How to approach communications surveillance in energy trading? In this blog we talk about the 5 levels of monitoring and what the best approach is for the energy and commodities sector.
The status-quo for energy firms
Energy trading firms often trade a wide and complex set of products, many do not fall under the direct supervision of regulators such as the FCA or ACER . Historically, this has led to the limited adoption of surveillance technology by energy firms leaving them somewhat behind the curve compared to their peers in other sectors such as the sell-side.
However, with the implementation of MAR and REMIT over the last few years, energy firms have invested in trade surveillance technology. With more in store for 2019 such as Brexit and the “REFIT” changes, firms are looking towards a wider set of monitoring tools. During a recent roundtable we hosted for energy and commodity compliance practitioners where we discussed communications monitoring in energy trading – Click here to download the conclusions
Still a lot of voice trading
We found that, although there is still a lot of voice trading in this sector, reviewing calls is typically performed manually. Spot checks are usually carried out or deeper dive investigations happening when perhaps a trade surveillance alert occurs. Otherwise calls are typically recorded and stored, available for compliance teams to review as and when.
This got us thinking about the different levels of communications monitoring and how each level brings added benefits.
Gradually increasing your monitoring capability enhances and improves your surveillance program.
Fonetic can guide you through this process and provide an intuitive and easy to use system powered by state-of-the-art AI & ML technology.
What are the 5 levels of communications monitoring?
When monitoring communications there are typically 5 levels, each increasing in complexity for a more complete and comprehensive surveillance program.
1. Recording calls
The first step to any surveillance program is to start recording, of course. However, that is where a lot of firms stop. Recording calls is a requirement under MiFID II, article 16.7. However, we don’t see any value in just recording and storing these audio files away on a server collecting dust. They hold vital information which can be utilized not just to achieve better compliance but also to improve certain business processes using insights gained from customer interaction.
2. Listening to calls on an ad hoc basis
The next stage is making these recordings available for compliance to review when necessary. This is otherwise known as sampling or manual sampling and has a very limited reach. While some smaller companies can make do with this method due to low call volumes, it is not an appropriate measure for larger firms.
3. Doing lexicon-based searches
There is a wide variety of communications channels used by traders as part of their daily business. These include channels such as ICE chat and Bloomberg chat as well as emails and texts. Typically, these channels are authorised for use by compliance and some lexicon-based monitoring occurs across these channels.
In some cases, lexicon searches have been able to highlight trader discussions worthy of further investigation. One roundtable attendee found monitoring swearing and abusive language was an important trigger for further investigation.
However, using only lexicon-based searching can result in a higher number of false positives. This is because the context is not being considered at the same time. A lot of software which centres around lexicon only analytics do their searching across communication metadata without analysing the communication’s content. However, when both the content and metadata is analysed together, this allows for more powerful analysis and more refined alert results.
4. Correlating communications surveillance with trade data
A big issue for many firms is achieving effective compliance across multiple data sources. Often there is gaps in knowledge due to various silos of information between communications data and business data. A more ‘holistic’ approach can be very beneficial for firms. Technology such as Trade Reconstruction is able to provide that bridge between communications data and trade and business data.
If compliance is able to analyse alerted communications and their associated trades and employees at the same time, this can result in faster and more comprehensive investigations and rapid resolution of both regulatory requests and internal audits.
5. Using AI to identify suspicious activity
The adoption of AI based tools to interrogate large data sets and uncover suspicious patterns is an efficient approach. In fact, technology has evolved a lot in recent years and when applied at the right stage of the analysis process, it can produce increased levels of accuracy and detect signs of intent. This is most advanced level of communications monitoring.
However, AI and ML technology for the trading floor should be deployed with the right approach and in the right measures. Compliance practitioners we spoke to were less interested in a black box generating alerts and preferred to have control over the collaboration of rules, much the same as their experience with the roll out of trade surveillance alerts.
It's widely known that applying AI to bad data is likely to lead to bad results. Getting data quality to the highest levels of accuracy before applying and AI techniques was the best approach.
To read more about the conclusions of our recent energy and commodities roundtable event, you can download the conclusions by clicking here.
If you found this interesting, you might also like: