Automated trading in the repo markets current and future trends

Transcription

1 CAPITAL MARKETS Jerome Cardon, Repo Product Manager, Broadridge Martin Seagroatt, Global Marketing Director, Broadridge August, 2018 Automated trading in the repo markets current and future trends

2 SUMMARY Automated trading and execution management is becoming more prevalent in the repo space. The use of automated trading and smart order routing technology has long been common in the equities and FX markets. In contrast, as a largely relationship, voice traded business, repo has lagged behind this trend. However, these techniques are now becoming more widely used by forward thinking firms looking to compete effectively in repo and collateral trading. A number of trends are driving this increasing uptake of algorithmic trading techniques. This paper covers the market and regulatory dynamics that are leading to this interest in algorithmic trading. It also discusses a range of techniques that automated trading solutions can employ to improve trading, including emerging technology such as artificial intelligence. The techniques covered include: Market analysis transaction volumes, market trends etc Trading across multiple markets Passive aggressive Price divergence from GC Trading off inventory Trading Algorithm Pattern Recognition and Randomisers Sniping tools The information comes from interactions between the Broadridge Securities Finance and Collateral Management team and market practitioners in the roll-out of the Broadridge Repo Order Quote solution a recently launched multi-market aggregation and execution tool for repo markets. TRENDS FRAGMENTATION OF LIQUIDITY AND GROWING COMPLEXITY Electronic repo markets are becoming ever more fragmented, with increasing scope and growing overlap across good to-trade brokerage platforms such as LSE s MTS, NEX s BrokerTec, and Deutsche Boerse Eurex. This results in added complexity for the repo trader. With firms now forced to guarantee best execution under MIFID II and more volume starting to appear on emerging peer-to- peer/all-to-all networks and CCPs, this complexity will only increase. With a need to sweep more markets to achieve best execution and prices changing rapidly, it is difficult for a human to keep up. Furthermore, ICMA noted in its latest European repo market report that there is a trend toward smaller trade sizes and more rapid turnover of dealer positions. Liquidity squeezes around year end reporting dates are also compounding the intricacy of the Gordian knot the repo trader must now untangle. NEED FOR EFFICIENCY GAINS Post-crisis, repo desks are now forced to do more with less. A shrinking resource base has occurred in parallel with an increase in the scope of the repo function, with repo traders now playing a more significant role in: Managing balance sheet, liquidity and capital more effectively The sources and uses of collateral across financing and margin based activities Firmwide collateral optimisation Best Execution & Order Slicing 2 BROADRIDGE

3 INCREASING USE OF AUTOMATED TRADING TOOLS To meet these challenges, repo trading is beginning to employ automated trading tools more traditionally used in electronic FX/equities trading. This allows traders to cut through complexity and take advantage of short lived trading opportunities in constantly moving markets. Coupled with closer integration with global inventory management and a move to a more aggregated view of markets in real time, it enables firms to maximise P&L from available inventory and helps source scarce market liquidity. Automation of the day-to-day General Collateral financing requirements frees up valuable time for larger, more complex trades and supports more strategic decisions about sourcing liquidity and optimising resources. We will now discuss a range of automated trading strategies traders can employ to improve effectiveness. AUTOMATED TRADING STRATEGIES BUILDING BLOCKS FOR AUTOMATION As a basic building block in employing automated trading strategies and improved execution management, it is important to first have in place a technology system such as Broadridge s Repo Order Quote Solution that can aggregate electronic and voice brokered quotes in real time to gain a single market view. MARKET ANALYSIS Basic techniques include analysing transaction volumes for a given security to gain a daily profile of trading for that specific security. This enables the trader to start identifying early move, first wave, second wave and stragglers. This also provides the ability to know what is coming to your market, what participants are saying about your price or what price they advertise, when is the best time to execute and what that price actually means. Collecting, handling and having the right data available is critical, but crucially, depends on your specific business, meaning that you need a complete but flexible platform. Comparing volumes today vs previous days can give an early indication of whether something is happening in the market. Likewise, looking at trading corridors, i.e. the difference between best bid and best offer, and how they evolve over the day can provide valuable insights that feed into trading strategy. AUTOMATED TRADING IN THE REPO MARKETS CURRENT AND FUTURE TRENDS 3

