is a function of the skill in executing the necessary deals in the market. A lack of skill in either discipline will result in a larger transfer of value to another counterparty and hence a lower investment return. But neither represent a cost in the context of consumer disclosures under MiFID and PRIIPs Aggregating Costs [p. 59] 20: Do you agree that a RIY or similar calculation method might be used for preparing total aggregate cost figures? Yes, we agree with using a RIY methodology to calculate total aggregate costs figures. We note that the PRIIPs Regulation requires such figures in order to show the compound effects of the total costs on the investment. Of the three basic methodologies identified by the ESAs, RIY is the only one that shows the compound effects of costs. We do not think it appropriate to use an adjusted ongoing charges methodology. The UCITS ongoing charges figure is now widely understood and is a reliable indicator of the charges taken by the various parties involved in managing and operating a fund. It is worth noting that the ongoing charges figure used in the UCITS KIID is required to be a reliable indicator of likely future charges (Article 24 of the KII Regulation). We think that combining ongoing charges with volatile costs (such as transaction costs and performance fees) would reduce the effectiveness of the figure as a reliable indicator and would reduce the PRIIP KID's effectiveness as a means of comparing the costs with similar products offered by other providers. It would also make it very difficult to target the disclosures recommended by ESMA to warn consumers of the potential for estimates to deviate from actual costs (ESMA advice p121 pp35). Targeting warnings in this way is essential to address the key questions in table 10 about the accuracy and variability of the information on costs. 21: Are you aware of any other calculation methodologies for costs that should be considered by the ESAs? No. 22: Do you agree that implicit or explicit growth rates should be assumed for the purpose of estimating total aggregate costs? How might these be set, and should these assumptions be adjusted so as to be consistent with information included on the performance scenarios? No, we do not agree with using assumed growth rates because it is not necessary. Although RIY does require assumptions to be made, in order to enhance the reliability of the figures produced, the use of assumptions should be minimised and should only occur when absolutely necessary. The purpose 10

11 of the total aggregate costs figure is to show the compound effects of the total costs on the investment which is different to the effect on investment returns. Therefore the RIY calculation should be constructed to illustrate the effect of costs on the amount invested in the absence of returns. A variant of this approach might be to set the growth rate to be the rate required in order that returns exactly cover costs, in other words, the rate required so that the investor gets back exactly the amount originally invested. Our answer to question 24 provides more detail on of views regarding the use of assumptions. 23: How do you think implicit portfolio transaction costs should be taken into account, bearing in mind also possible methods for assessing implicit costs for structured products? In our answers to questions 22 and 24 we advocate assuming zero return or a return sufficient only to preserve the capital invested. In effect, this approach takes the view that all variables are held constant except costs and charges. If it is assumed that returns are nil, or that capital remains constant, it is appropriate also to assume there is no portfolio trading activity and therefore there will be no portfolio transaction costs. The basis for disregarding returns from the calculation of the compound effect of costs is that past returns are not necessarily a guide to future returns. On the same basis we recommend disregarding portfolio trading activity, because we can demonstrate (Meaningful disclosure of costs and charges (February 2015) Annex 4) that past portfolio trading activity is not necessarily a guide to future activity. The nature of portfolio transaction costs is very different to implicit costs for structured products so we do not see a problem in disregarding the former whilst including the latter. Portfolio transaction costs are contingent in that they arise only when the portfolio is traded in order to implement the investment objective. The investment manager earns a fee (part of the ongoing charge) regardless of whether the investment decision is to hold, not hold, buy or sell a particular stock. The return from implementing the decision accrues to the portfolio. In contrast, the manager of a structured product creates the product in conjunction with, for example, an investment bank. The product manager takes a fee from the product offered whilst the investment bank earns its profit from the initial set up and the ongoing pricing parameters of the instruments, typically derivatives, created in order to deliver the structured return. The investment bank s profit is not contingent on future activity Parameters and assumptions [p. 61] 24: Do you have any views on possible assumptions that should be made, and how these might be calibrated or set? Our view is that, in order to avoid giving false expectations and misleading consumers, assumptions should be minimised and used only when unavoidable. During the consumer testing of the UCITS KIID it was identified that consumers have a tendency to view any figures shown as facts rather than indications or guidelines despite the fact that the text accompanying the figures state that they rely on a number of assumptions (UCITS Disclosure Testing Research Report (June 2009) p129 pp7.50). Therefore it is essential to separate factual and illustrative disclosures and that sufficient factual and reliable information is given in the summary indicators to enable advisers and distributors to fulfil 11

12 their obligations under MiFID II. Assumptions should be used only in calculating the compound effect of charges. ESMA has emphasised (ESMA advice p120 pp32) that, for the purposes of MIFID II cost disclosure, the only information that can be provided on a generic basis is the assumed investment amount. Moreover, ESMA has clarified (ESMA advice p121 pp35) that volatile costs (such as transaction costs and performance fees) should be estimated on a best effort basis and accompanied by a warning that assumptions have been used and the costs may not be indicative of likely future costs. These observations are relevant to our comments about each of the key areas for further work identified by the ESAs that follow. Ex-post vs ex-ante The ESAs observe that, in general, the UCITS ongoing charges figure is calculated on an ex-post basis and raise the concern that past costs are not necessarily relevant to what might happen in the future. However, the KII Regulation (Article 24) requires that the ongoing charges figure should remain reliable as an indicator of the amount likely to be charged in the future. We think this was a key improvement to the TER metric and ensures the ongoing charges figure is a reliable ex-ante indicator. We recommend that the ESA should take the same approach for ongoing charges in the PRIIP KID. The ESAs also note that it may be misleading in some cases to use ex-post data when providing details of the cumulative impact of costs over time. We think that the UCITS approach of ensuring that the summary indicators are a reliable indicator of future costs would help to address this issue. It follows that volatile costs (such as transaction costs and performance fees) should not be incorporated into aggregated costs figures. We expand on this point in our answers to question 23 in respect of the trading activity that determines transaction costs and question 28 in respect of performance fees. Holding period It is necessary to assume holding periods in order to calculate the compound effect of costs. We do not agree with using a recommended holding period because this could potentially be manipulated to show periods for which costs are artificially low. We agree with the ESAs suggestion of using certain standardised time horizons and we recommend using 1, 3, 5 and 10 years plus, where relevant, the full term of a fixed-life product. We think this gives sufficient granularity in the early years to illustrate clearly the potential effect of high entry or exit charges. Rate of return / growth rates We do not agree that it is necessary or appropriate to assume a rate of return. The PRIIPs Regulation requires the compound effect of costs on the investment to be shown and this does not need to include a rate of return. If an assumed rate were to be used, comparability would be enhanced if the ESAs specified a rate that all PRIIPs should use. However, the diversity of PRIIPs means it is unlikely that a single rate would be appropriate for all PRIIPs and would lead to comparisons based on unrepresentative scenarios and potentially misleading expectations. Therefore, we recommend disregarding this factor, either by assuming zero growth or by assuming the consumer will get back exactly the amount they put in. In effect, this assumes a growth rate exactly sufficient to cover the 12

13 costs, and is our preferred option because zero growth is prone to portray an overly pessimistic view of the effect of costs. Assumed amount invested Total aggregate costs are required to be expressed in monetary and percentage terms which means that it is necessary to assume a monetary amount invested. We agree with the ESAs that it may be appropriate to differentiate between single and regular contribution products but, in the interests of comparability, it is essential that the ESAs should specify the assumed amounts in each case. Inevitably there will be products where the specified amount is not appropriate but the PRIIP KID is designed for retail investors so most products are likely to be within the financial reach of the general public. If it is necessary to specify multiple assumed amounts they should at least be readily scaleable, for example, in multiples of 10 ( 1,000, 10,000 or 100,000). Rebates The PRIIPs Regulation requires there to be a clear indication that distribution costs not included in the costs specified in the KID will be provided by other parties. We think it should be made clear that such parties also will identify separately their costs where these costs are included within the KID and subject to a rebate arrangement in accordance with ESMA s technical advice (ESMA advice p123 pp11). 4.5 Presentation of cost disclosure in the KID [p. 63] Options for presenting costs [p. 63] 25: What do you think are the key challenges in standardising the format of cost information across different PRIIPs, e.g. funds, derivatives, life insurance contracts? We think the key challenges in standardising the format of cost information relate to the conflict between the need for the KID to enable retail investors to understand and compare PRIIPs. Understanding requires granular and nuanced information whereas comparison is enhanced by standardisation and simplification. In ESMA s view, (ESMA advice p115 pp15), strengthening investor protection means providing clients with a more detailed and granular picture, but that this needs to be balanced with consumers ability to process information. The challenge for the ESAs is to strike the appropriate balance between these factors. In our opinion the cost section of the UCITS KIID provides an excellent starting point. The KIID was thoroughly tested on consumers during its evolution and it provides a fair, clear and not misleading view of charges taken from a UCITS. It contains summary indicators of sufficient granularity to understand the nature of the charges, sufficient reliability to ensure it is indicative of likely future charges and sufficient simplicity to ensure it remains salient. The PRIIP KID could replicate the UCITS summary indicators and add an additional indicator in respect of portfolio transaction costs. The advantages of this approach are that PRIIPs and UCITS would be reasonably comparable during the transitional exemption under article 32 and would preserve the meaning of terms that are now established and familiar to UCITS investors and beyond. 13

14 We agree with the ESAs that the information on costs should be coordinated to ensure it is considered complete and sufficient for the purposes of MiFID II disclosures. In this respect we note that in finalising the technical advice, ESMA removed reference to one single figure. In our view, this clarifies the definition of aggregation as being the adding together the costs of investment/ancillary services and the equivalent costs within the financial instrument; it does not require different cost items to be combined into a single figure. This is consistent with the illustrative examples included in ESMA s consultation paper. ESMA also modified the tables in annex (ESMA advice p ) to make it clear that performance fees are not an example of ongoing charges. We think this is evidence that ESMA regard it as necessary to disclose an aggregated figure for each of the different cost items in the tables in the annex. Therefore we recommend that the PRIIP KID includes a separate summary indicator for each of the cost items listed in table 2 of the ESMA s annex and we illustrate such an approach in our answer to question : Do you have a marked preference or any objection for any of the presentational examples? If so, why? Please provide any alternative examples which you believe could be useful. Our preferred approach would be to present summary indicators that are factual and as free from assumptions as possible together total aggregate costs in the form of reduction in yield figures similar to option 6. Clearly it will be necessary to provide explanations of the figures to help consumers understand them. We recommend the following set of summary indicators should be required to be presented, expressed in percentage terms: Entry costs Exit costs Ongoing costs Performance fees Portfolio transaction costs [ ]% of the amount contributed before any costs are deducted [ ]% of the proceeds realised before being paid to the client [ ]% of the value of the client s interest in the PRIIP [ ]% of the return above the [benchmark] [ ]% of the underlying transaction value We recommend that total aggregate costs should be required to be presented for a number of specified holding periods, expressed in monetary and percentage terms, in the form of a table, as follows. The ESAs should specify the holding periods as being 1, 3, 5 and 10 years and should allow the full term of a fixed-life product to be given where relevant. However, we do not think a recommended holding period should be included as this could be manipulated to show periods for which costs are artificially low. Holding period Effect of costs Years % per year % % % % Term % We think our approach is appropriate because: 14

15 It aids understanding and comparability and is accurate, fair, clear and not misleading. It complies all of the relevant requirements of the PRIIPs Regulation. It helps to answer the key consumer perspective questions in table 10. Based on ESMA s technical advice, it provides all of the information necessary for aggregating costs in accordance with MiFID II. It facilitates the necessary warnings (ESMA advice p121 pp35) about estimates of volatile costs (transaction costs and performance fees) in a targeted manner. It maximises comparability with UCITS during the transitional exemption period. It preserves UCITS terminology that is now in common usage and is familiar to consumers. We do not think any of the options presented achieve this. We expect that consumer testing is likely to reveal a preference for the simpler presentation options and a tendency to misinterpret the more detailed options. However, there needs to be sufficient information for advisers and distributors to fulfil their MiFID II obligations and to interpret the PRIIP KID in order to advise their clients. 27: In terms of a possible breakdown of costs, are you aware of cost structures for which a split between entry or exit costs, ongoing costs, and costs only paid in specific situations or under specific conditions, would not work? We are not aware of any such cost structures. 28: How do you think contingent costs should be addressed when showing total aggregated costs? In our answers to questions 22 and 24 we advocate assuming zero return or a return sufficient only to preserve the capital invested. In effect, this approach takes the view that all variables are held constant except costs and charges. If it is assumed that returns are nil, or that capital remains constant, it is appropriate also to assume that there is no return in excess of a benchmark or target and therefore there will be no performance fees. The basis for disregarding returns from the calculation of the compound effect of costs is that past performance is not necessarily a guide to the future. On the same basis past outperformance is not necessarily a guide to future outperformance. An understanding of a performance fee can only be achieved by being aware of its terms for example, the manager gets to keep 20% of any outperformance. 29: How do you think should cumulative costs be shown? We think cumulative costs should not be shown in the PRIIP KID. Our recommended approach in our answer to question 26 satisfies the requirements of the PRIIPs Regulation. Our recommended summary indicators provide all the information necessary for advisers and distributors to aggregate with their own costs and charges data in order to meet their MiFID II obligation to show the cumulative effect of costs on the return of the investment. 5 Other Sections of the KID [p. 73] 5.2 Identity Article 8(3)(a) [p. 73] 15

