October 05, 2015
While Best Execution and Transaction Cost Analysis (TCA) are well-established in equity trading, other asset classes have been slower to adopt such techniques due to limitations in market data and market structure characteristics. In Over-thecounter (OTC) markets there has typically been no requirement for central reporting, making it difficult to demonstrate best execution in the same way as for equities. This is beginning to change due to pressure from regulators and end investors who require higher standards of information. Market structure changes, with more electronic platforms taking increasing shares of trading, are also enabling more precise analysis. Over the last three or four years, Foreign Exchange (FX) TCA has become increasingly mainstream for asset managers, while one recent survey shows that in the past year, Fixed Income TCA has become the fastest growing category of analysis1. These trends are expected to continue, not least in the light of MiFID II regulations.
It is a truism that an asset manager should be expected to execute orders on terms most favourable to the client as opposed to the benefit of themselves or any third party. It is also reasonable that they should be expected to demonstrate this on a systematic basis across all types of trading that they undertake, whatever the asset class. In the real world this has not been as easy as it sounds. Increasingly, regulation is raising the bar of expectations and imposing obligations which the industry is struggling to implement, particularly in markets which have historically traded over-the-counter.
Under article 27 of the EU Directive 2014/65/EU,
The asset manager is obliged to take these key factors into account as part of their measurement and monitoring processes to be able to show that their trading has indeed met the regulatory requirements. Random sampling of a few trades per quarter is unlikely to be considered adequate in the way it would have been in the past. In this new regime firms will have to introduce processes to record details of every trade and its characteristics, together with the market conditions, and any relevant instructions received from the client.
The largest asset managers have in some cases developed in-house processes to undertake this kind of monitoring, but for many firms the cost of such developments is prohibitive. Hence, third party suppliers have increasingly been called upon to assist, not just in the well-developed field of TCA for equities, but in other asset classes as well.
It is a major challenge even for such vendors to offer the same granularity and accuracy of analysis in all asset classes. Although much can be learned from the equity experience2, it is simply not possible to apply exactly the same techniques that have been developed in equity markets to FX, Fixed Income or Derivatives, where market structures are less transparent and electronic platforms only penetrate certain parts of the market.
While the details do vary, the broad approach is consistent from regulators, and the expectation that the asset manager should assess the different execution factors still applies, based on a stated execution policy, taking into account the context of the trade. The asset manager should aim to minimise the implicit and explicit cost of trading for the end investor, other things being equal, while avoiding any conflicts of interest or hidden charges.
Asset managers have typically been required for many years to have a stated execution policy. The new regulations go further however, stating:
Crucially, the policy should deal with all asset classes that fall within the MiFID II definition of a financial instrument, and should be more than simply a routine reporting process. It should be the subject of prior discussion and understanding on the part of the investor. It should include an explanation of trading both on execution venues (stock exchanges and MTFs) and outside a trading venue (typically over-thecounter). It is the asset manager’s responsibility to monitor it on an ongoing basis, and report on it periodically to demonstrate its effectiveness.
Typically, such monitoring entails the identification of outliers or anomalies in trading, with an expectation that the manager will investigate such outliers in more detail, recording the circumstances which led to the good or bad performance. A key point about best execution is not that every outcome is within a narrowly predictable band, but that those occasions which are identified as outliers can be explained based on either the execution strategy employed or the market conditions in which they occurred. It is at the overall level of performance that best execution is ultimately evidenced, with individual trades (and even more so individual fills) being subordinate to the process.
Hence it is the policy that is key, and the ability to demonstrate compliance with it to clients. This is expected to ensure a closer and more open level of communication between asset manager and client than in the past, bringing enhanced transparency throughout the investment process.
The regulation is more specific about the reporting of execution venues than in the past:
This particular statement has been a cause for concern on the part of some asset managers who initially believed it to mean that they needed to install expensive systems to record every single fill to determine where their brokers were routing their orders, whether to primary stock exchange or MTF, whether in the lit or dark markets. The interpretation has evolved however, and it is now widely considered that the buy side need only to monitor the brokers that they use. The brokers in turn are expected to monitor the underlying execution venues on which the trades occur, and make this information available to the buy side on request.
Nevertheless, the more progressive (and financially able) buy side firms are already recording the execution venue data as well as the brokers. This enables them to understand in far greater detail the ways in which their broker is routing the order, whether to their own liquidity pools or elsewhere, and the relative performance due to such routing. It can also potentially help identify the value of any rebate or other third party payment or non-monetary benefit from an execution venue to an investment firm.
