By now most of the financial services industry is aware of the Markets in Financial Instruments Directive (MiFID), the replacement directive to the Investment Services Directive (ISD) due to be implemented on 1st November 2007. In fact, most organizations have initiated their MiFID programme and have already decided whether they are going to simply comply with the Regulation and Directives that were transposed into UK law at the end of January, or whether they will look to develop competitive solutions to take a leading position in the market. But, unfortunately, the process isn’t as simple as trying to hit a deadline or trying to determine which pieces of MiFID could deliver a competitive advantage.

First the investment firm has to try and gain a thorough understanding of the meaning of the MiFID text and this isn’t as easy as one would think. MiFID is far from prescriptive and in the UK, where the Financial Services Authority (FSA) promotes principles based regulation, MiFID is fast becoming a minefield of interpretation and counter interpretation.

First the investment firm has to try and gain a thorough understanding of the meaning of the MiFID text and this isn’t as easy as one would think. MiFID is far from prescriptive and in the UK, where the Financial Services Authority (FSA) promotes principles based regulation, MiFID is fast becoming a minefield of interpretation and counter interpretation.

Article 21 states, “the obligation to execute orders on terms most favourable to the client”:1. Member States shall require that investment firms take all reasonable steps to obtain, when executing orders, the best possible result for their clients taking into account price, costs, speed, likelihood of execution and settlement, size, nature or any other consideration relevant to the execution of the order. Nevertheless, whenever there is a specific instruction from the client the investment firm shall execute the order following the specific instruction.

So on the one hand there is the obligation of means2 where firms have to interpret the meaning of taking all “reasonable steps to obtain……the best possible result” for each order but there is also an obligation based on result where firms specifically assess “price, costs, speed, likelihood of execution and settlement, size, nature or any other consideration relevant to the execution of the order”. From this they must then create an execution policy that the client must agree to and to sign-up to and then measure this in real-time to see if the best execution policy is being delivered. Not only does this demand a highly capable technical solution but there are new processes, legal, compliance and commercial aspects to consider.

So on the one hand there is the obligation of means2 where firms have to interpret the meaning of taking all “reasonable steps to obtain……the best possible result” for each order but there is also an obligation based on result where firms specifically assess “price, costs, speed, likelihood of execution and settlement, size, nature or any other consideration relevant to the execution of the order”. From this they must then create an execution policy that the client must agree to and to sign-up to and then measure this in real-time to see if the best execution policy is being delivered. Notonly does this demand a highly capable technical solution but there are new processes, legal, compliance and commercial aspects to consider.

Since the first Level 1 documents were delivered in November 2004 there have been questions raised about many parts of MiFID. These have not been based on simple interpretation of a phrase or minor sentence but the workability of whole fundamental concepts such as Best Execution. In some quarters it has led to a sense of gloom around MiFID with some industry commentators claiming it will be the death of the UK market and far from making Europe more competitive it will give an impetus to the migration of capital from Europe to other markets around the world.

Whilst one should refrain from the simple use of superlatives there is some seriousness in such statements and one only has to examine a post-Sarbanes-Oxley American Bond market to see how regulation can stifle growth. However, the ideas behind MiFID are laudable: increased competition has to be commended as does the idea of protecting consumers from mis-selling and other forms of market abuse. Unfortunately, something as laudable as Best Execution, supposedly a cornerstone of consumer protection, is fast becoming seen as complex, ambiguous and unworkable by many firms.

Whilst one should refrain from the simple use of superlatives there is some seriousness in such statements and one only has to examine a post-Sarbanes-Oxley American Bond market to see how regulation can stifle growth. However, the ideas behind MiFID are laudable: increased competition has to be commended as does the idea of protecting consumers from mis-selling and other forms of market abuse. Unfortunately, something as laudable as Best Execution, supposedly a cornerstone of consumer protection, is fast becoming seen as complex, ambiguous and unworkable by many firms.