16 30: Do you have any views on the identity information that should be included? Regarding relevant information about the competent authority, the format in CESR s template for the KIID is suitable, ie. [Name of PRIIP manufacturer] is authorised in [name of Member State] and regulated by [name of NCI]. The inclusion of market identifiers, such as ISINs, would be irrelevant or confusing for most retail investors. 5.3 Comprehension Alert [p. 74] 31: Do you consider that the criteria set out in recital 18 are sufficiently clear, or would you see some merit in ESAs clarifying them further? We welcome the ESAs initial suggestions on how to better capture which PRIIPs should be excluded from the comprehension alert, and the reference to Article 50 of the UCITS Directive as a basis for what could or could not be included in the comprehension alert. While this should serve as a basis for further discussion, there are other types of investments that retail investors may seek exposure to, such as real estate and precious metals, and the comprehension alert should also exclude these investment types. 5.4 What is this product? [p. 76] Type of PRIIP [p. 77] 32: Do you agree that principles on how a PRIIP might be assigned a type will be needed, and do you have views on how these might be set? A PRIIP should be assigned a type based on the underlying regulatory types. These are: (i) UCITS or AIFs (open-ended or closed-ended); (ii) insurance (unit-linked or with profits); (iii) structured bonds; (iv) structured deposits; and (v) derivatives (as defined on page 13 of the Discussion Paper). We do not see any merit in using the classification of AIFs related to the reporting requirements under AIFMD as many are not available to retail investors. Moreover, the underlying assets of the PRIIP are covered in the risk section of the KID. 33: Are you aware of classifications other than by legal type that you think should be considered? A PRIIP should be assigned a type based on the underlying regulatory types. These are: (i) UCITS or AIFs (open-ended or closed-ended); (ii) insurance (unit-linked or with profits); (iii) structured bonds; (iv) structured deposits; and (v) derivatives (as defined on page 13 of the Discussion Paper) Objectives and means of achieving them [p. 76] 34: Do you agree that general principles and as necessary prescribed statements might be needed for completing this section of the KID? 16

17 A combination of general principles and clear guidelines (Level 3) would be helpful. The objectives and investment policy section of CESR s template for the KIID is a good starting point. Generally, we agree that general principles and, as necessary, prescribed statements might be needed for completing this section of the KID. 35: Are you aware of other measures that might be taken to improve the quality of the section from the perspective of the retail investor? We believe that the objectives and investment policy section of CESR s template for the KIID is a good starting point and that it is unnecessary to go further than this Consumer types [p. 89] 36: Do you have views on the information PRIIPs manufacturers should provide on consumer types? The problem with identifying a consumer type at which any PRIIP is aimed is that it will have to be generic and high level because it cannot factor in what other investments the consumer already owns, nor other circumstances particular to the investor. A PRIIP with a focussed investment objective, eg. Asian emerging markets, could be suitable for any investor as long as it was part of a diversified portfolio which, overall, matched the investor s risk appetite Insurance benefits [p. 80] 37: What is the key information that needs to be given to the retail investor on insurance benefits, and how should this be presented? Term [p. 81] 38: Are you aware of PRIIPs where the term may not be readily described, or where there are other issues? Standard sentences should be specified for fixed length or open-ended investments. 5.5 What happens if [the name of the PRIIP manufacturer] is unable to pay out? [p. 81] 39: Are you aware of specific challenges arising for specific PRIIPs in completing this section? There are PRIIPs where a bankruptcy of the manufacturer is largely irrelevant because the investment is insolvency remote (eg. funds with separate depositary arrangements are effectively ring-fenced from the financial position of the PRIIP manufacturer, so the question of whether the PRIIP manufacturer is able to pay out is not relevant). This is an important factor which should be mentioned in the KID. 17

18 5.6 How long should I hold it and can I take money out early? [p. 82] 40: Are you aware of specific challenges arising for specific PRIIPs in completing this section? The answer to the first part of this question ( How long should I hold it? ) will have to be generic and high level in nature because it cannot factor in circumstances particular to the investor, including other investments held. This section should include reference to factors know at the time of the investment, such as detrimental effects of an early exit due to front loading of charges in certain PRIIPs. 5.7 How can I complain? [p. 83] 41: Are you aware of specific challenges arising for specific PRIIPs in completing this section? This information needs to be generic, otherwise manufacturers might need to create different KIDs for the same PRIIP in order to address different distribution channels, contrary to the concept of a single KID per PRIIP. This can be accomplished by including a reference to where further information may be found. 5.8 Other relevant information [p. 84] 42: Do you agree that this section should link to a webpage of the manufacturer? We agree with the publication of additional information on the manufacturer s website. 6 Products offering many options [p. 85] 43: Do you agree with the assessment of when PRIIPs might be concerned by article 6(3)? Beside a general KID, there should also be specific KIDs for each of the options provided. 6.3 Scale of market [p. 86] 44: In your market, taking into account the list of criteria in the above section, what products would be concerned by article 6(2a)? What market share do these represent? 45: Please provide sufficient information about these products to illustrate why they would be concerned? 6.4 Impact of article 6(3) [p. 86] 18

19 46: Do you have views on how you think the KID should be adapted for article 6(3) products, taking into account the options outlined by the ESAs? Using examples of representative investment options for illustrating risks and rewards or costs of an investment could lead to poor consumer outcomes. If a product offers investors a choice between different investment strategies, use of a representative investment could mislead investors. The use of ranges for the risk/reward information and cost data could be considered. 47: How do you consider that the product manufacturer should meet the requirements to describe and detail the investment options available? Any description of the different investment options should correspond with the information required in the section What is this product for individual PRIIPs. 48: Are you aware of further challenges that should be taken into account? In the case of products offering many options, it could be difficult for investors to relate the information about the product wrapper to the information about specific options. This may be particularly the case with any descriptions of overall risk profile and overall costs, as well as performance scenarios net of costs. 7 Review, Revision and Republication [p.89] 49: Do you agree with the measures outlined for periodic review, revision and republication of the KID where material changes are found? The UCITS KII Regulation should be used as a guide. It is important to remember that the KID has been designed as pre-contractual information and any attempt to extend this to inform existing investors of changes (other than through a broadcast or passive model) would create serious complications (including overlapping existing periodic disclosure requirements) and should be avoided. Further clarification on the meaning of material changes would be helpful in this regard. 50: Where a PRIIP is being sold or traded on a secondary market, do you foresee particular challenges in keeping the KID up-to-date? Closed-ended funds have signing periods during which the product is offered. If a product is no longer offered by the PRIIPs manufacturer, there should be no obligation to keep the KID up-to-date. 51: Where a PRIIP is offering a wide range of investment options, do you foresee any particular challenges in keeping the KID up-to-date? 52: Are there circumstances where an active communication model should be provided? 19

20 Publication of the revised KID on the manufacturer s website ( passive communication ) should be sufficient. There should be no duty to actively communicate modifications of a PRIIPs KID to existing investors. 8 Timing of delivery [p. 93] 53: Do you agree that Recital 83 of the MiFID II might be used as a model for technical standards on the timing of the delivery of the KID? While it is valid that technical standards address the issue of retail investors being given the time they need to consider the KID before making any decision, it is an equally valid investor protection concern that they should not be prevented from investing at the point they want to because of constraints imposed by overly prescriptive regulation. PRIIPs KID technical standards should align with the Distance Marketing Directive and MiFID II. 54: Are you aware of any other criteria or details that might be taken into account? 9 General aspects of the KID [p. 97] 9.2 Use of templates to establish consistent look and feel or visual style [p. 97] 55: Do you think that the ESAs should aim to develop one or more overall templates for the KID? The KIID template published by CESR (CESR/ ) has helped consumers by making sure all KIIDs are similar in format and it has helped firms by clearly indicating how to construct a compliant document. A similar approach for the KID would be welcome. 9.3 Single payment and regular payment products [p. 98] 56: Do you think the KID should be adjusted to reflect the impact of regular payment options (on costs, performance, risk) where these are offered? If so, how? As the KID is designed to be a generic disclosure document, multiple versions reflecting different payment options should be avoided as such multiplication will just make it harder for the retail investor to locate the right version. One solution would be to produce the KID for any PRIIP on the basis of single payment in single payment out, unless that payment option is not available in the specific PRIIP concerned, in which case the KID should reflect the reality of the payment profile. 10 Impact assessment [p. 101] 57: Are there other cost or benefit drivers that you are aware of that have not been mentioned? Please consider both one-off and ongoing costs. 20

21 58: Do you have any evidence on the specific costs or benefits that might be linked to the options already explored earlier in this Discussion Paper? Please provide specific information or references broken down by the specific options on which you wish to comment. 59: Are you aware of situations in which costs might be disproportionate for particular options, for instance borne by a specific group of manufacturers to a far greater degree in terms relative to the turnover of that group of manufacturers, compared to other manufacturers? 21

22 February 2015 Meaningful disclosure of costs and charges Discussion Paper Page 0 of 60

23 Contents Page Overview 3 Next Steps 4 Chapter Summaries 5 Chapter 1 5 Chapter 2 5 Chapter 3 5 Appendices 5 Chapter 1: Defining charges and transaction costs Background Clarity of charge disclosure Nature of transaction costs Transaction cost variability: evidence Types of transaction costs Explicit costs Implicit costs Quantitative vs qualitative implicit costs Further considerations 15 Chapter 2: Portfolio Turnover Rates Background Key Issues UCITS methodology SEC methodology Analysis of the methodologies UCITS methodology SEC methodology Refining the existing methodologies Securities Purchases and Sales Issues and cancellations Average net asset value Average value of portfolio securities Empirical analysis Method 1, UCITS methodology divided by two Method 2, method 1 without an adjustment to reflect investor flows Method 3, modified SEC methodology Conclusions 21 Chapter 3: Practical proposals for product neutral disclosure Summary framework for disclosure Definitions and methodology for product-neutral approach Wider application in UK pension context Towards a common template? PTR presentation 30 Page 1 of 60

24 3.6. Next steps 31 Appendices 32 Appendix 1: Principles of disclosure 32 Appendix 2: Summary of initiatives to improve transparency 33 Appendix 3: The importance of pricing policy 35 Appendix 4: The variability of explicit transaction costs 37 Appendix 5: Comparison of PTR methodologies 39 Page 2 of 60

25 OVERVIEW This technical position paper builds on The Investment Association s work over the past three years, which has already seen the provision of additional information on transaction costs and a framework for unit-level historic accountability in pounds and pence. It has a number of overlapping objectives, relating both to our own aims for improving industry transparency and specific forthcoming UK and EU regulation: To articulate an approach to disclosure of charges and costs for investment products, which can be used by key decision-makers, such as pension scheme trustees, and end consumers. This also includes a conceptual approach that could extend to how total charges and costs are expressed for individual pension accounts. To facilitate consistent charge disclosure by fund and investment managers to schemes in the context of the charge cap on default arrangements required under the Pensions Act To provide a template for transaction cost disclosure that can inform current regulatory work on the provision of transaction cost information required by the Pensions Act To inform the technical debate at European level about the aggregation of charges and costs to be required under the revised Markets in Financial Instruments Directive (MiFID II) and the regulation on Key Information Documents (KIDs) for Packaged Retail and Insurance-based Investment Products (PRIIPs). To develop additional information about activity levels within portfolios via more robust measures of portfolio turnover rate (PTR) than the withdrawn UCITS measure. The output in practical terms is a definitional framework for charge and transaction cost disclosure that would provide the underlying component pieces for good disclosure. It can be used across different products, and could be developed as a common template once the shape of regulation in this area becomes clear. The paper is informed by a set of principles for good disclosure (see Chapter One and Appendix One) and proposes: A consistent definition and expression of product charges, based on the Ongoing Charges Figure (OCF) in conjunction with other charges as relevant. A definition of explicit costs that can be expressed in pounds and pence or percentage terms on a historic basis, and on an indicative estimated basis looking forward. A definition of implicit costs that would see the bid-offer spread within the relevant markets available on at least a quarterly basis. A recommendation for PTR methodology that is based on a modification of the Securities and Exchange Commission (SEC) approach. A format for graphic representation of PTR and transaction costs that would allow consumers to have a better appreciation of the performance hurdle created. We look forward to working with Government, regulators, industry and consumer groups to determine precisely what form this information provision should take. With the right narrative and graphic support, it will be possible to shed more light on investment services, whatever the end product Ultimately, the success of all proposals whether industry or regulatory led should be judged by three key criteria: Disclosure should facilitate informed decision-making, by or on behalf of savers and investors. Disclosure should allow users to compare a range of product types and providers of like products. Disclosure should be designed to serve the best interests of savers and investors. Page 3 of 60

26 NEXT STEPS This paper summarises The Investment Association s thinking to date, and we would welcome comments and contributions about any of the issues raised. In terms of practical implementation, the issues discussed will also be the subject of regulatory outputs in the UK and EU in the near future: We are working with the Financial Conduct Authority (FCA) and Department for Work and Pensions (DWP) to assist the formation of cost disclosure regulation to be introduced in 2015 for the automatic enrolment pensions market. Under Pensions Act 2014, this will result in specific rules for pension scheme assessment of transaction costs, expected by summer 2015 and subject to forthcoming consultation. We would also like to see a common approach developed for charge disclosure across different products as part of this work. At EU level, MiFID and PRIIPs are at different stages of the legislative process. A path towards charge and cost aggregation is in development, but how transaction costs are defined and accounted for, and the future shape of the OCF, will be a major technical topic of consultation over the next 12 months. We believe that any aggregation should still allow charge and transaction cost information to be shown separately, and this paper is therefore also intended to contribute to the European technical debate. Comments should be sent to positionpaper@theinvestmentassociation.org Page 4 of 60

27 CHAPTER SUMMARIES Chapter 1: Defining charges and transaction costs We introduce The Investment Association s approach to disclosure and outline a way of providing charge and transaction cost information. This is based on principles that aim to establish a meaningful and consistent disclosure regime. We set out the difference between product charges and transaction costs, and why both should be disclosed separately, even if there is a requirement to aggregate at an overall level. We explain why we continue to believe a distinction in disclosure should be made between future disclosure ahead of any product sale, and historic disclosure at the end of an accounting period. We explain why we consider explicit costs to be different from implicit costs, and why we continue to believe there are two categories of implicit costs which should not be confused. We suggest a way in which implicit cost information could be made available based on current unit pricing approaches in underlying funds. Chapter 2: Measuring activity levels using Portfolio Turnover Rate We provide an analysis of portfolio turnover rate (PTR) metrics. While these are not cost measures in and of themselves, they can help to provide an indicator of the activity levels within portfolios (ie. how much buying and selling of stocks and securities is taking place). We identify strengths and weaknesses of alternative approaches, including the two dominant approaches to date (from European and US regulators). We propose a single methodology, based on a modified version of the Securities and Exchange Commission (SEC) methodology, which we consider could be introduced in disclosures to clients. Such disclosure would need to include clear narrative statements about the role any metrics should play in assessments by consumers. Chapter 3: Practical approach to charge, cost and PTR disclosure We propose a product neutral approach for charge and transaction cost disclosure that we believe could be used in the context of the UK pension reforms. This is informed by the principles outlined in Chapter One. It would include both quantitative and narrative information. Depending on the final shape of regulation in this area, The Investment Association intends to explore the scope for a common template to facilitate consistency of disclosure. We set out an example of a graphical representation of the relationship between PTR and transaction costs that could help consumers understand better the implications of portfolio turnover, subject to appropriate narrative. Appendices Appendix 1 provides a summary of the principles set out in the paper. Appendix 2 summarises previous Investment Association activity in this area. Appendix 3 discusses pricing policies for open-ended funds to assist the wider market understand how this presently occurs and how spreads are already factored into pricing. Appendix 4 looks at a range of examples to show explicit transaction costs variability. This provides empirical evidence showing support for not presenting a single aggregated number to cover all charges and costs. Appendix 5 presents empirical data comparing methodologies for calculating portfolio turnover rates Page 5 of 60