While this type of granular analysis is likely to become the norm it is something to be used with care. Individual fills are rather like looking at each of the dots of a colour TV: individually, they can tell a picture of sorts, but the real story is only understandable by standing back and looking at the overall effect. The fact that an order was traded in one thousand pieces is all very well, but the timing, the sizes and the locations of the fills when taken in aggregate are what determines the overall result. Best execution is not about where and at what price 100 shares traded out of an order for 100,000 shares. In addition, research has shown that factors such as the strategy used to access the venue are critical to the overall evaluation of the execution venue3.
The proposed regulations require execution venues to publish significantly more detailed information on trades that take place on their venue. This includes expectations of the required fields, format and frequency, plus additional requirements for quote/order driven execution venues. This requirement for more data is part of a move towards greater pre- and post-trade transparency that continues to be a core theme in the regulatory direction, both explicitly spelt out in the regulations and as interpreted by market practitioners.
One phrase which is open to interpretation in the draft regulations is the definition of the “quality of executions obtained”. This is intrinsically a qualitative benchmark which requires interpretation by the user. One proposal which is emerging is the possibility of an industry-led definition of best execution which could include some outline of the qualitative features that should form the basis of such a definition. The FIX Trading Community published a document of Best Practice in TCA in 2014, and there is scope for the document to be extended to incorporate key definitions regarding Best Execution as well. Within such a framework the onus would still remain with the asset manager to define its own policy based on their investment process, asset mix and circumstances.
So perhaps a useful approach to defining the road map to meet MiFID II requirements for other asset classes is to look at the current state of preparedness of a typical (or “median”) asset manager versus those at the leading edge in the field of equity trading, where typically such analysis is most advanced. Clearly not all firms have the resources to be at the leading edge in terms of IT infrastructure, personnel or external service provision, but those firms who are able to pioneer the use of such approaches perhaps indicate a direction which others can reasonably be expected to follow over time. And as such trends become mainstream, the standards tend to become expectations from end clients and regulators alike. Best practice can become the minimum expected standard (or a “hygiene factor”) remarkably quickly.
From the table it is clear to see that typical managers have the ability to monitor the costs at parent order level, with aggregate metrics available to assess broad factors in reports. Outliers are monitored at least quarterly, including assessment of broker performance. Leading edge firms go beyond simple reporting, using data on a daily basis to monitor and document outliers, drilling down to execution venue at fill level to understand how brokers are routing the order, and using the various outputs to fine tune the investment process.
While the concept of Best Execution has developed over the last two or three decades in the world of equity trading, in other asset classes it is arguably only in more recent years that it has come into sharp focus. The structure of OTC markets, without centralised recording and publication of trades and little or no reporting of volumes, has made it far more challenging to develop the kinds of sophisticated monitoring and analytical tools taken for granted in equities. As noted by Aite in their 2014 survey of the FX TCA market, “one immediate challenge is to first define what ‘best execution’ in foreign exchange is all about, and then to reach a consensus agreement on that definition among all market participants on the buy-side and the sell-side alike.4”
Spot FX trading is not within the scope of MiFID regulations, either currently or in the proposals for MiFID II, but the various scandals of rigged markets and law suits against custodian banks have led many participants in FX trading to look for ways to systematically monitor the achievement of best execution – whether as a client or as a provider of the service. Thus, for the purposes of this qualitative assessment, spot FX is included.
Electronic platforms are increasingly used in FX trading, and can provide good data for analysis, but they are not a panacea as far as providing a panoptical view of the market place. Competing quotes provide some limited context against which to assess a trade, but they do not on their own go far enough to ensure that the selected subset of counter-parties are a true reflection of the wider market. The best of three quotes may simply be the third worst quote of dozens available. And if an asset manager maintains an unchanging list of counter-parties over time they will have no way of gauging whether their selection remains optimal over time.
It is not uncommon for firms to have such competitive quote comparisons recorded for spot trades, although fewer currently have a systematic process in place for assessing forwards or swaps. Nor do firms typically tend to link the timing and cost of the FX transaction to the related transaction – for instance the buying or selling of a stock or a bond. This can hide implicit costs or inefficiencies in the investment process which over time can lead to a significantly weaker investment return than would otherwise be the case. Leading edge firms have started to assess forwards and swaps, but true linking of FX to a related event is in its infancy.
While those firms that have monitored FX trading have tended to do so on a monthly or quarterly cycle it is highly likely that this will evolve into a daily process. It will also entail far more granular analysis based on data inputs from a variety of sources, not just representing those that are currently employed. The different channels will be assessed to determine whether ECNs or banks or algorithmic strategies are most appropriate for a given set of trade characteristics, bringing together post-trade and pre-trade analytics. Precise timestamps and a record of the investment objective will allow more accurate benchmarking and fine tuning of the process.