One would expect that in an environment as fast moving as trading is, firms would be keen to position themselves as offering competitive solutions, drawing more clients based on multiple client specific execution policies but the industry is strangely quiet about what progress is being made. One obvious reason for this guardedness is that the banks are still refining their understanding of MiFID’s Best Execution requirements but this is also exacerbated by vendors who have also delayed producing relevant technology solutions – delayed because they too have had to take time understanding the relevant text before building something clients may or may not want. However, these solutions are now starting to gain momentum and discussions with some of the main tier 1 firms as well as a search on the web indicates that vendors are beginning to surface. But, before choosing a vendor one has to understand exactly what it is that should be delivered by November 2007.

As mentioned earlier, Best Execution is split into 3 main components: the obligation of means to deliver the best result for the client; the creation and documentation of an execution policy that incorporates venue selection; and the obligation of firms to demonstrate they are meeting their requirements under MiFID. The next few sections investigate these in more detail.

The obligation to deliver best result

The first question facing an organisation is to define exactly how Article 21 should be interpreted, on which instruments it should be imposed and the relative importance of the factors identified.

Basically, as the FSA outlines in its Consultation Paper CP0619: COB Reform (October 2006), the rule applies only “where there is a customer order, which means only where a transaction involves an agency or an agency-like relationship between the customer and the firm that is performing the transaction”. As a result MiFID’s Best Execution requirements are unlikely to apply to dealer markets but what is clear from Directive 200673EC (Level 2)3 is that it covers all instruments.

Some would say that Best Execution is intrinsically linked to transparency and whilst there is a certain requirement to provide information to prove Best Execution MiFID is very clear that a firms’ transparency obligation and Best Execution are not tied together. With this is mind, the next stage of the process is to attempt to understand and define what ‘best result’ actually means. First and foremost one has to establish the relative importance of each factor in the Best Execution equation with the variables being price, costs, speed, likelihood of execution and settlement, size, nature or any other consideration relevant to the execution of the order. What is absolutely clear in the MiFID text is that best price isn’t necessarily the most important. Each factor should be weighted according to the “characteristics of the client”, “the characteristics of the client order”, “the characteristics of the financial instruments” and “the characteristics of the execution venue” and this must be examined on an order by order basis.

In some cases price may well be most important but in many cases speed or costs may well come out on top. If, as many people believe, the equity execution market fragments post- MiFID and there is a proliferation of venues offering pan-European stocks it may well be that speed becomes the primary driver – especially in fast moving, liquid markets increasingly dominated by algorithmic trading. In less liquid markets or the credit market for instance likelihood may be the primary driver ensuring that a firm can offset its risk. Ultimately there is a basis for price but other factors become more important to ensure a member can get the best result.

In CESR’s recent Consultation Paper (07-050b) they are very clear that for Retail customers “the best possible result shall be determined in terms of: “total consideration, representing the price of the financial instruments and the costs related to execution, which shall include all expenses incurred by the client which are directly related to the execution of the order…..””5

Interestingly CESR is also clear that implied costs may not be a consideration for most retail orders as they are likely to be average or small orders in quite liquid stocks. They discuss the implicit costs such as speed, likelihood of settlement and moving the market and the way that these may not be of interest to retail customers. Any order below Standard Market Size (SMS) is unlikely to move the market significantly and as such these implicit costs are not likely to be a consideration. In the OTC derivative market price is rarely seen as the most important factor and in fact as each contract is drawn up on a case-by-case basis one could argue that Best Execution becomes irrelevant. The IMA in its response to the FSA’s DP063 on Best Execution put forward the point that: “it is simply not possible to know whether execution is ‘best’ in relation to any specific OTC derivative contract. Investment Managers use various techniques to track the quality of their fills, to ensure that execution achieved is reasonable by their assessment.”6

This statement followed some considerable concern from the industry that the Best Execution obligation was unworkable in less liquid markets and specifically because the FSA proposed that in such markets benchmarking of prices could be used to measure against. The FSA ultimately retracted its stance on benchmarking but still contends that price is important. It remains to be seen how important price will become as firms grapple with the best result but what is certain is that the main way of assessing Best Execution will be through the Execution Policy, a firm’s competitive position and the selection of appropriate execution venues.