28 CHAPTER 1: DEFINING CHARGES AND TRANSACTION COSTS To help explain the framework that we set out in Chapter Three, this chapter explores the following areas: The definition of charges and transaction costs. Forward-looking (ex ante) versus historic (ex post) reporting. The nature of transaction costs, notably the difference between explicit and implicit costs. The distinction between quantification of transaction costs and a qualitative assessment of the trading process. We set out our views, based on a series of eight principles. We also outline our work to date in this area and suggest specifically how enhanced disclosure of implicit costs might operate Background The UK and EU debates on transparency in financial services have in recent years intensified and the scope with respect to asset management is wide, covering: All client groups, particularly retail investors and savers in Defined Contribution (DC) pension schemes. All product types, whether segregated (bespoke agreements between client and asset managers) or pooled vehicles such as investment funds (unit trusts, open-ended investment companies (OEICs)), investment trusts or insurance funds. Investment management charges as well as the transaction costs incurred in delivering a given investment strategy. Specific elements within the cost chain, notably payments for research made as part of dealing commission. This raises a range of challenges: Different client groups have different information needs and capabilities with respect to the interpretation of information. An inconsistent range of reporting mechanisms exist according to the nature of a product, partly reflecting different regulatory requirements. 1 Charges and costs behave in different ways, and transaction costs have different degrees of visibility depending upon the nature of the relevant market(s). Point of sale disclosure is intended to provide information to clients about essential characteristics of a product whose actual performance (and associated costs) cannot be known in advance Clarity of charge disclosure Asset - and fund - management services provided by investment managers operate on an agency model. Charges are levied for a specific professional service, in the course of which market costs and taxes (relating to transactions to buy and sell investments) are encountered. Other providers of financial services will also levy charges relating to the professional service they offer (for example, a group personal pension). We emphasise firstly the need for consistent calculation methodologies for charges within and across product sets. We observe the ongoing confusion within the UK long-term savings and pensions market, noting that there are currently multiple terms and methodologies for expressing something that ought in essence to be very 1 Segregated mandates will have client-specific reporting, usually on a quarterly basis, and the Pension Fund Disclosure Code provides a detailed template to provide transparency on dealing commission and research payments. Pooled vehicles have a range of standardised disclosures reflecting the collective nature of such vehicles. Investment funds, for example, have a fundlevel reporting via half-yearly reports and accounts. Page 6 of 60

29 simple: what you are paying to a given agent to provide a professional service on your behalf. We would like to see an end to this situation and move towards standardisation. This is the driver behind Principle 1. PRINCIPLE 1: Charges paid to an agent to invest on behalf of a client should be expressed and calculated in a consistent manner. The long-term savings and pensions industry should adopt a clear and inclusive terminology for all product groups, moving away from the confusion that has characterised dominant terms, notably Annual Management Charge and Total Expense Ratio. To pursue this goal, we issued guidance 2 in 2012 to the UK fund management industry (ie. operators of unit trusts and OEICs) which encouraged a focus on the Ongoing Charges Figure (OCF) rather than Annual Management Charge. In 2014, the FCA also took the same position. 3 Unlike Total Expense Ratio (TER), which was a term that was not intuitive, we believe that the OCF does what its nomenclature suggests: provides information about a product charge incurred on an ongoing and fairly predictable basis, subject to regulatory guidance on the detailed calculation. 4 Crucially, this approach also allows consumers to look across competing funds and see what the product providers are charging them for their services, and compare the costs of the service on a meaningful basis. Within the UCITS Key Investor Information Document (KIID) framework, other charges are also displayed clearly in a prescribed format (entry, exit, and performance fee if relevant). The importance we attach to being able to understand the product charge itself leads to Principle 2 below (separating charge and transaction cost information) and our belief that the UCITS format is one that is worth continuing and applying more broadly in the market. One specific advantage of the OCF over the TER is that it splits out charges that are contingent (performance fees were previously in the TER) from those that are known with a degree of certainty. An example of this approach can be seen in the sample KIID in Figure 1. This issue of contingency versus certainty is a critical part of our argument later about the difference between historic and forward-looking disclosure of transaction costs. Figure 1: Presentation of charges in UCITS KIID The Packaged Retail and Insurance-Based Investment Product (PRIIP) Key Information Document (KID), which will apply to UCITS funds in due course, will move to a fuller disclosure format, including transaction costs. However, we believe that it is desirable that the design work for the new KID also preserve the clear charge disclosure of the current KIID. 2 IMA Enhanced disclosure of fund charges and costs (September 2012) 3 FCA Thematic Review: Clarity of fund charges (May 2014) 4 CESR s guidelines on the methodology for calculation of the ongoing charges figure in the Key Investor Information Document (July 2010) Page 7 of 60

30 1.3. Nature of transaction costs Transaction costs are inextricably connected with the process of investing in order to generate a return on behalf of a client. To buy shares, bonds, real estate or any other type of asset, there is a cost of investing. This arises because of the need for an intermediary structure that can move assets from those who wish to sell to those who wish to buy. At the same time, exposure to markets can also be gained via derivative instruments, whose value relates to the behaviour of the underlying assets. But these, too, include in-built costs. All investors whether individuals buying shares in companies on the stock market or institutions investing in corporate debt are subject to costs. There is no such thing as a cost-free market return. A market index such as the FTSE 100 measures the aggregate change in value of its constituents; it does not indicate the realisable return for owning the constituent shares. If the index increases by 10% over a given period, it is impossible for any investor to obtain the full 10% return because they necessarily incur some form of cost to acquire access to the market. Transaction costs are quite distinct from a charge paid to the manager. Two asset managers may charge the same fee for their service, but invest in very different ways with a view to delivering the best return for a client given the objective of the fund or investment mandate. An active manager needs to be able to alter the different fund holdings, for example shares or bonds, during the year. There are a variety of reasons why sales and purchases will need to be made, for example, some of the holdings may no longer be good investments, or others may be more attractively priced. Transaction costs can also vary significantly across markets and across time, and will also depend upon how economic exposure is achieved (eg. use of derivatives). A fund investing in small companies will typically encounter markets where the spread (ie. the difference between bid and offer price) is significant due to absence of the same level of market liquidity as would typically be seen by a fund investing in large companies. However, not trading will leave clients without the exposure, and hence without the return, they expect. Historically, the approach used internationally has been to separate the charge levied for the service from the transaction costs incurred in delivering the service. There are a number of reasons for this: In an agency model, we believe that customers of the industry must be able to isolate exactly what they are paying the agents for their services. Clients need to be able to assess and compare the charges levied by different investment management companies for providing what may be a broadly similar objective. These charges for services contribute to investment managers profits and, all things being equal, can only be a drag on performance. Transaction costs behave in a different way to charges. At a total portfolio level, they are not necessarily predictable and fluctuate in response to the amount of buying and selling of assets that takes place. Furthermore, as we discuss below, they can be challenging to quantify with precision. The Investment Association believes that both charge and transaction cost information should be available, but presented in a way that is meaningful to allow understanding and effective usage. This leads us to Principle Two below and also informs the others which follow. PRINCIPLE 2: The charges figures should be distinguishable from transaction costs, recognising the need to be able to compare fees between managers. This is particularly important for point of sale disclosure, to avoid confusion across different asset types with different market cost structures, and to recognise the variability of transaction costs. Ideally, information about transaction costs incurred in delivering a return should be presented in the context of performance. In providing information about transaction costs, we believe that it is neither practicable nor meaningful to suggest that disclosure could be made with the same certainty on a forward-looking basis compared to full retrospective disclosure at the end of an accounting period: The forward-looking objective should be to understand the significance of potential transaction costs, so far as practicable. Page 8 of 60

31 After the fact information about transaction costs should support the objectives of full historic accountability. Our proposed disclosure approach in Chapter Three makes this distinction. In this respect, we note the recent Discussion Paper published by the Financial Services Consumer Panel, which suggested a single number should be used as a basis for product charging, and look forward to continuing the debate in light of the observations made in our respective papers Transaction cost variability: evidence A key point to emphasise is that there is a significant difference in nature between product charges, which operate on a relatively fixed and stable basis, and transaction costs, which may vary significantly across different time periods. It is important to have a debate about transaction costs with reference to real data. We have therefore analysed ten funds in detail (see Appendix 4) in order to provide a reasonable empirical basis for examining the question of predictability. This analysis shows that: While considerable consistency can be seen in terms of actual transaction costs as a proportion of transaction value, this is not a way of expressing total costs incurred. In contrast, expressing total transaction costs incurred as a proportion of Net Asset Value (NAV) (similar to the way in which ongoing charges are expressed) shows significant variation in some funds. Our view is that historic data is therefore not a reliable indicator of future costs. In fact, it could be actively misleading. We are committed to promoting transparency and full accountability for all costs and charges. However, we question whether attempting to combine (relatively stable) product charges, such as the OCF, with metrics that cannot be known with any precision on a forward-looking basis will provide meaningful and reliable disclosure for consumers. These less predictable metrics are not just confined to transaction costs, but would also include contingent charges, notably performance fees. This is why our disclosure approach proposes separation. Principles 3 and 4 capture both the intent and these distinctions. PRINCIPLE 3: Transaction costs should be disclosed to help clients whether retail, institutional or other understand the economic experience of monies invested in a given market on their behalf. PRINCIPLE 4: The distinction between ex ante and ex post disclosures should be recognised. While estimates can be used for point of sale documents, only ex post disclosure provides full accountability For a variety of reasons, transaction costs can be highly variable and care should be taken when presenting them on a basis that may appear to be predictive. Historic data is the key source of information and can be presented both for indicative purposes (accompanied, as appropriate, by narrative or graphic explanation) and as a part of full historic accountability for a realised performance Types of transaction costs It is necessary to differentiate between explicit and implicit costs. By definition, explicit costs can be identified by association to real cash payments moving from the fund, or other product, to someone else. Implicit costs, however, cover a variety of impacts not all of which are measurable with any high level of certainty. 5 FSCP Discussion Paper: Investment costs more than meets the eye (November 2014). Page 9 of 60

32 Table 1: Main costs of trading Type of cost Main features Commission (explicit) A fee paid to a broker on the purchase or sale of stocks or securities Transaction tax (explicit) Tax levied on the underlying transaction (eg. 0.5% stamp duty levied on UK shares) Bid/Offer spread (implicit) Explicit costs The difference between the price for buying (offer) and the price for selling (bid) a stock or security Table 1 above provides a simplified overview of the costs of trading. Explicit costs are straightforward to identify; they are monies paid to other agents in the market (brokerage fees) and to government (transaction taxes such as 0.5% stamp duty on equity purchases in the UK). These are known monetary amounts and, in the case of funds (unit trusts and OEICs), have been required to be disclosed in the audited report and accounts since Figure 2: Illustrative performance of a unit in issue throughout the year (SORP) Share Class Name 2013 Change in net assets per unit (p/unit) Opening net asset value per unit 1, Return before operating charges* Operating charges Return after operating charges* Distributions on income units Closing net asset value per unit 1, *after direct transaction costs of: Performance Return after charges 8.14% Other information Closing net asset value ( '000) 3,347 Closing number of units 300,000 Operating charges 0.89% Direct transaction costs 0.11% Prices Highest unit price 1, Lowest unit price 1, Page 10 of 60

33 Our work over the past three years has sought to make this information more accessible: The 2012 Enhanced Disclosure Guidance (see Appendix Two) provided for average brokerage costs and transaction taxes to be available alongside information on bid/offer spread. The 2014 Statement of Recommended Practice (SORP) also requires a table of financial highlights that includes figures for ongoing charges and direct transaction costs expressed in the context of the size and performance of the fund. An illustration is presented below in Figure 2. The SORP also requires the spread on the portfolio at the end of the financial year to be disclosed in the notes to the accounts Implicit costs Implicit costs are those that do not result from any fees being paid as a separately identifiable amount by one party to another. The bid/offer spread is a commonly encountered implicit cost, as most financial instruments are available at any moment to be bought at one price, but to be sold at a lower one. Accordingly, at some moment in time, a fund seeking to buy a bond would need to pay 110p and a fund selling at that self-same moment would only receive 108p. This cost of translating cash into a financial instrument is not explicit but it is a real cost borne by investors and which reduces the total amount of capital that they have that is being put to work. No cash has been transferred by the fund for an identifiable transaction fee separately from the purchase price, but a brokerage firm or investment bank has earned revenue in the process. This distinction is captured in Principle 5. PRINCIPLE 5: The distinction between quantifiable explicit costs and indicative implicit costs which can only be estimated should be recognised. Explicit and implicit costs cannot be quantified with the same degree of accuracy due to their profoundly different nature. This is, however, a necessary cost. It is not possible to access the capital markets and deal with market-makers (which is the predominant model in bond trading, for example) and expect to be able to do so without a cost. The market-maker provides the service of connecting buyers and sellers without these parties having to seek each other out and the spread is the cost of this service. In the absence of this service, the transaction would be much harder to complete and the costs of doing so would likely be higher. This is an important point that distinguishes spread from other types of implicit cost as we explain below. Spread is considered to be observable at an agreed valuation moment but, importantly, it is rarely observed 6 (let alone actually quantified) at the very moment at which trades are executed. The size of the spread is, obviously, also dependent on the size of the order. Small orders for equities may fit within an exchange s best quoted bid and offer prices, but large orders would not, since it may be more difficult for a broker to find another party(ies) for a given trade. Generally, the nature of the market structure means that it is much harder to measure spread for bond transactions. A bank willing to sell a particular bond may not choose to reveal at what price it might be willing to buy that bond. And of course if it is wanting to sell, it is likely that the price at which it is willing to buy will be worse than the price which some other bank might be wanting to buy that same bond. At this moment, there is not always complete information in the market which is publicly available from which all these numbers can be certainly or easily ascertained. Nevertheless, as we explain in Appendix 3, it is already the case that fund managers carry out a valuation process on every asset in the portfolio and periodically make an assessment, on a particular date, of the average spread for the portfolio. This information is used as part of the pricing process and will be assessed through the proper systems and controls for assessing valuations at least quarterly and perhaps more frequently for some funds. The 2014 SORP also requires spread to be disclosed as part of the audited accounts of authorised funds (unit trusts and OEICs). This is a snapshot at the year end, but could be considered no less valid than historical data as an indicator of what might happen in a future period. However, we note that both approaches should be used with care in terms of predictability. 6 Detailed forensic analysis of the transactions that took place in the market at the time a trade took place is sometimes performed and from this the apparent spread can be deduced. The cost of such third party expert analysis means it is usually only commissioned by institutional investors. Page 11 of 60