4 These techniques can start to give the trader a much better understanding of market activity, and successfully replace trying to piece together data from disparate sources such as trading terminals, repo rates, clients and counterparties. As with the game of poker, knowing what is happening sooner can make all the difference. MANAGING MULTIPLE MARKETS Simple execution management can be as basic as executing in a way that avoids multiple hits when trading across multiple markets. This can also extend to managing an integrated quote across the markets, rebalancing un-executed quantity on perceived available liquidity. PASSIVE AGGRESSIVE Another technique is the Passive Aggressive approach across multiple markets. Say you want to buy 100m and put a quote out to market, but the liquidity is not there. If you find a way to source an amount close to the volume you want to buy, you can then execute this almost size. The execution system then reduces the quote amount in the market automatically without trader intervention. This allows you to trade on the basis of your overall objective rather than on a quote by quote basis, and to manage this goal across markets. PRICE DIVERGENCE FROM GC Gaining an immediate notification of what s trading away from GC gives a real time view of when a component of the basket is diverging from the basket price. Gaining this understanding more explicitly across markets can provide various opportunities depending on the trading objective. TRADING OFF INVENTORY Integration between the trading system and the global inventory manager can provide major benefits in defining the trading objective in relation to a position, where the position can be updated by another party, for example a fund manager, or a cash desk. For example, the position shows a long position of 100m, and the desk objective is to finance that position. As the day progresses, the position changes to 80m, and the trading objective is automatically updated to reflect this. This link to inventory can also be enhanced with off-system (behavioural) information: for example, the desk knows that the client will roll-over a position, but the roll-over date is in the future. TRADING ALGORITHM PATTERN RECOGNITION AND RANDOMISERS Counterparty trading activity, including automated trading can sometimes create a trail that makes it possible to identify the trading strategy. Solutions that can use pattern recognition (something that machine learning is particularly good at) to spot counterparty strategies can provide value to traders. Conversely, randomisers built into one s own trading algorithms can cloak one s strategy, meaning counterparts are unable to spot any discernible logic to the firm s trading activity and cannot therefore start to trade against you. SNIPING TOOLS Anyone who has bid for anything on ebay will know the frustration of sitting watching an item about to close. Your bid is winning! But at the last second, another bid suddenly exceeds yours. How is this possible?! You ve been sniped. Automated sniping tools, widely available on the internet, can auto-exceed the highest bid within a defined limit, allowing the user to avoid having to sit at their PC waiting for a bid to close. These tools are now coming to the repo market, and mean that correctly timing trading strategies becomes ever more important. Not using these tools can lose you trades but can also put the firm s liquidity management under duress during times of market stress and year end liquidity crunches as the firm repeatedly gets pipped to the post when trying to source liquidity. 4 BROADRIDGE

5 AUTOMATED TRADING IN THE REPO MARKETS CURRENT AND FUTURE TRENDS 5

6 BEST EXECUTION & ORDER SLICING Best Execution can be defined using different dimensions, for example, price, liquidity, cost, speed, execution likelihood etc. An automated execution tool could therefore optimise for whichever of these parameters are most important or some combination of them. Likewise breaking orders into smaller chunks that will avoid moving the market and then timing those orders in a way that ensures optimum execution can also provide benefits. Market impact models, increasingly employing artificial intelligence can evaluate the effect of previous trades on a trade and how the impact from each trade decays over time. This allows traders to avoid executing certain trades too closely together, leading to market impact effects that reduce P&L. ARGUMENTS AGAINST AUTOMATED TRADING The repo world is something of an anomaly vs other electronically traded asset classes. Daily trading volumes are not as large and profit margins on each trade are not as wafer thin as in the FX and equity markets. The repo markets will require substantive changes to clearing and settlement infrastructure, something that blockchain promised but has yet to deliver, to become a high frequency business. Furthermore, when applying tools such as artificial intelligence, algorithms perform better when they have larger datasets to analyse. It may be that volumes in the repo markets are not sufficiently large to enable successful training of AI solutions. However, should direct buy side participation become more prevalent, then volumes would perhaps increase to the point where trading solutions evolve to make use of new techniques and become even more essential for successful trading. While avoiding using a sledgehammer to crack a nut, the deployment of algorithmic execution tools can provide significant benefits to traders, allowing them to focus on defining strategies to deal with the increasing range of functions the repo desk now needs to perform. WHERE DOES ARTIFICIAL INTELLIGENCE FIT IN? Machine learning has a role to play in execution management as techniques evolve and investments in AI driven technology increase. For example, JP Morgan recently announced a machine learning solution that optimises the execution of large block trades based on prevailing market conditions for its equities business. There is no reason why a similar system could not be employed in due course in the repo markets. Reinforcement learning, which uses a large number of simulations to train an AI system on optimum strategies to achieve a given outcome in trading scenarios is another area that could help to boost traders performance. Following its initial training, the machine can continually adapt and improve its strategy as it gathers more data from real life trading. Of course, the machine driven decision support will still need to be validated by human traders. The authors expect that it is only a matter of time before AI is employed in Repo trading. Greater data standardisation and transparency brought about by regulatory reporting mandates such as SFTR will help to provide larger data sets for AI algorithms to consume and will accelerate this trend. 6 BROADRIDGE