The Execution Policy

In practice, each firm’s Best Execution provision will have to take into account the transparency that is available in each particular market and in doing so will be immediately prioritising its weightings for the criteria mentioned earlier in this whitepaper.

In certain markets, such as credit derivatives and the corporate bond market, a certain amount of opaqueness is demanded to enable clients to operate effectively. As a result, the execution policy will need to take this into account. However, in those markets where a lack of transparency has been the norm, it may suddenly become a competitive advantage to offer increased transparency. The Best Execution policy does not demand market wide transparency in price formation but where this is delivered the market may drive more liquidity in this direction. Dealers may attract a wider customer base because trading under a more transparent regime carries more certainty and is more attractive7. It also highlights the fact that alternatives are available to clients. This is likely to take place first in the Government Bond market where reference price data is much more widely available but increasingly the market is likely to see greater transparency in other areas. As a result, were intermediaries such as ICAP or Euro MTS to offer more transparent markets and greater likelihood of Best Execution, they could take market share from competitors.

In the equities market the situation is somewhat different as the MiFID pre-trade and post-trade obligations impact on the price discovery process. As firms can expect to have full reference data available in real-time they can also expect to benefit from the full extent of real-time Best Execution. Firms who offer the ability to carry out client orders will be expected to have in place complete execution policies covering all elements mentioned earlier.

These policies will have to be linked to an execution venue list and the selection of each venue will need to be carried out on the basis of the weighted criteria. And all this will need to be done in real-time. The technical challenge is significant and only a few specialised vendors are realistically going to be able to deliver smart order routing systems that can handle the rules required to route orders.

Initially, it is likely that firms will offer a single execution policy that contains each of the set criteria but as clients become more demanding, or as traders become more competitive the number of policies within a single firm is likely to grow. Clients (buy side firms) will need to have consistent policies to pass through to their own retail clients and it is unlikely to be long before they are asking the sell side to adopt their policies. An example of how the Best Execution criteria could be weighted is shown in Table 1. In this example – a retail client, placing an at market order in equities on an electronic order book – it can be seen that price is the most heavily weighted parameter in the policy and the order routing algorithm will drive towards selecting the best price from its venue list, but this may not always be the case.

Based on the total weightings, in the example above, Price is the most important parameter, followed by Speed, then Size and then Cost and Preference; Nature or any other consideration is not important in this example.

After looking at the policy in this way a firm’s execution system could be configured to select venues based on the weighted parameters.

Say client places an order to Sell 10,000 Vodafone shares At Market (FOK)

The routing engine would have to calculate the most appropriate execution venue and in this case it could be done on the basis that LSE scores 17, Chi-X 39 and Equiduct 39. The trade in this instance should be routed to Equiduct or Chi-X. There are clearly other ways that this routing could be done, especially as the type of order could begin to play a very important part – a Fill and Kill order would mean that size becomes less important and a split fill would mean that multiple venues may be selected. In the example above cost is seenas of low importance compared to other factors but if cost became a determining factor the algorithm would need to maintain a running, real-time total consideration to ensure the order could be routed accordingly. This just serves to show how complicated the routing algorithms could become and the client will have to be made aware of exactly what rules are being followed.

In addition there are a few other criteria the technical solution will need to take into account when processing orders. These include, establishing whether a static or dynamic routing process is required. This is clearly a business decision and MiFID does not attempt to give such direction but moving towards a dynamic process could be seen as a valuable competitive advantage. Few vendors are explicit about the definition of dynamic order routing but it could be seen as one where either the client could dictate what is most important for each order or in fact one where depending upon liquidity, type of stock or even other parameters, such as historic reference data, the weightings change accordingly. This could be the case in a VWAP trade where a reference price is used to determine performance and as the market changes, speed of execution becomes more important – certainly a possibility in a volatile market and perhaps where algorithmic trading is being used. This means a bank or investment firm may ultimately move towards a situation where they have a multitude of execution policies; not just one for Professional clients and one for Retail clients. In this situation, the question of how the industry would be policed must be raised.