34 The rules about ex ante disclosure for funds are highly prescribed by regulation. However, we think that it would be appropriate for firms to disclose their estimates of the expected impact of transaction costs (including spreads, commissions and stamp duty) used in their pricing process 7. We recommended disclosure of these estimates as part of the 2012 guidance on enhanced disclosure of fund charges and costs. Greater information on bid-offer spread Spread as an implicit cost is of particular significance since nearly all bond trading costs are financed within the spread and it would be counter-intuitive to suggest bond funds incur no transaction costs. The average spread across the financial instruments in a fund is therefore a relevant piece of information for investors. The Investment Association proposes that funds should now make public, the following: 1. The estimate of the bid/offer spread on the underlying portfolio being used for fund pricing at least quarterly and as soon as practicable after it is changed. 2. The estimate of the dealing commission being used for fund pricing at least quarterly and as soon as practicable after it is changed. 3. A narrative disclosure explaining these estimations. 4. Details of the prior disclosures and a comparison of actual transaction costs audited in the last period with the estimated costs disclosed under this proposal. We are conscious that in the context of MiFID and PRIIPs, the pre-sale disclosure of charges and transaction costs, and the way in which they could be aggregated, is being closely examined. We do not believe that there is currently a robust methodology to deliver such an output for implicit costs: The weaknesses of portfolio turnover metrics (explored in the next chapter) mean that multiplying historic Portfolio Turnover Rate (PTR) by a cost metric based on estimated spread is likely to give in many cases a very poor estimate of future outcomes. An alternative might be to capture the value traded and apply an estimated cost metric. This would mirror the approach suggested for ex-post disclosures by a Federation of Dutch Pension Funds paper in While this is technically possible, we are concerned that this will not provide meaningful or reliable information about future costs. Another issue relates to the completeness and accuracy of trading data. To achieve higher levels of accuracy for implicit costs, then there will need to be high quality order recording in all markets, including for FX. This area of disclosure will be a key part of the technical debate. The steps proposed above to make more information available on spread using existing pricing methodologies is a practical further step forward Quantitative vs qualitative implicit costs There is an additional family of costs discussed by commentators. These arise variously from the response of the market to any trading or known intent to trade and also from any impacts from the timing of a trade or delay in getting an order to market once it has been decided upon by the fund manager. This has been discussed in the US and the EU. The Plexus iceberg is often the example used to describe how there are many components to cost; and in the conditions then being considered by Plexus, they are of significantly different sizes. As an example of this discussion, a paper last year by The Pensions Institute 9 included the following under a reference to hidden costs: 1. bid-ask spread 2. transactions costs in underlying funds 7 This information is already required in the prospectus of single priced funds in the form of an estimate of the swing or dilution levy. 8 Federation of the Dutch Pension Funds, Further Elaboration on asset management costs, David Blake, On the Disclosure of the Costs of Investment Management, Pensions Institute Discussion Paper PI-1407, May Page 12 of 60

35 3. undisclosed revenue 4. market impact 5. information leakage 6. market exposure 7. missed trade opportunity or market timing costs 8. delay costs Citing the Plexus Research, the paper stated that the visible costs were 9 basis points (bps), bid-offer spread, market impact and information leakage were 12 bps, delay costs 21 bps, and missed opportunity costs 9 bps. This led to a summary of visible costs (commissions, taxes, fees and custody charges) at 18% of total costs and hidden costs at 82% of total costs. We continue to believe such an approach mixes very different costs. It does so by discussing what Wayne Wagner (former CEO of Plexus) described as economic costs that occur before the trade interacts with the market in the same terms as the costs of trading itself: It is part of our argument that item 1 (bid/offer spread) is an implicit quantitative cost that should and can be disclosed. Indeed, it is already in the SORP and it informs fund pricing. The inclusion of item 2 (transaction costs in underlying funds) is, in our view, a category error as all the components of the cost of those underlying funds must necessarily fall into one of the other categories of costs identified. The treatment and disclosure of costs in products such as funds of funds is worthy of discussion and we return to this subject in Chapter 3. The receipt of undisclosed revenue as in item 3 is not acceptable and we do not recognise this as occurring in UK funds. In the rare event that a fund may share revenue with its manager, such as for stock lending, the 2014 SORP makes it clear that the revenue should be shown gross and not used to reduce apparent costs by off-set. We believe that items 4 to 8 are of the implicit qualitative kind that do not form a part of a quantitative disclosure of costs. However 4 to 8 are described, what they have in common is that they are all estimates of how well a given decision to buy or sell an investment is implemented in the relevant markets. What makes these costs different from other implicit costs like spread is that these costs represent value lost to the market as a consequence of how and when the decision to trade was brought to market. The cost of spread represents value transferred to counterparties transacting with the fund as a consequence of the actual execution of a trade. A simple example of the difference might be where a fund manager wants to sell a large holding in a company's shares. It may in the usual course of things take many hours or even days to be able to find enough buyers. If participants in the market observe that the manager is a large seller, but has only managed to sell a small amount so far, they will lower the prices at which they are willing to buy. So if the manager is careful and does not let the market guess its intentions, then the overall price obtained may well be better. Accordingly, where the trading is "clumsy" there is an economic cost in the sense that the fund does not do as well as it might - for example because the sale raises less than it might have in the prevailing market conditions. This explains why, in our view, debates about implicit costs - and consequently demands that suggest a single number could be used - conflate two different but connected questions of interest to consumers, both of which can be formulated as the question What have the costs of transactions been? The first way of putting this question is to ask: What has been the actual cost of the transactions carried out by the fund? The second is to ask: Were the prices obtained, or paid, as good as they could have been? Importantly, we consider that explicit costs and spreads fall into the first type of disclosure - and disclosure about market impact falls into a second distinct category of information relating to best execution. This is captured in Principle 6. Page 13 of 60

36 PRINCIPLE 6: Standardised charge and cost disclosure requirements should deal only with tangible costs arising due to actual events. Intangible opportunity costs such as market impact should not be part of such requirements and can be looked at separately. We think that obligations to make disclosure about costs and charges should distinguish between those which describe: The quantity of actual prices incurred (including within spread) and paid to market participants to access the capital markets; and, The quality of trading and so relating to market impact and opportunity cost. At the heart of The Investment Association s position is that there should be full disclosure of all charges and costs in that first category, but that distinctions have to be made between what can be accurately disclosed in advance and the full historic accountability that comes with actual trading activity. Table 2 summarises our view of the challenges in this area. Table 2: Quantifying different kinds of transaction cost Type of cost Forward-looking Historic Commission Bid-Offer Spread Market Impact Quantum difficult to predict with certainty Difficult to predict with certainty; easier as time horizon to trade narrows Impossible to predict with any meaningful degree of accuracy, unless close in time to trade Hard costs can be exactly quantified Good general approximations can be made for specific situations Detailed forensic analysis can give useful generalisations against a range of possible industry benchmarks There should be disclosure about the quality of trading as well, but it is of a different nature (Principle 7). Despite the large number of papers that have been written about costs, we have seen no persuasive methodology for addressing all costs in standardised client disclosure documents, in particular implicit costs. We think that is because there is no single approach that can or should be applied across these different category issues that we describe as quantity and quality categories. There are a number of commercial services available to assess quality. For example, one Investment Association member firm uses an expert third party to take data on all their trades. Overnight, it produces a bell curve analysis, providing data on the quality of trades, by the slippage in price between passing to the trader and execution. All those trades outside two standard deviations from the mean are queried and trader rationales and comments recorded. Another firm uses reversion to measure how good a job the trader has done. PRINCIPLE 7: A calculation of total quantum of transaction costs can tell you little about how well or how badly a manager is trading within a given market. Different forms of quantitative assessment exist and may be needed by some clients. We do not consider the iceberg analogy used by Plexus - and so widely quoted by other bodies supports the view that investors should be given one number, or that the metrics about quality of trading have the same level of certainty as those which relate to commission payments. That said, asking and answering the question about how well a trade idea was formulated and brought to market and whether the timing and nature of that process was less efficient than it reasonably should have been, is a legitimate concern of investors. But it is not the same as disclosing the quantity of costs incurred in trading and it should not be aggregated as if it were. Page 14 of 60

37 1.5. Further considerations The qualitative dimension we discuss with respect to implicit costs also has a parallel in the assessment of explicit costs. Total quantum of trading costs may be of less interest to some than the overall quality of the execution. Market impact and associated issues are one aspect. Another is the cost of brokerage, where commission is being paid by the fund, as would typically be the case with equity trading (although equities can be traded with the broker acting as principle). As we show in our analysis of transaction cost variability, there are ways of looking at this that can offer a simple and comparable approach to the issue (expressing total transaction costs relative to total transaction value) as well as more sophisticated measurements and benchmarking services. Page 15 of 60

38 CHAPTER 2: PORTFOLIO TURNOVER RATES In the context of a debate about transaction costs, we also think it is useful to consider the role of Portfolio Turnover Rate (PTR) metrics. Such metrics have been used by regulators in various jurisdictions internationally to provide fund investors and their advisers with a framework to help understand activity levels within a fund (ie. how much buying and selling of stocks and securities is taking place). This has partly been driven by a concern over the connection between turnover and transaction costs, and specifically in the US by the potential taxation implications for individual investors. There are two well-known methodologies for calculating portfolio turnover rate; the European Commission methodology for UCITS and the US Securities and Exchange Commission (SEC) methodology. Also, there are variations to both methods that are sometimes used. This chapter looks in detail at the how the European Commission and SEC PTR metrics operate. It also tests three alternatives based on the UCITS and SEC methodologies and concludes by proposing a modified version of the SEC approach Background The European Commission introduced the PTR in its 2004 Recommendation 10 on the contents of the simplified prospectus as an additional indicator of the relevance of transaction costs. However, in developing the 2010 KIID Regulation, it was recognised 11 that a Portfolio Turnover Rate (PTR) is likely to be too opaque for the retail investor to interpret without detailed explanation of its significance and it was excluded from the KIID. In its place a requirement was introduced to give a narrative warning where the impact of portfolio transaction costs on returns is likely to be material. In the US, the SEC enhanced its PTR disclosure requirements 12 in 2009 by adding a requirement to include the latest PTR figure, together with a brief explanation of the effect of portfolio turnover on transaction costs and fund performance, in a summary prospectus. The explanation was to be included to address concerns regarding the degree to which investors understand the effect of portfolio turnover and the resulting transaction costs. The pre-existing requirement to present a five-year record of the PTR in the full prospectus remained intact, as did the methodology for calculating it. In developing the summary prospectus the SEC considered including the impact of transaction costs in a fund s expense ratio but concluded that it was not feasible at that time because of the absence of an adequate basis for prescribing a specific and accurate methodology for reflecting transaction costs in a fund s expense ratio. In publishing their rules for the summary prospectus the SEC stated that there is not necessarily a direct correlation between portfolio turnover rate and portfolio transaction costs. Nonetheless, in the absence of a basis for prescribing a better measure, we believe that portfolio turnover rate, though imperfect, is an appropriate indicator of transaction costs Key issues It is evident that the PTR is designed to be an indicator of transaction costs. (ie. the higher the PTR, the higher the transaction costs incurred by the fund). However, both in Europe and the US, its use has been controversial and contested. It is widely acknowledged that it is difficult for investors to understand and therefore, if it is used, it requires explanation. Where it remains in use, it is recognised that it is an imperfect measure and is not sufficiently reliable to be used to quantify costs. There is also a particular need for clarity about what the PTR is designed to show, and how it might be applied. As we discuss below, both the UCITS and SEC methodologies are, in effect, attempting to identify discretionary levels of turnover: ie the extent to which the manager has chosen to transact. As discussed below, the extent to which this is appropriate depends on the fund manager s pricing policy. A more detailed overview of pricing policies is also presented in Appendix 3. Importantly, the inherent imperfections in the methodologies for calculating PTRs and the interaction with the pricing policy necessarily means that a PTR metric cannot be used as a multiplier in a quantification of total transaction costs. 10 Source: European Commission Recommendation 2004/384/EC 11 Source: CESR consultation paper CESR/ Source: SEC release Page 16 of 60

