With so many vendors offering a variety of tools for just about every compliance process, some which deal with very discrete processes only, is there a danger of firms missing a key trick in their efforts to detect and deter damaging market behaviour?
In this first post, I will cover some key points at a very high level for firms looking to take control of their data and will expand themes which warrant further detail in follow up posts.
A surveillance system can play a key role in detecting some dubious behaviours and patterns, but if used in isolation are you really seeing the bigger picture? With some vendors now expanding their offering to incorporate voice detection in their trade surveillance systems it seems the industry is crawling to where it needs to be.
While reviewing alerts could highlight some suspect patterns, they are alerts on a subset of your data and from what I hear can still yield a lot of noise. Many vendors also offer end users the ability to build tests (queries) within the vendor application and adjust thresholds. Useful, but when looking at one process in isolation you could be missing valuable pieces of information (context) which could be the difference between "might be" and "definitely" abusive.
Another way of using all of your data could be to take a more holistic approach. One where if, for example, you can follow the flow of material non public information (MNPI) through your firm you might learn of breaches in your Chinese Wall or suspect position building.
An approach where you put your data to work for you.
Holistic surveillance – taking all of your data into account needs to start with:
Creating pots of golden source data
Being able to strategically read across those pots of data
Carry out analysis on anything that doesn’t look right
1. Examples of golden source data
This almost goes without saying. Unless your data is good and in good order, your output will suffer. I’ve always been of the view that many eyes on one data set is better than fewer pairs of eyes on many data sets. Besides local teams are more likely to pick up on anomalies sooner with the former approach. It’s a case of consolidating so that for example you have one (reference) pot for trade data. One pot for MNPI. One pot for PA deals etc…
2. The power (and purpose) of reading across data pots
This will ultimately form part of your data strategy which I encourage firms to think about before designing any overarching system and always keep in mind when enhancing the system (i.e. does the strategy still work?) Once you have your trade data in database A, your HR data in database B and telephony data in database C and you’re able to read across all data sets then all of a sudden you start to bring your data to life. You can start to piece together a chronology of events from data you already have. (This will be more powerful with market data, but I will save that for another blog post).
3. Getting clever with picking up trends
You have good clean data, you have a strategy and you realise the powerful insights you can gain. So what next? Fundamentally you need to know what normal versus suspect looks like. Learn normal and use your data to pick up on suspect trends. For example three members of your corporate broking team return from seeing a new potential client. All of a sudden you see three new PA deal requests. All buys, all in the shares of that company and none of the requestors have any history of interest in that issuer. Unless your pipeline tool is linked to your PA dealing system, compliance approvers will be none the wiser.
4. Human capital
The final piece to effective surveillance is an enquiring mind. ESMA in their RTS for the Market Abuse Regulation (coming into effect July 2016) state the following:
“Effective monitoring involves much more than just a surveillance system and must include comprehensive training genuinely dedicated to monitoring, detecting and reporting suspicions of market abuse or attempted market abuse. Training plays a key role in staff’s ability to detect suspicious behaviour”
It is interesting ESMA have included this in their final text and highlight the need for more than just a surveillance system. It also adds the argument that not only do colleagues in your team absolutely need to know what they are looking for, but also how it could occur and have the tools necessary for detecting it (i.e. a holistic approach?)
Though banks, brokers and fund managers are primarily concerned with raising and managing capital, it may be the case the ones who don’t also consider themselves managers of data will always be on the back foot when regulatory change (e.g. MAR, MiFID 2) comes about.