Volatility spikes and impacts on liquidity

Predicting events, such as spikes in volatility, in financial markets is notoriously difficult.

At BestX we have been researching the use of Hawkes processes to model events, and our latest paper is a follow up to our publication on this subject in August 2020.

We extend the framework to model the impact on liquidity, and therefore transaction costs, with the use of an unsupervised machine learning method to categorise the number of observed liquidity states for a security. This provides the potential for an ‘early warning system’, resulting from monitoring, for example the VIX index, to indicate if sudden changes in liquidity are expected.

If you are a BestX client and would like a copy of the paper please contact us at support@bestx.co.uk.

Measuring an Algo Fill Profile – Centre of Mass

In our latest research article we present a new method for visualising the execution path of an algo and whether it executed more of the total notional at the start of execution period (front loading) or at the end of the period (back loading).

We condense this fill profile into a new metric, Centre of Mass, that can classify individual executions as either front or back loading. There isn’t necessarily any right or wrong as it clearly depends on the algo’s objective and the prevailing market conditions, but it can be useful to understand the behaviour when analysing an algo’s performance. This metric can therefore be used in conjunction with other BestX metrics (e.g. performance versus Arrival, market impact etc) to assess the performance differences across front and back loading algos.

Please email contact@bestx.co.uk if you are a BestX client and would like to receive a copy of the paper.

Momentum & Range Regimes – Adding & Conserving Alpha

In our latest research article we expand the regime framework we have previously used to predict volatility and liquidity regimes, and apply it to price direction. Using a novel machine learning method, we categorise the market into Momentum (up or down) and Range regimes, and demonstrate applications of the regime predictions to both alpha generation, and also alpha conservation. The latter is done via the comparison of execution performance in the different regimes, with Range regimes generally demonstrating lower spread costs.

Please email contact@bestx.co.uk if you are a BestX client and would like to receive a copy of the paper.

Update on the BestX Event Risk Model

Our latest research article provides an update from the BestX event risk model that we introduced in August this year. Recent moves in the VIX, and the forthcoming US election, have resulted in the model triggering. This means that the probability of further shocks has increased, and any additional spikes leading up to the election would cause the model to start predicting an impact on transaction costs. We will continue to monitor closely.

Please email contact@bestx.co.uk if you are a BestX client and would like to receive a copy of the paper.

BestX NFP Sentiment Index

In our latest research article we turn our attention to the deployment of big data in order to try to predict NFP. As an alternative approach to the traditional econometric models, we have researched the use of a simple linear model, but one that is based on the incredibly rich data set that is now available via Google Analytics. Testing indicates that the model produces improved predictions versus the market consensus. Such a prediction could obviously add value in helping trade timing when executing on NFP days, and also provide an additional sentiment indicator for more generic risk management applications.

Please email contact@bestx.co.uk if you are a BestX client and would like to receive a copy of the paper.

Balance and compromise within the Best Execution Process

There are no solutions; there are only trade-offs.
Thomas Sowell

Best execution as a concept continues to be the subject of much debate across the industry, with both liquidity providers and takers, and asset owners, having different opinions and definitions. This is no surprise given the relatively amorphous nature of the subject, it really does just depend on what you are trying to achieve, how you are trying to achieve it and why. Best execution means very different things to different types of market participants. For example, a passive index manager will typically have a best ex policy that differs considerably to that of a pure alpha driven macro hedge fund, or a corporate treasury. Clearly, it is not possible to define a ‘one size fits all’ definition for such a broad spectrum of market participants, and therefore any technology solution needs to be flexible and configurable.

As we have previously discussed here, best execution is a multi-faceted process, that is impossible to distil into a single metric or formula. Many decisions are made when conducting the execution of financial transactions, either by humans and/or by algorithms of varying sophistication, and invariably each decision has an element of ambiguity embedded in it. This ambiguity is the result of a world where there is no one ‘right answer’, but a spectrum of outcomes depending on what you are wanting and trying to achieve. Each of these results is often dependent on balancing conflicting outcomes in the decision making process, and in this article we will explore and summarise some of the key themes where compromises have to be made to achieve your definition of best execution.

Market structure and execution method

Fundamentally, the market structure within which you are executing can already provide an over-arching dilemma at the outset of the process. If you are executing in the Fixed Income cash markets, where the majority of flow is still governed by a quote driven structure, there may be few options but to trade in a principal fashion (although this is changing rapidly). However, in a market such as FX, you can choose to trade on a principal basis, or on an agency basis through the issuing of an order, or a mixture of the two. There is clearly a compromise to be resolved here. Trading on a risk transfer principal basis results in immediate certainty of execution, but does not provide opportunity to participate and benefit in the provision of liquidity. Trading as a pure agent passes the responsibility for your order to another party, thereby removing the execution certainty, but does allow you to benefit from spread earning through passive order placement. Participating in a hybrid fashion, blending principal and agent i.e. through the use of an algo, may offer the optimal compromise, although it all depends on what you are trying to achieve.