With highly complicated algorithmic policies acting dynamically and in real-time, even with real-time verification that is outlined in MiFID the measurement of whether a client actually received Best Execution or not on an individual transaction is going to be a tremendous technical challenge. The system will have to take a snap-shot of the all information available at the time of execution including the price, size, cost, etc of every available venue that the firm links to (or is included in the firm’s execution policy). If historic data was taken into account in say a fast moving falling market then scrutinising the delivery of best result will be a very interesting challenge.

Another dimension that the execution policy will need to take into account is the venues – selected clearing and settlement options. Some venues will have decided that a single clearing and settlement regime is appropriate whereas others, such as Virt-x give their members an option from several.

Obviously, these “options from several” concepts are not dynamic in that they can be made on an order by order basis but the decision taken by the member will have to be included in the execution policy. Some considerations for firms’ execution policies are:

• What happens with non-domestic settlement?
• Can the venue cope with cross border execution and the probable increase in cost
  this may bring?
• What happens if a firm places a buy order for a particular stock on one venue that clears and settles through one CCP and CSD and this is immediately followed by a sell order for the same stock through a different venue that clears and settles through a different CCP and CSD?

Unwinding back to back positions in this way will create potentially expensive and time consuming complications for the firm. Therefore, the more custodian connections that a firms CSD has the better. Alternatively the firm will have to provide upfront funding for the settlement. This potential for additional cost should filter back through the chain of execution implying that the more options a venue has the more likely it is to be placed on the Execution Venue List of firms.

Ultimately, questions surrounding the best result for the client will be based on whether the client agreed to the best execution policy, did they understand it and were they aware of the risks – but these are not obvious components of best result. The problem here is simply compounded by the potential for a lack of understanding in an industry that thrives on a competitive edge. Because the Best Execution obligation is almost schizophrenic in its nature, being an obligation of means as well as an obligation of result, the regulators will be permanently policing a multitude of non standardised and unstructured policies. As D’Hondt and Giraud say in their white paper on the subject, there were two possible routes in drafting the provision:

• “impose an obligation of means and leave the industry to develop standards to which clients could refer when confronting an investment firm;
• Impose an obligation of result based on a clear and measurable objective.

With such an unbalanced provision, we believe that there is absolutely no chance that either the industry or the investor will develop a sense of confidence that their regulation is fair and protective of their rights.”8

This situation is again made more complicated for the Regulator to police as mentioned earlier, because many subjective factors could also be required to be included as part of the policy. In fixed income trading anonymity is seen as critical – how could this be measured in best result? Across the industry other factors are often taken into account, for instance deadlines, price improvement and risk – again where are these placed in a best result algorithm?

The industry is likely to see a situation where the obligation of result is usurped by the obligation of means and although this accords with the FSA’s principles based approach it makes it tremendously difficult to measure. In an industry where knowledge is capital those informed investors – the Professionals – are likely to gain significantly or at least be more aware of their risk.

The concern is that retail investors are not going to have the luxury of either time or information to assess how close to a best result they ever actually got to. If they are not able to assess this they are unlikely to complain9 and with thousands of execution policies operating it is highly unlikely that the regulator will ever be in a position to effectively police the market; the risk of unscrupulous firms taking advantage of this situation has to be very real.

The Execution Venue List

Once the execution policy has been determined it will need to be populated by relevant venues creating an execution venue list from which the policy can select the destination for the order. MiFID is quite clear about the obligation on firms in the selection of venues and in addition it also makes it very clear about the need for firms to review their execution policies on a regular basis.

Article 21 (3) says: “The order execution policy shall include, in respect of each class of instruments, information on the different venues where the investment firm executes its client orders and the factors affecting the choice of execution venue. It shall at least include those venues that enable the investment firm to obtain on a consistent basis the best possible result for the execution of client orders.”10

It follows that a full and complete execution policy will have to include a reasonable list of execution venues, including the rationale for the selection of such venues. The Execution Venue List will become an important piece of collateral in the banks’ ability to demonstrate Best Execution. However, market data vendors, such as Reuters, Bloomberg and Thomson, as well as the venues themselves will play an important part in disseminating this information to allow firms to monitor relevant execution venues.