Where an external full service vendor is used, they should be able to incorporate price and volume information from a variety of execution venues and price feeds, including leading ECNs, bank platforms, interbank platforms and so on. The vendor should also be able to analyse not just spot trading but also forwards, swaps and NDFs, as well as being capable of granular intra-day analytics, including algorithmic analysis. Benchmarks should be size-adjusted (as opposed to simply referencing top of book) and be capable of measuring performance against multiple different timestamps to reflect the different stages in the investment process. Leading firms already expect to use data to evaluate the benefit of active trading versus netting, and such analysis is expected to become the norm across the majority of asset managers.
Similarly, there is growing interest in market movements in OTC markets immediately after a trade (sometimes referred to as footprint analysis). And given the scandals of recent years the ability to assess the benefit or cost of trading on the 4.00pm London fix is expected as a standard metric in any best execution or TCA process.
Currently, listed derivatives are largely treated in a similar way to equities for TCA purposes, although under the current reporting regime not all trades are reported, meaning that certain metrics such as Volume-Weighted Average Price (VWAP) are unreliable. In the future, it is expected that reporting requirements will be more rigorous, allowing more accurate analysis. It is also likely that asset managers will start evaluating the performance as an offset to other asset transactions and the related delay costs which are currently being incurred. An example might be the switching of an index-based instrument to the underlying constituents: it may be that a cost on one side is offset by a gain on the other, or if handled badly, costs may be incurred on both sides of the trade.
OTC derivatives pose a greater challenge, as with all OTC markets. Currently they are at best measured manually using laborious recording which is not always reliable or insightful, for instance comparing competing quotes. Leading edge firms are starting to look at RFQ (Request For Quote) automated software to provide audit trail and analytical records to provide more robust surveillance and the ability to adjust processes to iron out inefficiencies. A range of metrics are already available for post-trade analysis, including hit ratios, product statistics, size statistics, and links to underlying assets within a basket. Notional hit ratios can be measured, as can response ratios, bid-offer spreads, alpha versus quote average, alpha versus listed price and broker assessment measures.
FIXED INCOME TCA
While FX TCA is in its early stages Fixed Income TCA is in its infancy. The challenges of an OTC market without a central record of prices or volumes, and in which large numbers of securities do not trade at all for weeks or months, are considerable. In fixed income, as in FX, a widely used approach to best execution is a manual comparison against competing quotes from a limited number of counter-parties. Alternatively the use of indicative valuation data is used, based on the prices used for portfolio valuations. The latter approach is indicative but not necessarily a tradable price. Despite the challenges there is a pressing need for good analytics in fixed income trading. A recent report by Greenwich Associates claims the use of TCA in fixed income markets is growing faster than in any other asset class, with over one third of asset managers interviewed using TCA in the sector, up from 19% two years earlier. Of those who do anything at all, 54% use internally developed tools, however. This is likely to move towards third party solutions over time5.
More advanced firms have developed automated comparisons of market quotes to their executions, or use the TCA reports provided by electronic platforms. Best execution can also be assessed for fairness compared with similar products, although such approaches are still relatively simplistic and lack granularity. For rarely traded securities the best available reference price is likely to be based on securities with a similar mix of characteristics (credit rating, coupon, maturity, liquidity and so on) and which have traded in similar size. At best this is likely to be patchy, a sort of shadow play which is as good as the bucketing of those characteristics. The more homogeneous the bucket, the more accurate the price indication is likely is to be.
A number of initiatives are underway which will help develop more sophisticated Fixed Income analytics and best execution reporting, starting with market structure. If more fixed income trading moves to centralized electronic venues and away from the OTC market makers this will provide more transparent sources of post-trade data on trades (although the effect on liquidity remains to be seen). Similarly, a move to more standardized instruments may to some extent mitigate the issue of bucketing similar securities, at least for commonly traded instruments.
The FIX Trading Community initiative Project Neptune is also expected to provide more granular data which could be used for best execution and TCA purposes.
It is likely that most firms will implement manual or partially automated processes for comparing rates to dealer quotes, using daily or potentially intraday data. More advanced firms are likely to want to compare performance against a range of price sources – actual market quotes and trades, peer data and intraday evaluative pricing. Cost measurement can be expected to develop similar to the implementation shortfall types of metric used widely in equities, along with difficulty-adjusted, liquidity-based assessment of counter-parties.
Challenges clearly remain in best execution for fixed income trading. In the near term the determination of a fair price will continue to entail a degree of inaccuracy. Over time it is possible that developments in market structure and data availability will finally allow the same level of analysis that is widely employed in equities.
This article was originally published in Best Execution