Counterparty panel size

For those markets that retain a significant proportion of quote driven transactions, i.e. Fixed Income and FX, the number of counterparties you ask for a quote is a key compromise to be rationalised. As we discussed in our research published in March 2019, there is a trade-off between achieving the ‘best’ price and tightest possible spread versus information leakage. There is a law of diminishing returns by simply asking more and more counterparties for a price as past a certain number, you will not generally achieve any better pricing but you do increase the risk of signalling to the market of your trading intentions. There is no simple answer, unfortunately. Annoyingly, the answer is ‘it depends’, and is governed by what you are actually trading, at what time of day and in what size.

Liquidity curation

A similar concept applies when deciding which execution venues to form part of your smart order router for those order driven transactions, a common conundrum in Equities, Futures and FX. Simply adding venues into the mix won’t necessarily result in the optimal execution outcome. Again, there is a trade off between accessing liquidity and achieving price discovery, versus signalling risk and information leakage. Are all of the venues that you wish to add to the liquidity pool contributing unique, additive sources of liquidity, or is there a risk of recycled or ‘phantom’ liquidity polluting the mix?

Spread vs impact vs opportunity

A classic. Sure, you can trade on the tightest price every time, but that won’t necessarily be guaranteeing you have achieved best execution depending on how you execute and what your objectives are. If you are breaking a block up into a number of clips and trading over the course of a few hours, a common practice in all asset classes, then hitting the tightest price for the first clip may result in a sub-optimal result when you look at the overall result for the block at the end of the day. Why could that possibly be, we hear large swathes of the market cry in anguish? Well, the counterparty with the tightest price may also be the one that creates the largest market impact through, for example, the lack of internalisation opportunities. This impact can result in price deterioration in the market for when you come to trade the subsequent clips, so when the overall price achieved for the block is analysed it is worse than if you had executed on slightly wider spreads with a counterparty who created less market impact. We discussed this in our first article in early 2016 using the analogy of buying a suit, an analogy which seems very dated now in these work from home pandemic times.

Executing via an order can further complicate this one by also including the probability of getting filled. If your order gets rejected, and the price moves against you whilst you are waiting for the order to be resubmitted and filled, then this represents an additional cost, often referred to as an opportunity cost. So, one venue may look like it is offering the tightest pricing, but if 50% of the time you try to hit these prices you get rejected, and also find that any subsequent price action to be moving against you, it may not necessarily be the optimum choice. Again, we explored this in our recent research that presented our ‘Total Cost’ framework, a model for bringing all of these different components together to help with quantifying and managing this 3-factor balancing act.

Speed of trading

Remaining focused on the subject of market impact, the speed that you, or for example, the algo you choose to use, execute is another trade-off. This may not be a problem if your trading objective is crystal clear and singular in nature, for example, buy 500m of USDJPY to get done as quickly as possible. In this case, you may not care at all if you create significant market impact. However, in our experience, we find that the trading objectives for the majority of buy-side transactions tend to be more nuanced and there is a desire to not create too much impact. At this point, decisions have to be made on how quickly to execute your order – too quickly and you increase the risk of impact, too slowly and you run the risk of the market moving away from you and once again creating an opportunity cost. As with constructing a counterparty panel, there is no one simple answer to how quickly you should trade. It depends.

Timing

Another potential conundrum arises with trade timing, at least for those that have discretion in their trading. There are a number of aspects to this, for example, the classic of whether to wait to net positions or not. Netting may produce overall less spread costs, but the waiting time for orders to flow through from the OMS to the EMS to allow netting creates our old friend, opportunity cost. If this potential opportunity cost is more than the spread saving, and you don’t have a strong market view, then it may not be worth waiting to net. Equally, the timing of trades to access better liquidity pools during the day, and hence lower spread costs, can be plagued with the same dilemmas. For example, a US asset manager who has a number of Asian NDF orders arrive on the desk late in the NY trading day may obtain a better outcome by waiting to execute the following morning in Asian trading hours to access better liquidity, but only if this spread saving outweighs the potential opportunity cost.

Conclusion

This article has sought to summarise some key themes within the best execution process that are all subject to balancing conflicting outcomes. Frustratingly, none of these have simple, solvable equations to provide a single result. We are not living and trading in the world of the Hitchhiker’s Guide to the Galaxy, where the answer to the Ultimate Question of Life, the Universe, and Everything is 42. So, how do you cope with these grey areas and complexity? The answer lies in the use of data and analytics, to help quantify the outcomes and thereby allow more informed decisions to be made that can be justified. At BestX we designed the product to deal with ‘shades of grey’ through the use of configurable BestX Factors and corresponding traffic lights to provide a qualitative summary of the quality of execution (although it all still looks like different shades of grey to the colour blind amongst us). Conscious of the potential complexity created by multiple execution factors, not all of which are explicitly obvious, and the sheer volume of data, we also provide automated outlier detection. Thus, once you have decided what is important to you, BestX will notify you of any transactions that breach your specified comfort zone for the value of a given execution factor.