What is not clear is what happens to those venues who are not included in a venue execution list but who through data vendors, or directly by the venues, demonstrate they are faster, can deliver a lower cost on an order by order basis and in general purport to deliver a better result for clients. MiFID would seem to indicate that firms are obligated, owing to results, to include such firms on any execution venue list but the obligation of means may well preclude this – adding further confusion to the provision. For instance, if the contract between the venue and the bank appears to add risk into execution, or perhaps if there is a question over financial stability of the venue or even some concern over the longevity of consistent performance should the venue be excluded? Managing such subjective parameters is extremely difficult but MiFID is clear in its drive to deliver consumer protection so one would think that longevity of operation is an important factor. However, it is also clear about its drive to deliver greater competition in venues.

So far during discussions across the industry, it has become clear that there are a number of varied viewpoints on how venues will be selected for execution venue lists. What is clear is that there are a large number to choose from across the EEA including the obvious ones such as the London Stock Exchange, Euronext and OMX but what about newer venues such as Equiduct, Chi-x and Turquoise? If Article 21(3) is taken at face value then firms should investigate all these in order to ensure that on a consistent basis they are able to get the best possible result for clients. Clearly a complete and regular review of every venue from Tallinn to London via Malta covering all execution parameters is unworkable but the question is what is reasonable?

One way to get around this would be to perhaps include a venue that guarantees to deliver Best Execution but this raises more questions. First, how can this be guaranteed given all that has been said so far? If such venues demonstrate they deliver the best net result based on transparent and measured characteristics then surely they should be included.

However, if other characteristics of the venue could be interpreted as a potential risk, such as experience in a market, then they, in commercial terms, are likely to be precluded. Secondly, one has to ask how a venue that delivers Best Execution could be included by a firm that has built a smart order routing system themselves. The two policies could well conflict and the policy configuration of the venue may not be as flexible as the firms’.

Moreover, as mentioned earlier firms are likely to drive towards more and more dynamic execution policies and this leads to the potential for inaccuracies between the Best Execution delivered by a venue and that delivered by a smart order routing system.

Again, it then depends on how configurable the execution algorithm is of those venues who deliver Best Execution and whether characteristics of the client, nature of the order etc can be included in the order. If they are extensible then there is a very strong case for ensuring firms include such venues on their lists but it demonstrates the importance of a well defined execution policy; one that covers more than just price, costs and latency etc

A venue selection checklist

This white paper has highlighted some of the wide range of criteria that may be used in the development of an execution policy and that through competitive force those execution policies could become quite complicated. Nevertheless there are some practical guidelines that may be useful in identifying a solid policy:

• Static or dynamic: A static policy, one that simply uses a standard weighting regime may be appropriate but a more dynamic policy that uses historic data may give a competitive advantage. So, the question is does the venue operate in a way that would support such trading?
• What instruments are being traded on the venue? Equities only, Derivatives, Commodities, ETFs, ETCs?
• How important is latency in venue selection? If it is very important then obviously the venues with lowest latency should be included in the venue selection checklist
• What types of orders are expected? Fill or kill, fill and kill, partial fill, at Market, Limit, Iceberg etc. The more types of orders a venue can support will mean the more likely it should be placed on the venue list.
• What type of order book is important? It is unlikely that with aggressive real-time execution policies anything other than electronic continuous order books will suffice. However, the order book algorithm is also worth investigating to see how orders are placed and what process is used for fills.
• Does the venue have a clear set of rules? The clarity, conciseness and consistency
  of venue rules and directives should be taken into account in venue selection. They should support the trading regime of the policy and be as transparent as possible.
• Is fill size important? Size of fill and ultimately the type of orders expected by the venue strategy should be investigated. It would be inappropriate to select a venue that prides itself on algorithmic trading, lending itself to small lot sizes when the policy may demand significant order sizes. Most venues outline what minimum fill size is acceptable on their system and this should be examined.
• How important is cost of trading? The cost of trading should be investigated as it will differ across all venues. The cost should be in basis points (bps) and cover clearing, settlement and any other potential venue fees. These should be transparent to the member of the venue. Flat fee structures for all trades are possible on some venues although tiered volume discounts could be more appropriate (assuming these are not determined as inducements to trade by CESR – CESR’s consultation paper 07-050b makes it clear that fee structures should not discriminate against other venues 11)
• Is pan-European settlement important? If the policy demands execution of orders across country boundaries the ability of the venue to deliver cost effective pan-European settlement may be very important indeed.

