Every day, millions of trades are made by financial institutions around the globe. Not all of them are in accordance with market rules and some of them might constitute market abuse. Recently, the European Union has issued the Market Abuse Regulation (MAR) to address market abuse.
MAR requires trading institutions to monitor their trading behaviour to detect regulatory breaches and subsequently report them. In this context MAR updates the regulatory regime for market abuse for easier enforcement.
Here at Capgemini, we have developed a Trade Surveillance System (TSS) to supply clients with a monitoring system that supports compliance functions to meet current and future regulatory requirements.
Indicators of market abuse
When complying with MAR, alerts and reports provide valuable feedback to increase precision, reduce risk and lessen the occurrence of false positives. Our TSS is designed to ensure compliance with MAR.
The regulation describes potentially abusive activities. It presents a list of indicators relating to false and misleading signals and price securing, as well as a list related to fictitious devices and other forms of deception.
The adaption of this list in our system includes 17 indicators of market manipulation. The indicators are evaluated to compute their probability of occurrence. These indicators include aspects of trading such as:
- an unusual concentration of transactions of arbitrary product types,
- simultaneous selling and buying of a product type through different brokers,
- large numbers of order cancellations.
To achieve business compatibility and easy interpretation of the results the indicators are combined to describe classic techniques of market abuse such as:
- marking the close,
- painting the tape,
- wash trading.
Concept of a Trade Surveillance System
Our TSS consists of four distinct layers that clearly separate the user interaction or front end and the underlying data handling and storage functionalities:
- Data source catalogue
- Computational layer
- Visualization layer
Fig. 1: Architecture of the Trade Surveillance System.
1.Data source catalogue
Data selected from a data source catalogue are imported into a central database. In this database information about trades, orders and market-specific information are stored.
Data required include trade data (prices, time stamps, volumes, etc.), order data (bid and offer prices, volumes, etc.) and market data (average bid and ask prices, total market volume, etc.).
Also data that is usually discarded, such as price forecasts, are listed here.
The database layer functions as central data storage of all MAR-related data. All data queries are executed in this database so that no live connections between the visualisation front end and other user defined sources are necessary.
There are three roles the analytics layer needs to fulfil:
- First, it models the client’s role in the market, since regularities in the trade data allow an assessment of anomalous behaviour.
- Then it computes indicators for market abuse through data mining, statistical analysis and machine learning
- Finally, it determines the likelihood of a compliance breach. Here, the simple probabilistic deviation from normal behaviour is insufficient. Instead, we employ advanced data science methods, e.g., supervised machine learning techniques.
The indicator system is linked to a list of model cases of market abuse and a traffic light system indicates possible compliance breaches of any MAR regulations.
The visualisation layer, finally, aims at presenting all information required to initially assess, if a trade that has been highlighted as potentially manipulative requires further investigation. Each dashboard includes filters that allow the user to pre-select information relevant to the investigation at hand.
The main advantage of this type of visualisation is easy and straight-forward adaptability to client requirements. Also user defined modifications are possible.
The main dashboard lists the trades that flag as yellow or red. Using this information the dashboard user can further investigate the rationale behind the trades. Additionally, a trading summary is presented that lists profit and loss, price and volume at the client compared to market values, etc.
Fig. 2: Market situation dashboard.
The trade dashboard displays all flagged activity at the day of the selected trade. Using this view the user is able to understand the environment of the trade and determine its plausibility in the business context.
Finally, the trader dashboard provides extra information about selected traders to understand their actions regarding their ordering and trading behaviour.
The TSS concept demonstrates two key strengths: cost efficiency and scalability.
By offering surveillance capabilities around a specific scope the TSS is cost efficient. It is aimed at small and medium sized enterprises that are incapable of spending large sums and are interested in co-development of Trade Surveillance around our experience.
The TSS infrastructure is a platform for other use cases. Examples within trade include assisting trading strategy, monitoring of trading mandates (tasks and allowances of traders), use of the TSS as a portfolio management tool, or first steps towards robo advisory or algorithmic trading.
By using the indicator system the TSS allows applications outside of trade, e.g., for other regulations or in reporting for risk or financial purposes. A comprehensive representation of trade activities for other departments is easily possible and could even provide the base of a system of quantitative and analytic trade strategies.
I thank Torben Schuster and Christian Petersen for their support writing this article.
This article was written by Christoph Euler from Capgemini: Business Analytics (UK) and was legally licensed through the NewsCred publisher network.