Implementing a data-driven technology solution for best execution is only the start. Once you have this in place it allows you to conduct controlled experiments, perform scenario analysis, leverage trader experience and intuition and then analyse the outcomes rigorously with no bias. This provides the foundation for a best execution process that can be iterated, refined and improved as the knowledge bank and datasets expand and deepen.

Event Days and the Pattern of Intraday Trading

Our latest research article investigates how the pattern of intraday liquidity in FX changes depending on the nature of the day in question. We analyse key economic data release days, such as NFP, as well as month ends. Interestingly, we do not find any significant differences in the pattern of trading based on the day of the week, and we also don’t find a peak in activity around the WMR 4pm London Fix, at least when proxying liquidity by using tick counts. The economic events that seem to have the largest impact on the intraday liquidity pattern are, unsurprisingly, NFP and FOMC days.

Please email contact@bestx.co.uk if you are a BestX client and would like to receive a copy of the paper.

A framework for aggregating total cost of execution in FX

Our latest research article introduces a theoretical framework for aggregating all key components of execution cost into one, holistic measure. Analysing spread cost alone, or the impact from information leakage in isolation, only provides a piece of the puzzle. In addition, it is key, where the data is available, to also compute and include the opportunity cost from unfilled orders. Managing costs and achieving best execution is a complex and delicate balance between spread, impact and opportunity cost. Aggregating all three components provides a total cost of execution, thereby allowing more meaningful comparisons and decisions to be made.

Please email contact@bestx.co.uk if you are a BestX client and would like to receive a copy of the paper.

The Impact of Extreme Events

In our latest research article we analyse the impact of volatility shocks in the financial markets. We move away from traditional methods to analysing volatility and use a technique originally deployed to model the impact of earthquakes and subsequent aftershocks.

Exploring equity, fixed income and FX markets provides interesting insights into how long shocks typically persist in each asset class, and what the probabilities are of subsequent shocks. For example, in equity markets we find that following a volatility shock there is a significant probability of a further shock over the ensuing 8 days. In fixed income, this lasting impact period is much longer, lingering for 45 days on average, but with a lower probability of subsequent shocks. The research has a number of potential applications, for example, as risk management signals, or potentially as predictors of changing liquidity conditions.

Please email contact@bestx.co.uk if you are a BestX client and would like to receive a copy of the paper.

Analysing Liquidity in Fixed Income

Our latest research article introduces our new BestX Liquidity Score, BXLS, as a measure for analysing and comparing liquidity within the fixed income markets.

BXLS produces a value of between 0 and 100, where 100 represents high relative liquidity. Such a measure has many applications, including assisting in post-trade analysis by providing further context around cost and benchmark metrics measured for a given trade, together with potential value in stratifying exception reporting. In addition, such a score could provide value in a pre-trade, or portfolio construction, environment, where a user could select the most liquid securities to provide the desired exposure.

Please email contact@bestx.co.uk if you are a BestX client and would like to receive a copy of the paper.

Assessing the Cost of Trading a Cross

Our latest research article attempts to answer one of FX’s long standing questions – is it better to trade the cross directly, or is it better to trade through the legs?

Analysing this problem empirically we find both intuitive and interesting results. For example, on average, we find that it is generally better to trade EURSEK rather than the USD legs, as many would assume. We also find that for some crosses, it can be more optimal from a pure cost perspective at least, that trading the cross (e.g. AUDJPY) can be better than trading the USD legs.

If you are a BestX user and would like a copy of the research article please contact us at support@bestx.co.uk.

FX Algo Performance during the Covid Crisis

In our latest research article, which is an update to one of our previous publications which analysed performance of different algo styles in a variety of market regimes, we specifically focus on algo performance in Q1 this year.

This period obviously incorporates the period of extreme volatility arising from the pandemic, and we compare performance to that experienced in more normal market conditions in Q4 last year. Interestingly, we find that the Get Done algo style when measured versus Arrival Price as a performance metric generally improved in Q1 compared to Q4 for EURUSD. However, Opportunistic algos appear to have underperformed compared to Q4, when measured versus the Interval TWAP benchmark.

We also find that the performance differences across regimes have increased in Q1, reiterating the need to make informed decisions when selecting algos to achieve best execution. In times of more volatile market conditions, the cliché of selecting the right tool for the job in hand is even more accentuated.

Please email contact@bestx.co.uk if you are a BestX client and would like to receive a copy of the paper.