39 2.3. UCITS methodology The UCITS calculation attempts to identify discretionary trading separately to the trading caused by inflows or outflows. To do this the approach deducts the sum of inflows and outflows from the sum of purchases and sales and then divides the result by the average net asset value. The effect is to maximise the extent to which trading activity is deemed to be attributable to inflows and outflows. The UCITS calculation 13 can be represented as: PTR = (Purchases of securities + Sales of securities) (Issues + Cancellations) Daily average Net Asset Value A variation of the UCITS approach is to ignore the adjustment for net flows. This reduces the calculation to being the sum of purchases and sales divided by the average net asset value SEC methodology The SEC approach divides the lesser of purchases or sales by the average portfolio value. It does not attempt to adjust for inflows or outflows, but implicit in the calculation is the assumption that the amount by which purchases or sales exceed the other does not contribute to turnover. The SEC approach 14 can be represented as: PTR = Lesser of Purchases or Sales of portfolio securities Monthly average Value of portfolio securities A variation of the SEC approach is to apply this formula on a monthly basis and then to sum the resulting monthly figures over a year Analysis of the methodologies Common to both methods is the challenge of identifying the purchases and sales that are relevant as indicators of transaction costs. For example, some funds operate a policy of sweeping uninvested cash balances to a cash fund on a daily basis in order to mitigate counterparty risk. The frequent purchase and sale of shares in the cash fund do not contribute to transaction costs but may inflate considerably the total purchase and sale figures used in the calculation and will therefore cause the PTR to be distorted. Moreover, where inflows or outflows are settled by transferring stock in specie (ie. without having to go to the market to transact), these transfers are recorded as purchases or sales but do not cause transaction costs to be incurred UCITS methodology The basis for the UCITS calculation is that purchases and sales caused by inflows and outflows can be separated from transactions used to express a fund manager s investment decisions. Excluding inflows and outflows means that the PTR is an indicator of the relevance of the transaction costs that arise only as a result of discretionary investment decisions. However, whether or not it is appropriate to discount the transaction costs arising as a result of inflows and outflows depends on the fund manager s pricing policy. Where the pricing policy is to use a mechanism 15 to protect the fund from the dilutive effect of inflows or outflows, the investors that cause these flows make a payment to the fund to cover the resultant transaction costs and the existing/ongoing investors do not experience these costs. Where the pricing mechanism provides such protection, it will be appropriate to exclude trading activity resulting from inflows and outflows from the PTR 13 Source: Annex II of Commission Recommendation 2004/384/EC 14 Source: Item 13 of Form N-1A 15 A discussion of pricing policies is presented in Appendix 3 and a full explanation of the relationship between the pricing mechanism and the effect of transaction costs can be found in the IMA Guidance: Enhanced disclosure of fund charges and costs. Page 17 of 60

40 calculation; this approach underpins Method 1 later in this chapter. However, where such a mechanism is not used, or is only used under certain conditions, the existing/ongoing investors do experience transaction costs arising from inflows or outflows and the adjustment for flows is not appropriate, or may only be appropriate in part. In such circumstances, the variation of the UCITS methodology that ignores the adjustment for net flows might become a more appropriate indicator of the relevance of the transaction costs; and underpins method 2. In reality, it is impossible to identify with certainty whether trading activity relates solely to flow or is the result of discretionary decisions. Fund managers will identify stocks they wish to buy or sell, and if net flows are in their favour, they will be able to implement their decisions using the net flows. For example: The managers of Fund A and Fund B both wish to invest in Stock X. Fund A has net inflows and Fund B has no flows. The manager of Fund A can use the inflows to pay for Stock X. Assuming the pricing policy is to protect the fund from dilution, the existing investors will experience no transaction costs because the incoming investors will have paid the cost of buying Stock X. However, the manager of Fund B must sell stocks in order to pay for Stock X. Investors will suffer the transaction costs both for selling stocks and for buying Stock X. By adding together purchases and sales, the result is twice what would be expected. For example, if the fund manager sells all the investments and replaces them with different holdings, it would be reasonable to expect a representation of turnover to be 100% ie. the whole portfolio has been turned over once. However, the UCITS calculation gives a PTR of 200%. Moreover, the calculation is theoretically prone to giving negative PTRs. For funds in the UK post RDR this will be a particular problem because of the proliferation of share classes with new classes open to new business and legacy classes effectively closed and in run-off mode. For example, a fund worth 80m has two share classes: a large legacy class and a small RDR class. Over the course of the year it issues 80m of RDR shares and cancels 40m of legacy shares. As a result, the fund purchases 40m of stock and there are no sales. The PTR will be - 80%. 16 It is widely accepted that a negative PTR is meaningless, 17 but the same effects will be at work when the resultant figure is positive SEC methodology The basis for the SEC approach is that turnover occurs when stocks are sold in order to buy alternative stocks and that this turnover is indicative of the level of transaction costs. The extent to which purchases or sales exceed the other is disregarded in the calculation. In broad terms the difference between purchases and sales is indicative of the net flow into, or out of, the fund during the year. This approach appears more reliable; if the entire portfolio is turned over once in the year the PTR will be 100% and the result always will be a positive number. However, the approach is unable to cope with circumstances where substantially all trading activity results from net flows and the pricing mechanism provides protection from dilution; typically this will be the case for passive funds that replicate an index. For example, consider a tracker fund with an average value of 100m that experiences inflows, and hence purchases, of 40m and outflows, and hence sales, of 25m. Turnover might be expected to be nil because there is no discretionary trading activity, however in this example the SEC calculation will give a PTR of 25% Purchases + sales = = 40. Issues + cancellations = = 120. Average NAV =100. PTR = (40 120)/100 = -80%. 17 The problem arises because each share class is a separate legal instrument and accounting standards prohibit offsetting for the purposes of the accounts. Therefore, the accounts must show the gross values of shares issued and cancelled even though they may have been issued and cancelled on the same day. However, for the purpose of calculating a PTR it should be recognised that trading activity is caused by the net flow on a given day. As a result, it is the net flow on each day that should be used to determine the amount of issues and cancellations in the year. In the example, using daily netting would have given a PTR of 0%. 18 Sales are less than purchases so PTR = 25/100 = 25%. Page 18 of 60

41 2.6. Refining the existing methodologies Our view is that the PTR should always be calculated for a full year and not extrapolated from a shorter period. It is not possible to annualise reliably figures calculated for periods other than a year because purchases and sales do not occur evenly throughout the year. It should be calculated for funds and not individual share classes because it is an indicator of portfolio activity. Many of the shortcomings identified in the analysis in the previous section can be addressed by defining more precisely the terms used in the calculations. In this section, we outline what we believe the definitions of those terms should be, presenting them as refinements to the two prevailing methodologies: Securities Both SEC and UCITS approaches refer to purchases and sales of securities. The SEC definition includes derivatives and makes adjustments to exclude holdings purchased with less than a year to maturity. Whichever calculation methodology is applied, our view is that the FCA s definition 19 of securities should be used. This includes shares, debentures, alternative debentures, government and public securities, warrants, certificates representing certain securities, units in collective investment schemes and rights to or interests in any of these securities. Securities do not include deposits, foreign exchange contracts, forward contracts or derivatives Purchases and Sales Purchases and sales that are relevant to the trading activity of a fund as an indicator of transaction costs should be included. Purchases, sales, primary issuance, corporate actions, calls, conversions, redemptions and maturities should be included. Where a fund operates a regular cash sweep to a money market fund or cash fund, these transactions will cause the PTR to be overstated. Although these are transactions in securities they should be excluded from the PTR calculation. Stock transferred in specie may be excluded provided, in the case of the UCITS calculation, the value transferred is also excluded from any deduction made for issues or cancellations Issues and cancellations The UCITS calculation deducts issues and cancellations which should be the sum of the daily net amounts for the fund. The summation of the amounts for each share class in the absence of daily netting would overstate these figures causing the PTR to be understated or even to be negative. Sales and redemption amounts should not be used as these figures are not representative of the cash inflows and outflows that cause trading activity in the fund. Conversions, or issues and cancellations used as a proxy for conversions, should be excluded. The value of stock transferred in specie should be excluded to the extent it is also excluded from purchases or sales Average net asset value The net asset value used in the denominator of the UCITS calculation should be the average of the net asset value calculated at every valuation point (typically daily) during the year. This figure will be the same as the denominator used in the calculation of the ongoing charges figure for the year Average value of portfolio securities The denominator in the SEC calculation is the value of portfolio securities calculated as an average of the opening value and the twelve month-end values for the year. It is adjusted to include only the securities included in the numerator. However, for the modified SEC methodology underpinning method 3 later in this chapter, the average net asset value calculated on the same basis as for the UCITS calculation should be used Empirical analysis Using modifications of the UCITS and SEC definitions, we compared the results of three potential methodologies by calculating PTRs using real data for ten funds (the same funds that we analysed in Appendix 4). We addressed the counter-intuitive outcomes under the UCITS based methodologies by dividing the results by two. 19 FCA Handbook glossary definition of securities. Page 19 of 60

42 The results are presented in graphical form in Appendix 5 and we produce one example below for ease of reference in Figure 3 Figure 3: Predominantly UK Equity Fund Method 1, UCITS methodology divided by two PTR = Half of (Purchases + Sales) (Issues + Cancellations) Average Net Asset Value The results of this method are shown as a navy line in the graphs in Example 1 and Appendix 5. This gives the lowest measure because it takes full account of inflows or outflows due to investors joining or leaving the fund. This represents the significance of discretionary turnover as a driver of transaction costs. In effect, it assumes that the pricing policy always ensures that all of the costs caused by investor flows are collected from those investors. This method relies on issues and cancellations for each share class being netted on a daily basis to give an accurate indicator of the net flow driving portfolio activity. Failure to do this will cause the PTR to be understated when flows in both directions occur on the same day. We have not been able to net flows in this way with the data available. The effect of this understating is most apparent in example two, where the navy line shows the turnover being negative. This is a tracker fund so there is negligible discretionary turnover. This phenomenon might also be in evidence in example seven Method 2, method 1 without an adjustment to reflect investor flows PTR = Half of Purchases + Sales Average Net Asset Value The results of this method are shown as a green line in the graphs in Example 1 and Appendix 5. This gives the highest measure because it takes no account of inflows or outflows due to investors joining or leaving the fund. This represents the significance of turnover as a driver of transaction costs because costs are incurred for trading regardless of whether the cause is discretionary or flow driven. However, it ignores the fact that the pricing policy of most funds will ensure that some or all of the costs caused by investor flows are collected from those investors. Page 20 of 60

43 Method 3, modified SEC methodology PTR = Lesser of Purchases or Sales Average Net Asset Value The results of this method are shown as a coral line in the graphs in Example 1 and Appendix 5. This gives a measure somewhere between Methods 1 and 2 because it takes account of the net inflow or outflow over the year (so is lower than Method 2) but does not take full account of inflows and outflows (so is higher than Method 1). The more a fund experiences one-way flows throughout the year, the nearer this method becomes to Method 1 (coral and navy are close). The more a fund experiences inflows and outflows during the year, the less effective this method is at taking the flows into account and the closer this method becomes to Method 2 (coral and green are close) Conclusions Notwithstanding the areas of weakness inherent in any methodology, all three lines in the charts tell a very similar story for each of the real world funds chosen. With proper daily netting of issues and cancellations it is likely the three lines would be even closer together (the navy line would rise towards the others). Therefore, it would appear that in practice, and since investors will wish to observe changes in the PTR over time and relative PTRs between otherwise similar funds, it may make little difference which method is used. This being the case the choice of methodology becomes one of operational pragmatism, which would lead to recommending Method 3: the modified SEC methodology. However, it is important to recognise the accuracy of the calculations whichever method is used. In reality trading due to flows and discretionary trading are often inextricably linked. The examples show the results of the three methods are typically in a range of about 20%. This margin of error is acceptable provided the measure is used to understand the relative level of trading (for example, high medium low). However, the error range is too large to use the number as an absolute measure of turnover, and care should be taken not to portray a spurious degree of accuracy. Accordingly, presentation will be of great importance and we discuss this further in Chapter 3. Page 21 of 60

44 CHAPTER 3: PRACTICAL PROPOSALS FOR PRODUCT NEUTRAL DISCLOSURE Building upon the principles established in Chapter 1 and respecting existing regulatory approaches, most notably the KIID regulation 20 and the supporting CESR guidelines for the calculation of the ongoing charges figure, we now propose a disclosure approach for discussion and comment. The intention is to answer, in the first instance, the need for Independent Governance Committees, trustees and other scheme decision-makers to obtain the necessary information and to calculate the charges and transaction costs borne by members of pension schemes during the scheme year. The approach is written with one eye on its application to: i. Defined Contribution (DC) pension contracts. ii. iii. Collective investment funds into which DC contributions may be invested. Funds of funds, and any other arrangement in which costs are experienced at one or more removes through the packaging of products. There are three key points to emphasise here: 1. Presentation of OCF. The simplicity and consistency of the ongoing charge approach should make it the starting point for charge disclosure, but will need to be adapted accordingly for the pensions environment (for example, to take account of appropriate administration and other costs). Figure 4 shows what is in and what is outside the existing regulation of ongoing charge disclosure for UCITS. As can be seen, the rules recognise that whilst a number of different categories of costs can be identified in advance, the quantum of some of those costs, and hence their effect on returns, is sometimes not capable of expression before the event, either as an absolute amount or as a percentage. Due to the bundled nature of transaction costs and some research expenditure, it is appropriate for now to include research expenditure within the transaction costs and so outside the ongoing charge figure. [1] However, this area is currently subject to regulatory change. 2. Treatment of unit-linked. We propose a framework which ought to be transposable to funds or products which exist as a set of liabilities on a balance sheet, as would be the case for unit-linked funds or other insurance contract based investments. Straightforwardly it requires any costs associated with the product to be treated as if that unit-linked fund were indeed a separate set of segregated assets. Any provider will be doing this in order to produce valuations as well as measuring and accounting for fees and charges. 3. Treatment of fund of funds. On the question of fund of funds, or strategies operating across multiple funds or asset classes, we note competing pressures. Our technical view remains that extracting transaction costs in an accurate manner across such structures could prove extremely challenging, and we would welcome views on the practicalities in this area. However, we recognise that in the DC pension environment, there needs to be a way to communicate to individuals that transaction costs are not zero simply as a function of an inability to aggregate across different vehicles or asset classes within a fund. One potential methodology would be to develop a weighted inclusion of costs in underlying products such as funds of funds. In this case, on every valuation date and so daily for most packaged products the weighted contribution of underlying funds costs will be measured. So in a fund of funds, on each valuation date, the fund of funds manager will record what percentage of the value of the fund is represented by each of the underlying funds; disclosure ex post would be made by applying that weighting to the costs and charges of each underlying fund and aggregating it. 20 And in future the PRIIPs KID regulation and ESMA guidelines thereunder [1] Detailed disclosure of research expenditure is currently provided in the Pension Fund Disclosure Code and its mirror for investment funds, the CIS Disclosure Code. As the regulatory landscape around research and dealing commission becomes clearer, we will revisit these Codes, which could also be repurposed in the context of the current discussion about transaction cost disclosure. Page 22 of 60