This list is not exhaustive but gives an indication of some of the aspects the execution policy and the venue selection process should cover.

Conclusion

The MiFID Best Execution provision has created quite a storm across the industry over the past year or so and it would seem that this is likely to intensify rather than simply go away. Firms are in the business of making money and this will drive them to test the extremities of the regime. In the UK, they are likely to use the FSA’s drive towards principles based regulation as a standard and therefore utilise the obligation of means in Article 21 as much as possible.

There remains a question about how the Best Execution regime will be interpreted elsewhere in Europe and with non-FSA regulated firms. In its recent consultation paper12 CESR makes it clear that there are “Implementation Issues” but that it hopes for harmonisation by urging competent authorities to follow the intelligent copy out concept: “Discussions in the CESR Implementation Forum suggest that many Member States and competent authorities will implement MiFID’s best execution requirements by introducing the terms of the Directives directly into their legislation or rulebooks.” Nevertheless the CP goes onto say that: “……important questions remain such as how competent authorities will interpret and supervise compliance with these rules in practice.”

What is also likely is that there will be a huge difference between the implementation of Best Execution for professionals and that delivered for retail customers. We will have to wait and see whether this ultimately delivers protection for retail customers or whether the complexity in policing the best result leaves many consumers wondering if MiFID really has been a benefit. On the other hand MiFID promotes increased competition and with many new venues positioning themselves to offer MiFID related services one has to hope that firms take advantage of this situation bringing about reduced cost and less risk in execution.

What is also likely is that there will be a huge difference between the implementation of Best Execution for professionals and that delivered for retail customers. We will have to wait and see whether this ultimately delivers protection for retail customers or whether the complexity in policing the best result leaves many consumers wondering if MiFID really has been a benefit. On the other hand MiFID promotes increased competition and with many new venues positioning themselves to offer MiFID related services one has to hope that firms take advantage of this situation bringing about reduced cost and less risk in execution.

1. Directive 200439EC
2. MiFID the (in)famous European Directive – Catherine D’Hondt and Jean-Rene Giraud
3. Recital (70) of Directive 200673EC (i.e. MiFID Level 2)
4. Article 44 (1) paragraphs (a), (b), (c) and (d) of Directive 200673EC (i.e. MiFID Level 2)
5. CESR/ 07-050b – Best Execution under MiFID (February 2007) – paragraph 26
6. IMA response to FSA DP063 dated, 14 August 2006, Page 4
7. However, there is a contrary view proposed in the FSA’s CP0614 and an earlier Equiduct whitepaper – “Is this
   the CESR Liquid List”, page 14. Essentially in markets where there is a drive for increased transparency there is
   the possibility that Market Makers suffer from reduction in their informational advantage. As a result, they may
   increase spreads and this may actually reduce liquidity.
8. MiFID the (in)famous European Directive -Catherine D’Hondt and Jean-Rene Giraud
9. There is a contrary view that one could take in that where the execution policy is not clear or transparent there
   is more likelihood of them complaining to the FSA.
10. Directive 200439EC
11. CESR/ 07-050b – Best Execution under MiFID (February 2007) – paragraph 37
12. CESR/ 07-050b – Best Execution under MiFID (February 2007), para 8 and 9