7 CONCLUSION With market participants moving to more advanced trading techniques and repo activities remaining critically important in the effective functioning of financial institutions, many participants have invested in or are evaluating these emerging tools. If your firm is not currently using more sophisticated and automated strategies for executing trades, your competitors almost certainly are. This can result in missed opportunities, rendering manual trading activity more difficult, with thinner margins. In a market meltdown or liquidity crunch, traders may struggle to obtain scarce liquidity as they repeatedly lose out to other, more forward thinking market participants. While helping to mitigate against the above circumstances, automation can also free up time that enables traders to focus on a more strategic role in sourcing scarce liquidity, optimising collateral and maximising P&L. It is becoming clearer that Repo broker dealers, and even buy-side participants are now divided in two camps; the ones with automation in progress, and the ones without. Which one is your organisation? FURTHER READING Article - The Changing Role of the Repo Trader: broadridge.com/article/the-changing-role-of-the-repo-trader Case Study - Global Inventory Management: broadridge.com/_assets/pdf/broadridge-global-inventorymanagement.pdf THE BROADRIDGE REPO ORDER QUOTE (ROQ) SOLUTION Broadridge has developed a state of the art Repo Electronic trading solution, with a European bank client going live at the beginning of The Broadridge Repo Order Quote solution offers a front office repo trading tool for multi-market aggregation and execution. The solution enables traders to gain a clearer view of liquidity and explore market colour across venues. It also offers the ability to quote and execute on an aggregate basis, with automated execution logic that allows users to rebalance across markets to make the best use of market availability. For further information please see: resource/repo-order-quote-rrq-brochure ABOUT THE AUTHORS Jerome Cardon is Repo Product Manager at Broadridge. Jerome has over 20 years of experience gained in both the Consultancy and Investment Banking industries, including senior level roles at Lloyds, RBS, Capco, ING and Deutsche Bank. Jerome has a specific focus on risk management, regulatory compliance and change management and in a previous role undertook the design and delivery of CVA, FVA, RWA, and Collateral Optimisation desks for a European financial institution. Martin Seagroatt is the Global Marketing Director for Broadridge s securities finance and collateral management business unit. Previously he was Marketing Director at 4sight Financial Software. In his 13 years in the industry he has specialised in securities lending, repo, and OTC/ listed derivatives collateral management solutions. Prior to that Martin worked as a business expert in technology systems for risk management in the energy industry. Whitepaper - What are the Applications for Artificial Intelligence in Securities Finance and Collateral Management?: Brochure- Broadridge Repo Order Quote Solution: broadridge.com/resource/repo-order-quote-rrq-brochure AUTOMATED TRADING IN THE REPO MARKETS CURRENT AND FUTURE TRENDS 7

8 Broadridge, a global fintech leader with over $4 billion in annual revenue listed on the S&P 500 index, provides communications, technology, data and analytics. We help drive business transformation for our clients with solutions for enriching client engagement, navigating risk, optimizing efficiency and generating revenue growth. broadridge.com Communications Technology Data and Analytics 2018 Broadridge Financial Solutions, Inc., Broadridge and the Broadridge logo are registered trademarks of Broadridge Financial Solutions, Inc. CM_00024_BR_18