45 Figure 4: Calculating the ongoing charges figure Page 23 of 60

46 3.1. Summary framework for disclosure An overriding principle behind this paper is to ensure that clients have access to fair, clear and meaningful information on charges being paid to investment managers and trading costs incurred in delivering services. The approach we use is consistent with our position in the European disclosure debate, but is also intended to provide an approach to inform the current consideration of transaction cost disclosure to pension scheme trustees and Independent Governance Committees as part of responsibilities under impending DWP and FCA rule-making under the Pensions Act The approach should apply to all investment products distributed to the UK retail and pensions markets, and could be extended to capture all costs borne by consumers in pension accounts. Table 3: Disclosure framework for charges and transaction costs 1. Ongoing charges shall be calculated and described on a consistent basis. We suggest that they can be calculated in accordance with the approach in paragraph 2.2 below. Where the ongoing charges figure results from a series of separate headings of charge, then separate disclosure of each item should be made available if requested. 2. Contingent (performance fees) or one-off charges (such as entry/contribution charges) shall be expressed separately. 3. Transaction costs shall be presented on a basis that makes a clear distinction between what can be known with certainty (historic figures) and what can only be estimated as a future experience. This is particularly important for point of sale disclosure, to avoid confusion across different asset types with different market cost structures, and to recognise the variability of transaction costs. 4. Explicit and implicit costs shall both be presented, recognising that not all transaction costs can be quantified with the same degree of accuracy due to their profoundly different nature. Different kinds of implicit cost are too often conflated and that has hampered the quality of the disclosure debate. Implicit costs that relate to the bid-offer spread should be disclosed, though not as if they were explicit costs. In contrast, the far more qualitative and intangible assessments of market impact should be addressed in work on the quality of execution and not as part of standardised costs and charges disclosure work. 5. Charges and transaction costs shall be separately identifiable. If a single figure combining charges and transaction costs is insisted upon in regulation, it must be on the express basis that it is still possible to separate key elements, notably the fee being paid for a portfolio management or fund service, from variable transaction costs incurred as part of the service. 6. Other disclosures shall include: stock lending and repurchase agreements; financing costs; and taxes. 7. On an historic basis, both charges and explicit transaction costs can be expressed in monetary terms. On a forward looking basis, where monetary amounts are not known (entry or contribution charge being a possible exception), expression should be as a percentage of average net asset value. 8. Portfolio turnover rate presented in accordance with a standard methodology, based on the SEC approach. 21 Sections 44(1) and 44(2) of the Act require the DWP Secretary of State and the FCA respectively to make regulations regarding the disclosure and publication of information on transaction costs incurred by a relevant scheme and of the administration charges imposed on members of a relevant scheme Page 24 of 60

47 Building on analysis in Chapter One, Table 5 gives an indication of how different components of charge and transaction cost information can be made available at different stages of the investment process, and with what degree of accuracy. Table 4: Availability of different components (ex post vs ex ante) One off charges Entry Exit Basis Ex ante % Ex ante Ex post % Ex post % of capital contributed % of proceeds realised Maybe known Maybe known Known Known Known Assumed proceeds Known Known Ongoing charges % of product value Known Assumed product value Known Known Performance fees % of outperformance Known Unknown Known Known Transaction costs Commission and trans. tax (unless included in taxes) % of transaction value Spread % of asset price Market-wide estimate Good estimate Unknown Known Known Unknown Market-wide estimate Forensic analysis Financing costs % of principal and fee based Good estimate Maybe known Known Known Taxes Various Unknown Unknown Known Known Page 25 of 60

48 3.2. Definitions and methodology for product-neutral approach The text below provides a proposed methodology to capture the requirements set out in the summary framework in the preceding section. 1. The overarching principle All necessary information shall be provided by a disclosing party periodically and in good time to investors so as to provide them with a fair, clear and informative statement about all charges and costs in connection with any relevant product. 2. The components Any disclosure shall identify separately the following components: 1. One-off charges 2. Ongoing charges 3. Charges levied in certain specific conditions (such as a performance fee) 4. Transaction costs 5. Financing costs 6. Taxes 2.1. One-off charges One off charges comprise (however described): (a) any entry charges or contribution charge which might be deducted from the investor's capital commitment to any relevant product; (b) any exit charges or withdrawal fees or penalties which might be deducted from the product s proceeds before they are paid to the investor; and (c) any charges for investor-initiated switching between investment or other options within the product. Each type of one-off charge shall be disclosed separately Ongoing charges Ongoing charges are payments deducted from the assets of any relevant product over the defined period where such deductions are required or permitted by law and regulation and any constitutional instrument of the relevant product. Ongoing charges shall include all types of cost borne by the relevant product, whether they represent expenses necessarily incurred in its operation, or the remuneration of any party connected with it or providing services to it, being: a) all payments to, or in respect of, the following persons (including legal persons), including any person to whom they have delegated or outsourced any function: - any person responsible for the governance, oversight or management of the relevant product; - any person responsible for the safeguarding or custody of assets of the relevant product; - any investment advisers or consultants; - providers of administrative services such as valuation and accounting services; - ownership service providers, such as transfer agents and registrars recording ownership of the relevant product, or entitlements under it; - providers of legal and professional services, including advisers, consultants, actuaries, lawyers and auditors; - providers of services related to the assets of funds; b) operating charges that should normally be included as ongoing charges but are met in whole or part by fee-sharing arrangements, such as with a custodian on stock-lending income; c) one-off and ongoing charges levied by, or in respect of, underlying products where the relevant product invests in other products; d) all registration fees, regulatory fees and similar charges; e) all costs for investor communications such as producing reports and statements, printing and postage costs f) any costs of marketing and distribution, including platform costs where the platform forms part of the service being offered; and g) any other ongoing charge. Page 26 of 60

49 All items included in (a) to (g) above shall be assessed including any irrecoverable sales taxes. The following shall not be included as ongoing charges: i. one-off charges included in 2.1 above ii. contingent charges such as performance fees included in 2.3 below; iii. transaction costs included in 2.4 below; iv. finance costs such as interest on borrowing included in 2.5 below; v. taxes, other than irrecoverable sales taxes, included in 2.6 below; vi. payments incurred for the continuance of open financial derivative positions (for example, margin calls) Contingent charges Charges levied in certain specific conditions, such as performance fees, should be disclosed separately alongside the figure for ongoing charges Transaction costs Any expense that is not a charge that will fall to be deducted from or borne by the relevant product, being: i. transaction costs necessarily incurred in connection with the acquisition or disposal of any asset for the relevant product, where those costs are explicit (e.g. brokerage charges and linked charges), except: a. transaction-based payments made to the persons listed in 2.2(a) above; and b. the costs of acquiring or disposing of other funds, which shall be included in 2.2(c) above; ii. spreads and mark ups embedded in the transaction price of any asset acquired for or disposed of for the relevant product (e.g. bid-offer spread for dealing in fixed interest securities); and iii. any other transaction costs consideration will be needed as to how to represent costs in illiquid alternatives, such as some private equity investments Financing costs Financing costs should be disclosed, including interest on borrowing, debt arrangement fees and early settlement penalties. Costs should be stated after taking account of the effect of any hedging arrangements, such as an interest rate swap Taxes Any taxes incurred within the fund, such as taxes on profits, irrecoverable withholding taxes and other taxes associated with the holding of assets, should be disclosed. 3. The Methodology 3.1. Information and reporting differences Forward-looking information and historic reporting and accountability are different. Forward-looking Forward-looking disclosure should reflect the qualitative difference between some charges and some costs for example, whether they are fixed or a percentage; and if a percentage, whether they are applied to known monetary amounts or to monetary amounts that cannot yet be known (such as performance fees); or whether they cannot be known in advance (such as transaction costs). These costs may be expressed or calculated in a way appropriate to their nature (for example, a flat fee, a proportion of assets, a charge per transaction). Forward-looking disclosure should be expressed in terms that look one year ahead and should not compound charges or costs over time, because many elements will be unknown. Historic Historic disclosure should include all charges and costs recorded during the period of account Period In every case costs (except for spread) must be expressed on an annualised basis Quantum The figure to be disclosed shall be based on the total of all such payments made (and allowable provisions against such payments) over the specified period. Nevertheless this will be technically challenging in relation to spreads as explained elsewhere; and indeed for illiquid investments, such as some real estate, infrastructure and private equity investments. Page 27 of 60

50 3.4. Currency Charges and costs should be expressed in the base currency as defined by law and regulation or, if permitted, in the constitutional instrument. Where the base currency differs from the functional currency, all exchange differences resulting from translation to the base currency should be separately disclosed Accounting standards All other information shall be presented according to declared and accepted accounting standards as if the fund or product were a standalone entity Intangible products Where the product is represented by account entries and is in essence a liability on the balance sheet of a financial institution then the product will so far as practicable be treated as if it were the assets and liabilities which are associated with it for the purposes of deriving asset values, distributions and returns for investors. 4. Other disclosures The Investment Association is committed to achieving full and meaningful disclosure of costs and charges. Consumers of financial services should be provided with consistent, clear, understandable and meaningful information on all costs and charges. We have explained why we think some implicit costs are about quality of trading. But in addition to the quantitative disclosures of cost we consider there are additional disclosures that could be made Stock lending and repurchase agreements Fees and expenses associated with stock lending and repurchase agreements should be disclosed showing the gross fees, related direct and indirect expenses and the net amount retained by the fund. The identity of the recipient(s) of the expenses and their relationship to the any of the parties listed in 2.2(a) should be given. The average value of securities on loan or subject to a repurchase agreement during the period should be stated together with the nature and value of collateral held Pricing policy The pricing policy should be explained, including how amounts may be recovered from incoming and outgoing investors in order to protect ongoing investors from transaction costs and an indication of how effective the policy is in protecting ongoing investors. The amounts expected to be recovered should be quantified as a percentage of the relevant price Narrative disclosure In addition asset managers should include a narrative disclosure and discussion of the transactions costs incurred with regard to the PTR in their annual report. Page 28 of 60

51 3.3. Wider application in UK pension context While we recognise that product wrappers such as a pension or an ISA are not the same as an investment fund, we see no reason why the OCF terminology could not be extended, subject to appropriate underpinning methodologies. With respect to the introduction of the charge cap for DC automatic enrolment default strategies, we note the emergence of a new term, Member Borne Deduction (MBD). We urge the Government, regulators and the pensions industry to use the term MBD only in the context of charge cap compliance, not consumer disclosure. This will help to facilitate the kind of consistent, intuitive messaging on charges that Government, regulators and industry agree is necessary. In practical terms, this affects two levels for UK DC automatic enrolment (see Figures 5 and 6): Disclosure by service providers to schemes, particularly as the latter work to ensure that they are charge cap compliant in the delivery of default arrangements. Disclosure by schemes to end consumers (ie. what is the overall charge for the pension scheme being borne by the individual). Figure 5: Disclosure to UK pension schemes for investment services Investment services Investment Association Proposal: Pooled vehicles (authorised funds, life funds, other) to use consistent charge methodology and terminology for recurring charges OCF to be standard. This debate mirrors that taking place on investment funds specifically. Firms may choose to continue using separate components within OCF. But presentation and calculation of total would be standardised. This would apply to all investment components, but where default arrangement (and subject to charge cap), standardisation should facilitate scheme decisionmaking and compliance. Transaction cost disclosure (outside cap) Investment Association Proposal: UK regulation in this area required by The Pensions Act 2014 (anticipated Summer 2015), but current Investment Association position paper intended to inform clear definitions of explicit and implicit costs, and provide accompanying metrics on portfolio activity levels Once UK and EU regulatory requirements (notably MiFID) are clear, The Investment Association to revisit current Disclosure Codes (Pension Fund Disclosure Code, Collective Investment Scheme Disclosure Code) to explore common template for information provision by industry. Figure 6: Disclosure by UK pension schemes to scheme members With respect to those charges borne by scheme members Investment Association Proposal: Disclosure should use standardised language and methodology. Adapt OCF for total charges in individual pension accounts. For some providers, other applicable fees alongside (contribution charge, separate administration fee). Transaction costs Schemes to take view with regulators on appropriate communication of transaction costs to members. MBD calculation should only be used for cap compliance (eg. is total MBD <0.75% using combination of contribution and ongoing charges?). Our current understanding of the MiFID and PRIIPs proposals (which are intended eventually to encompass UCITS funds) is that the overall approach to disclosure recommended in this document is compatible. Where the European regulatory authorities envisage aggregate figures for both ex ante and ex post disclosure, we continue to believe that the only way to ensure meaningful information is to disaggregate the components. In other Page 29 of 60

52 words, over time, both the totals and their components will be available. In any event, UK pensions law now requires disaggregation, which effectively imposes a set of disclosure requirements that are distinct from MiFID and PRIIPs. This mirrors the component disclosure approach in the UCITS KIID, and could look something like the charge template in that document. An example of a basic standardised template is given below. As per evolving regulatory requirements and scheme preferences, transaction costs can also be included in the template. Schemes would determine whether charge cap compliance information would also be shown where a combination of charges is used to determine compliance (eg. a contribution charge plus OCF). Figure 7: Possible approach for forward-looking charge disclosure for UK pension accounts Charges for this pension One-off charges taken before you save: Contribution charge [] [Other] Charges taken from the fund over a year Ongoing charges [%] Administration charge [] Estimated transaction costs in delivering investment return Framework to be agreed. Investment Association view that charges and costs most meaningful in the context of performance generated; may require narrative 3.4. Towards a common template? In the context of UK pension reform specifically, the intention of this paper has been to inform a debate where the outcome will ultimately be determined by UK regulators through For that reason, it is premature to produce a firm template (ie. agreed common format) for disclosure at this stage. However, once the shape of the proposed FCA/DWP regulation is clear, our intention is to work with regulators, industry and pension schemes themselves to try to ensure that a consistent approach can be developed. One possible framework to use is the Pension Fund Disclosure Code (mirrored for investment funds in the CIS Disclosure Code). The Disclosure Codes currently provide extensive quantitative and qualitative information on the level and impact of dealing costs and particularly on the use of dealing commission to pay for research (where the Codes have a role recognised in regulation of ensuring appropriate disclosure). Additionally, the Codes describe the manager s best execution policy, procedures and controls. The Codes will be subject to revision given further changes in the rules governing use of dealing commission, notably MiFID II implementation. If this revision can incorporate the elements described in this paper, then information would be available to clients in a single place. A key challenge here for UK regulators and the industry will be to ensure that all of this aligns with evolving European legislation governing portfolio management and investment fund services, notably at this juncture MiFID and PRIIPs. Page 30 of 60

53 3.5. PTR presentation It is apparent that the PTR can be an appropriate guide to the significance of transaction costs. However, as a number it has spurious accuracy and it should be used only with extreme caution. Moreover it serves as an indicator of transaction costs and is not a reliable indicator of holding periods. 22 One alternative possibility to presenting percentages is to provide consumers with a visual representation of levels of turnover and the relationship to transaction costs. We propose such a presentation in Figure 8. Taken together, the estimates of dealing costs disclosed in relation to the anti-dilution mechanism as recommended in Chapter One and the PTR figure provide a meaningful representation of the significance of transaction costs for a fund. Figure 8: Proposed presentation for illustrating portfolio turnover in relation to transaction costs The Investment Association recognises that it is the role of regulators, after consumer testing, to determine appropriate forms of disclosure and wishes to make it clear that such a proposal has not been consumer tested. How to disclose to different types of client and decision-maker is an area that requires significant research and consumer testing and is properly a matter for the relevant regulators. An institutional decision-making body, such as a board of trustees or IGC may have a completely different aptitude and set of criteria with respect to understanding and questioning transaction costs than a normal consumer investing in a unit trust or group personal pension. We note the significant work that was undertaken at European level as part of the creation of the UCITS KIID. 23 This research demonstrated the difficulty that consumers have in understanding a number of aspects of the investment process. The Investment Association encourages Government and regulators to ensure that charge and cost metrics are presented in a way that consumers will understand, and looks forward to playing a role in such a workstream. 22 Further analysis of the meaning of turnover can be found in an IMA research paper. 23 IFF Research and YouGov, UCITS Disclosure Testing Research Report, June 2009 Page 31 of 60

54 3.6. Next steps Our position paper summarises Investment Association thinking to date, and we would welcome comments and contributions about any of the issues raised. In terms of practical implementation, the issues discussed will also be the subject of regulatory outputs in the UK and EU in the near future: We are working with the FCA and DWP to assist the formation of cost disclosure regulation to be introduced in 2015 for the automatic enrolment pensions market. Under Pensions Act 2014, this will result in specific rules for pension scheme assessment of transaction costs, expected by summer 2015 and subject to forthcoming consultation. At EU level, MiFID and PRIIPs are at different stages of the legislative process. A path towards charge and cost aggregation is increasingly clear, but how transaction costs are defined and accounted for, and the future shape of the OCF, will be a major technical topic of consultation over the next 12 months. Comments should be sent to positionpaper@theinvestmentassociation.org. Page 32 of 60

55 APPENDIX 1: PRINCIPLES OF DISCLOSURE In the course of developing this paper, we also developed eight specific principles which help to explain our approach and these are reflected clearly in the recommendations: PRINCIPLE 1: Charges paid to an agent who invests money on behalf of a client should be expressed and calculated in a consistent manner. PRINCIPLE 2: The charges figures should be distinguishable from transaction costs, recognising the need to be able to compare fees between managers. PRINCIPLE 3: Transaction costs should be disclosed to help clients whether retail, institutional or other understand the economic experience of monies invested in a given market on their behalf. PRINCIPLE 4: The distinction between forward-looking and historic disclosures should be recognised. While estimates can be used for point of sale documents, only historic disclosure provides full accountability. PRINCIPLE 5: The distinction between quantifiable explicit costs and indicative implicit costs, which can only be estimated, should be recognised. PRINCIPLE 6: Standardised charge and cost disclosure requirements should deal only with tangible costs arising due to actual events. Intangible opportunity costs such as market impact should not be part of such requirements and can be looked at separately. PRINCIPLE 7: A calculation of total quantum of transaction costs can tell you little about how well or how badly a manager is trading within a given market. Different forms of quantitative assessment exist and may be needed by some clients. PRINCIPLE 8: Data is not the same thing as information and narrative should be used as needed to provide additional information about any of the quantitative information disclosed. Page 33 of 60

56 APPENDIX 2: SUMMARY OF INITIATIVES TO IMPROVE TRANSPARENCY May 2014 IMA publishes 2014 edition of the SORP Revisions included a number of new transparency requirements for inclusion in the Reports and Accounts of Authorised Funds, specifically: Enhancement of the 2005 disclosure requirements for portfolio transaction costs to include: o an analysis by asset class of direct transaction costs; o a figure for the average portfolio dealing spread; and o an explanation of where there are indirect transaction costs or transaction costs incurred by underlying funds. A new comparative table for each class of unit showing the financial highlights for the year, including: o the return generated over the year; o the amount distributed to investors as income; o the total operating charges for the year; o the direct transaction costs incurred in the year; and o an explanation of where there are indirect transaction costs or transaction costs incurred by underlying funds. The 2014 SORP became effective for periods commencing on or after 1 January We have recommended to the FCA that they should require compliance with the new comparative table for all periods ending after March September IMA publishes enhanced disclosure of fund charges and costs guidance Guidance recommends that firms make enhanced disclosures readily accessible in marketing material and on websites, including: Only disclosing the ongoing charges figure instead of the annual management charge; Providing explanations of the types of charges and their purpose. Disclosing figures for broker commissions and transfer taxes as percentages of the fund value. Disclosing the average dealing spread on the underlying portfolio. Providing explanations that for some types of investment the transaction costs form part of the dealing spread. Providing an explanation of the pricing policy and how effective the policy is for mitigating transaction costs arising from investor inflows or outflows. The guidance became effective from March July Revised UCITS Directive became effective Revisions included: The introduction of the KIID, which replaced the Total Expense Ratio (TER) with the Ongoing Charges Figure and removed any requirement to disclose the Portfolio Turnover Rate (PTR) because European regulators felt the PTR was likely to be too opaque for retail investors to interpret. A requirement to disclose transaction costs in the annual report. Page 34 of 60

57 June 2008 IMA and DATA publish Collective Investment Schemes (CIS) Disclosure Code, which mirrors Pension Fund Disclosure Code (see below). The objective was to promote accountability of fund managers through increased transparency, by providing the Depositary, in an assumed capacity as investor representative with the transaction costs levied on the fund s assets. September 2007 IMA publishes third edition of Pension Fund Disclosure Code (first produced May 2002, amended March 2005) The objective of the Pension Fund Disclosure Code ( the Code ) was to promote accountability of fund managers to their clients through increased transparency and to assist pension fund trustees understanding of the charges and costs levied on the pension fund assets for which they have responsibility. The Code provides pension fund trustees with information on how their investment managers make choices between trading counterparties and trading venues, more detailed information on how the resulting commission spend is built up, and what services are met out of commission spend, in particular such execution and research services as are permitted by the Financial Conduct Authority (FCA). It also provides a comparison of client specific information on costs and trading with similar fund management firm-wide information. December IMA publishes 2005 edition of the SORP Revisions included a number of new transparency requirements for inclusion in the Reports and Accounts of Authorised Funds, specifically: The total amounts of broker commissions, taxes and any other charges included within portfolio purchases and sales. The gross amount of stock lending income and any fees that have been paid to arrive at the net stock lending income amount retained by the fund. Details of any fee sharing arrangements. The basis for calculating any performance fees. The total expense ratio (TER). The 2005 SORP became effective for periods commencing on or after 1 January Page 35 of 60

58 APPENDIX 3: THE IMPORTANCE OF PRICING POLICY In Chapter One, we drew a number of distinctions: Between product charges and transaction costs. Between explicit and implicit transaction costs. Between those implicit costs which should properly form part of a quantitative disclosure on costs and those of a qualitative type. While explicit costs are easily quantifiable, we also discuss the challenges of quantifying implicit costs. In this respect, The Investment Association does not believe there is a need to reinvent how valuations and estimations of spread are reached. There is a whole discipline concerning fair value and asset pricing; these well tested methods of valuation should be used. However we explain how pricing policies already use spread estimations and how we propose that those processes could be used for enhanced disclosures. Pricing a fund With most funds, investors invest by having units issued to them and they realise their investments by surrendering those units back to the fund. To work out at what price to issue or redeem units, the fund manager starts by valuing all the assets of the fund. Some of those assets will have, on any given day, a very certain price and a clear spread at the valuation point. Equities on a stock exchange might be an example of such. Others will have a less immediately identifiable value but fair value methodologies will arrive at a price and a spread for the entire portfolio. Importantly, the depositary of a fund will have oversight of this process and this helps ensure the fund manager follows a consistent, evidence-based and coherent process. It is likely with many of the assets being valued that the fund manager will only make a detailed snapshot of spread periodically, commonly at least quarterly, or when market events mean that the fund manager no longer has confidence in existing estimations. The financial crisis was an obvious example of fast moving market conditions where estimations of spread rapidly became stale, but there can be shorter term dislocations in narrower sectors of the market, such as where the spreads in some Euro-denominated Government bonds widened with fears of a Greek exit from the Euro. Figure 9: Construction of dealing prices in funds Page 36 of 60

59 The offer/bid or the mid are used to derive internal fund prices. Importantly these are derived by approximating the transaction costs associated with purchasing or realising assets held by the fund. More formally, the prices are obtained as follows: Issue price = Offer valuation + dealing commissions + any stamp duty. The offer valuation is the same as the mid valuation + half of the securities bid/offer spread. The cancellation price = Bid valuation less dealing commissions (there being no stamp duty on selling equity assets). The bid valuation is the same as the mid valuation less half of the securities bid/offer spread. As can be seen therefore the issue and cancellation prices within the fund do take account of transaction costs. This has to be so to ensure that existing, and continuing, investors in the fund do not bear the costs associated with the purchase or realisation of assets consequent upon the investment of funds by new investors or the withdrawal of funds by departing investors. The impact of joining or leaving is known as dilution and unless the joiners and leavers make due contribution to the costs of trading then the fund (and so the existing/continuing investors) would suffer that cost. The pricing policies should operate to minimise the risk of one or other group unfairly benefiting from, or paying for, fund flows. The actual price at which investors can deal depends upon the pricing policy. For a swinging single priced fund represented by the prices in figure X, it will be either p or 98.75p or a price anywhere between these limits. Essentially if all or most of the flows are into the fund then many funds will issue and redeem at p. This can be the same for dual priced funds though the rules allow them to issue at this higher price but to lower the price for redemptions. This paper necessarily simplifies pricing policies and issues what is relevant here is that spread estimations and measurements are already calculated and used. Page 37 of 60

60 APPENDIX 4: THE VARIABILITY OF EXPLICIT TRANSACTION COSTS This Appendix is designed to provide some examples from investment funds to inform the debate about the way in which transaction costs, as measured by brokerage payments and transaction taxes, can be explained and quantified. The examples are drawn from actual data drawn from a cross-section of available funds. As we emphasise in Chapter One, there is a significant difference in nature between product charges, which operate on a relatively fixed and stable basis, and transaction costs, which may vary significantly across different time periods. The data presented here shows that: Considerable consistency can be seen in terms of actual transaction costs as a proportion of transaction value. Comparing this across similar funds could also help to establish the degree of dealing efficiency with respect to brokerage fees. However, this is not a way of expressing total costs incurred. In contrast, expressing total transaction costs incurred as a proportion of NAV (similar to the way in which ongoing charges are expressed) shows significant variation in some funds. Our view is that historic data is therefore not a reliable indicator of future costs. In fact, it could be actively misleading. An additional technical issue arises in expressing costs incurred by the fund at an individual unitholder level. Depending on the pricing policy of funds, transaction costs may be experienced very differently by different unitholders (for example, where a fund swings its unit price to protect existing unitholders from additional transaction costs created by inflows or outflows). To better explain why The Investment Association thinks as it does, the following examples are provided. They were produced by analysing actual data on an anonymised basis from ten funds using available times periods, all of at least five years. They show different transaction cost metrics with a view to seeing what could be meaningful both as an account of what has happened historically (ex-post) and as a means of setting future expectations about costs (ex-ante). In each case: The first chart shows the extent to which transaction value is the key driver of total transaction costs. The second chart shows that expressing transaction costs relative to transaction value (navy line) provides a reasonably consistent metric which is potentially useful as a general indicator of dealing efficiency. Where applicable, brokerage costs (excluding transaction taxes) are isolated in a dotted coral line. Given the asymmetric nature of taxes such as stamp duty (only payable on purchases), the latter measure is the more stable. However, expressed relative to average NAV (green line), transaction costs can be seen to be much more volatile from year to year and, although it gives a fair account of total transaction costs on a historical basis, this measure is a poor indicator of likely future transaction costs. The third chart shows the NAV-based metric on its own. The key challenge is that the NAV based metric would be the most likely to be used alongside charges figures, which are also expressed relative to average NAV. Combining an ongoing charges figure (OCF) of, say, 1.0% with total transaction costs would present an indication of costs wholly different to the experience of some or all of the other years charted. Page 38 of 60

61 Example 1: Predominantly UK equity fund; c.25% of portfolio in Europe and North America As Chart 1.1 demonstrates, transaction value (shown on the right hand scale) is the key driver of total transaction costs (left hand axis). This is what we would expect given the ad valorem nature of brokerage fees and transaction taxes. The peak in 2011 is due to a higher proportion of transactions being purchases of UK stock which carry a significant transaction tax (stamp duty at 0.5%). Chart 1.1: 75% UK equity fund - total fund transaction costs vs total transaction value As Chart 1.2 shows, transaction costs expressed in terms of transaction value are reasonably constant (navy line), with a range from highest (0.34%) to lowest (0.26%) of 8bps. In this example this variation is due mostly to transaction taxes (6bps) because the proportion of UK purchases varies. Brokerage commissions (dotted coral line) are within a range of just 2bps across the entire 6 year sample. Despite a degree of complexity in this example (a fund that combines different equity markets, each with different transaction costs), expressing transaction costs in this way is a potentially useful general indicator of transaction efficiency. In other words, a significant move either upwards or downwards is a direct expression of cost relative to activity. In funds that were trading in identical markets, a considerable discrepancy may indicate that the fund with the higher costs was trading less efficiently. The green line in Chart 1.2 shows transaction costs expressed in terms of average NAV and gives an account of the costs incurred it is broadly consistent with the line showing total value of transactions in Chart 1.1. However, it is considerably more volatile (39bp range) than the navy line (8bp range). Page 39 of 60

62 Chart 1.2: 75% UK equity fund - transaction costs as % of total transaction value vs transaction costs as % of average net asset value The navy and coral lines in Chart 1.2 are good indicators of likely future costs of trading, but not total trading costs. This green line gives a fair account of what has happened historically, but its volatility makes it a poor indicator of likely future costs, which is what policymakers and regulators are keen to capture. Combining an OCF of 1.0% in 2009 with transaction costs of 0.63%, would present an idea of cost wholly different from the experience over the next five years (see Chart 1.3). Chart 1.3: 75% UK equity fund - transaction costs as % of average net asset value Page 40 of 60

63 Example 2: UK equity tracker fund By their nature, tracker funds deal only in response to net inflows or outflows, although there will be a small amount of rebalancing necessary in response to any changes in the composition of the underlying index. As Chart 2.1 demonstrates, transaction value (shown on the right hand scale) is the key driver of total transaction costs (left hand axis). This is what we would expect given the ad valorem nature of brokerage fees and transaction taxes. The dip in 2013 is due to the asymmetry in UK transaction costs (stamp duty at 0.5% is charged on purchases but not sales). This fund experienced strong net inflows in every year except 2013 when there were significant net outflows. Therefore the proportion of transactions that were sales was much higher in Chart 2.1: UK tracker fund - total fund transaction costs vs total transaction value Chart 2.2 shows a reasonable degree of volatility (20bp range) when transaction costs are expressed in terms of transaction value (navy line). This variation is due to transaction taxes because the proportion of transactions that are purchases varies. Brokerage commissions are very low (tracker funds don t buy research and deal on execution only rates) with a range of just 2 to 3bps across the 4 years for which data is available. The green line in Chart 2.2 also shows some sensitivity to the trading pattern. This is a fair indication of the total trading costs but is unlikely to be indicative of an investor s experience. For a tracker fund there is no discretionary trading and, other than a modest level of rebalancing, trading only occurs in response to net inflows and outflows. The resultant transaction costs will give rise to tracking error unless there is a policy in place to protect the fund from trading costs using a mechanism such as swinging pricing. In a swinging pricing system, new and departing investors pay the trading costs that their activity causes. This example demonstrates that trading costs are influenced by trading patterns and may be mitigated by the pricing policy. However, although less apparent, these effects also will be at play in actively managed funds. Page 41 of 60

64 Chart 2.2: UK tracker fund - transaction costs as % of total transaction value vs transaction costs as % of average net asset value Although the range in the transaction costs in Chart 2.3 is small (6bps), it would be unfair to present these figures as costs to investors in the fund throughout the period concerned when these costs were paid solely by incoming and outgoing investors via a swinging pricing mechanism. Chart 2.3: UK tracker fund - transaction costs as % of average net asset value Page 42 of 60

65 Example 3: North American equity fund This fund has with almost all its investments in US stocks. Chart 3.1 shows a dramatic fall in transaction costs in 2011 without a corresponding fall in transaction value. The American market is unusual because commission is calculated as a percentage of the number of shares traded (in other markets it is a percentage of the value of shares traded). As a result, when stock prices rise, the apparent commission rates will appear to be lower. This would account for the dramatic drop in transaction costs after 2010 as American markets recovered from the financial crisis. Also, this effect will cause the commission payable to trade the same value of two different stocks to be different if the price per share of the two stocks is considerably different. Chart 3.1: American equity fund - total fund transaction costs vs total transaction value The effect is also apparent for both measures of transaction costs in Chart 3.2, but is most marked for the measure relative to the average NAV (green line). Transaction costs expressed in terms of average NAV are fair indicator of what has happened on a historical basis, but a poor indicator of likely future costs. Unlike other markets, the navy line in Chart 3.2 is also a fairly poor indicator of dealing efficiency which would need to be assessed as a commission rate relative to the number of shares traded to be meaningful. Chart 3.2: American equity fund - transaction costs as % of total transaction value vs transaction costs as % of average net asset value Page 43 of 60

66 As Chart 3.3 shows, transaction costs in any given year are unlikely to be representative of costs in other years. Chart 3.3: American equity fund - transaction costs as % of average net asset value Page 44 of 60

67 Example 4: South East Asian equity fund This fund has the ability to invest throughout the region, both in developed and emerging economies. As Chart 4.1 demonstrates, transaction value (shown on the right hand scale) is the key driver of total transaction costs (left hand axis). This is what we would expect given the ad valorem nature of brokerage fees and transaction taxes. The dip in 2012 is likely to be due to the mix of jurisdictions in which trading took place. Commission rates vary across the region and some jurisdictions operate transaction taxes on purchases and/or sales. Chart 4.1: SE Asian equity fund - total fund transaction costs vs total transaction value As Chart 4.2 shows, transaction costs expressed in terms of transaction value are reasonably constant (navy line), with a range from highest (0.27%) to lowest (0.19%) of 8bps. In this example this variation is due mostly to commission rates given the multi-jurisdictional nature of this fund. The green line in Chart 4.2 shows transaction costs expressed in terms of average NAV and gives an account of the costs incurred. However, it is heavily influenced in the earliest two years by massive growth in the NAV. Overall transaction costs appear to have fallen despite Chart 4.1 showing actual costs rising. However, in the latter four years it is broadly consistent with the line showing total value of transactions in Chart 4.1. The green line is considerably more volatile (56bp range) than the navy line (15bp range) in Chart 4.2. Chart 4.2: SE Asian equity fund - transaction costs as % of total transaction value vs transaction costs as % of average net asset value Page 45 of 60

68 As Chart 4.3 shows, transaction costs in any given year are unlikely to be representative of costs in other years. Chart 4.3: SE Asian equity fund - transaction costs as % of average net asset value Page 46 of 60

69 Example 5: European equity fund This example fund has the ability to invest throughout continental Europe. As Chart 5.1 demonstrates, transaction value (shown on the right hand scale) is the key driver of total transaction costs (left hand axis). This is what we would expect given the ad valorem nature of brokerage fees and transaction taxes. This fund had a change of portfolio manager and it is likely that a change of approach will be responsible for the massive fall in trading volumes between 2010 and Chart 5.1: European equity fund - total fund transaction costs vs total transaction value As Chart 5.2 shows, transaction costs expressed in terms of transaction value are very consistent (navy line), with a range from highest (0.14%) to lowest (0.12%) of just 2bps. The green line in Chart 5.2 shows transaction costs expressed in terms of average NAV and gives an account of the costs incurred it is consistent with the line showing total value of transactions in Chart 5.1. It has also been consistent since the change of portfolio manager, with a range of just 3bps. Chart 5.2: European equity fund - transaction costs as % of total transaction value vs transaction costs as % of average net asset value Page 47 of 60

70 Chart 5.3 shows that for this fund transaction costs as expressed in terms of average NAV have been consistent since the change of portfolio manager and the figures in the latter 4 years are representative of each other. Nevertheless this example shows very clearly the relationship between transaction costs and transaction values, and the significance of major changes to the fund s circumstances. Chart 5.3: European equity fund - transaction costs as % of average net asset value Page 48 of 60

71 Example 6: UK equity fund with an income objective As Chart 6.1 demonstrates, transaction value (shown on the right hand scale) is the key driver of total transaction costs (left hand axis). This is what we would expect given the ad valorem nature of brokerage fees and transaction taxes. As with other UK funds the effect of transaction cost asymmetry is apparent here due to the stamp duty on purchases. In 2010 to 2012 strong inflows shifted the trading pattern strongly towards purchases and in the subsequent years the opposite occurred with strong outflows driving a higher proportion of sales. As a result, transaction cost growth outstripped transaction value growth in 2010 to 2012, but thereafter remained relatively constant despite the continued growth in trading activity. Chart 6.1: UK equity income fund 1 - total fund transaction costs vs total transaction value As Chart 6.2 shows, transaction costs expressed in terms of transaction value (navy line) vary in a range from highest (0.44%) to lowest (0.29%) of 15bps. In this example this variation is due almost entirely to transaction taxes (13bps) because the proportion of UK purchases to sales varies. Brokerage commissions (dotted coral line) are very consistent, with a range from highest (0.14%) to lowest (0.12%) of just 2bps across the 4 years for which data is available. The green line in Chart 6.2 shows transaction costs expressed in terms of average NAV and gives an account of the costs incurred. It has an interesting relationship with the line showing total value of transactions in Chart 6.1 in 2008 there was a market increase in trading activity and hence transaction costs due to market conditions at the time. However, the NAV fell, due to both market movements and outflows so the numerator peaked just as the denominator dipped. The result is an exceptionally high figure of over 1%. Conversely, trading volumes from 2010 onwards are driven by net flows so as the numerator (transaction costs) shows considerable growth, so does the denominator (NAV). The result is that transaction costs appear level by comparison to 2008 although the range of 17bps excluding 2008 is not especially low. Page 49 of 60

72 Chart 6.2: UK equity income fund 1 - transaction costs as % of total transaction value vs transaction costs as % of average net asset value The navy and coral lines in Chart 6.2 are good indicators of likely future costs of trading, but not total trading costs. This green line gives a fair account of what has happened historically, but its volatility makes it a poor indicator of likely future costs, which is what policymakers and regulators are keen to capture. As Chart 6.3 shows, even ignoring the spike in 2008 transaction costs in any given year are unlikely to be representative of costs in other years. Chart 6.3: UK equity income fund 1 - transaction costs as % of average net asset value Page 50 of 60

73 Example 7: Global equity fund This fund has a mandate to invest in developed markets around the globe. Chart 7.1 shows a significant divergence between transaction costs and transaction value. The global nature of this fund makes it likely that a number of factors are at play here. For example, the 2011 peak could be due to significant activity being attributable to buying UK stock. Moreover the dramatic falls in 2012 and 2013 could be attributable to the commission basis being different (see American fund example). The diverse range of markets also could account for a degree of divergence. There may be other factors at work. Chart 7.1: Global equity fund - total fund transaction costs vs total transaction value Chart 7.2 shows transaction costs expressed in terms of transaction value (navy line), varying with a range from highest (0.24%) to lowest (0.11%) of 13bps. With more detailed geographical analysis of trading patterns further conclusions could be drawn about transaction efficiency in each market. The green line in Chart 7.2 shows transaction costs expressed in terms of average NAV and gives an account of the costs incurred - it is more volatile (29bp range) than the navy line (13bp range). Chart 7.2: Global equity fund - transaction costs as % of total transaction value vs transaction costs as % of average net asset value Page 51 of 60

74 Notwithstanding our lack of detailed explanation of the factors affecting the numbers in this example, it is apparent that a number of factors affect transaction costs in this fund and it can be seen that transaction costs in any given year are unlikely to be representative of costs in other years (see Chart 7.3). Chart 7.3: Global equity fund - transaction costs as % of average net asset value Page 52 of 60

75 Example 8: UK equity fund with skew towards mid and small caps This example is drawn from a UK equity fund with a portfolio skewed towards mid and small cap holdings. In particular, about 20% is invested in small cap holdings. As Chart 8.1 demonstrated, transaction value (shown on the right hand scale) is the key driver of total transaction costs (left hand axis). This is what we would expect given the ad valorem nature of brokerage fees and transaction taxes. We can also see the effect of the asymmetry of UK transaction costs due to 0.5% stamp duty being payable on purchase but not on sales. In particular, sales exceed purchases in all years except 2007 and Chart 8.1: UK equity fund - total fund transaction costs vs total transaction value Chart 8.2 shows transaction costs expressed in terms of transaction value (navy line), varying with a range from highest (0.45%) to lowest (0.33%) of 12bps with peaks in 2007 and In this example this variation is equally attributable to transaction taxes (8bps), because the proportion of UK purchases varies, and brokerage commissions (dotted coral line, 8bps), probably due to the extent to which trading activity focuses on larger or smaller ends of the market (small cap stock is more expensive to trade). The green line in Chart 8.2 shows transaction costs expressed in terms of average NAV and gives an account of the costs incurred it is broadly consistent with the line showing total value of transactions in Chart 8.1. However, it is considerably more volatile (51bp range) than the navy line (12bp range). Page 53 of 60

76 Chart 8.2: UK equity fund - transaction costs as % of total transaction value vs transaction costs as % of average net asset value As Chart 8.3 shows, transaction costs in any given year are unlikely to be representative of costs in other years. Chart 8.3: UK equity fund - transaction costs as % of average net asset value Page 54 of 60

77 Example 9: UK equity fund with an income objective. As Chart 9.1 demonstrates, transaction value (shown on the right hand scale) is the key driver of total transaction costs (left hand axis). This is what we would expect given the ad valorem nature of brokerage fees and transaction taxes. Chart 9.1: UK equity income fund 2 - total fund transaction costs vs total transaction value As Chart 9.2 shows, transaction costs expressed in terms of transaction value are consistent (navy line), with a range from highest (0.38%) to lowest (0.35%) of 3bps. Brokerage commissions (dotted coral line) are within a range of just 2bps across the entire 7 year sample. The green line in Chart 9.2 shows transaction costs expressed in terms of average NAV and gives an account of the costs incurred it steadily diverges from the total transaction costs experience in Chart 9.1 because the fund NAV is growing faster than transaction costs due to both inflows (which trigger purchasing activity) and market movement (which does not generate trading). However, it is considerably more volatile (17bp range) than the navy line (3bp range). Chart 9.2: UK equity income fund 2 - transaction costs as % of total transaction value vs transaction costs as % of average net asset value Page 55 of 60

78 As Chart 9.3 shows, transaction costs in any given year are unlikely to be representative of costs in other years. Chart 9.3: UK equity income fund 2 - transaction costs as % of average net asset value Page 56 of 60

79 Example 10: UK high yield fund This fund has an outcome objective invested 80% in bonds and remainder in equity. Chart 10.1 shows considerable deviation of transaction costs from transaction values. This is because measurable transaction costs will arise only on equity transactions. Therefore transaction costs will peak when trading activity is focused on equities. Chart 10.1: High yield fund - total fund transaction costs vs total transaction value Chart 10.2 shows transaction costs are significantly diluted by bond transactions and are volatile as a result of trading patterns. Little if anything can be deduced from these figures. To derive any useful information about future costs would require an analysis of transaction costs relative to only equity transactions. This analysis would be useful for accessing the dealing efficiency of the equity portion of the portfolio. However, the green line does give a fair account of historical transaction costs. Chart 10.2: High yield fund - transaction costs as % of total transaction value vs transaction costs as % of average net asset value Page 57 of 60

80 As Chart 10.3 shows, transaction costs in any given year are unlikely to be representative of costs in other years. Chart 10.3: High yield fund - transaction costs as % of average net asset value Page 58 of 60