BUSINESS

OVERCOMING INFORMATION OVERLOAD IN THE FINANCIAL SECTOR

OVERCOMING INFORMATION OVERLOAD IN THE FINANCIAL SECTOR

Simon Price, managing director, Recommind Limited

With the financial services industry becoming ever more connected, financial organisations are increasingly facing challenges arising due to the resulting data boom. The sheer volume of unstructured data being created, coupled with the stringent regulations encroaching on the banking sector such as EMIR and Basel III, makes it imperative for banks to control collateral costs and optimise capital allocation.

Recent scandals like LIBOR indicate troubling times ahead for any organisations without adequate internal compliance programmes and tools to handle large volumes of company data for eDisclosure and internal investigations. Furthemore, investment banks now hold more Over-the-Counter (OTC) International Swaps and Derivatives Association (ISDA) agreements than at any time in history, and managing them is only getting more difficult given each OTC derivatives contract now has hundreds of data points. In some cases, banks have lost between $5 million and $25 million in a single trade after using the wrong interest rates, posting the wrong type of collateral or being arbitraged by counterparties.

Need for change

Failures in financial controls have cost banks large sums of money in fines that impact everyday operations and their ability to perform. For instance, the recent Forex investigation involving five banks cost the banks $1.7 billion in fines. This fine falls on top of money already set aside to fund the investigations and other legal expenses, which is often in the millions. It highlights that in today’s evolving regulatory landscape, financial institutions need to quickly get their hands on the facts to assess their exposure and respond appropriately to demands.

To mitigate against the risk of fines, banks are beginning to rely on solutions that automatically sift through tens of thousands of counterparty agreements and identify relevant data that traders need to drive profits – eligible collateral, interest rates, termination events, netting, thresholds and independent amounts. Increasingly, banks are moving from traditional processing methods that can take months to complete to using machine learning technology that automatically categorises data and extracts counterparty agreements for early insight. Overall, this drastically reduces data review time and helps eliminate error-prone manual processes that can lead to regulation headaches and substantial financial losses.

Unstructured chaos

A key example of data that is increasingly proving a challenge for banks is instant messaging (IM) conversation data. In 2015 alone, the number of global IM accounts totals to more than 3.2 billion – and in the time it took to read that statistic roughly 100,000 chat messages were exchanged. This is expected to further grow at an annual average rate of 4 percent in the next four years. When it comes to enterprise chat solutions, Instant Bloomberg is the heavyweight champion with some 320,000 users in the global financial industry, even as competitors like Slack and Symphony continue to grow as well.

IM conversations in particular present a unique problem in the highly unstructured data they contain. Unlike emails, IMs are replete with irregular formatting and extraneous metadata. Again, unlike emails with their comparably convenient subject lines and breaks, IM conversations continue indefinitely and may wind though multiple topics, making reviewing chat data that bit more challenging. The informal nature of IMs also means chat messages tend to exhibit more typos and slang than other types of written communication.

Bloomberg Chat evidence played a pivotal role in the LIBOR investigation. Meanwhile, professionals are already more acutely aware that chat is discoverable and have changed the way they communicate on chat by using more code names, slang, and omissions. As such, legal and compliance teams have routinely struggled to find the information that matters among large volumes of chat communications across the  organisation. In the past, the industry used compliance officers to process information, but today’s banks are producing billions of messages so it’s no longer feasible to manually review all of the information within the time required to build a case that satisfies regulatory demands.

 Teams can spend months wading through data, often against the clock. In the case of extracting the information from those conversations that matter, sometimes buried in a mess of IM data, accessing it quickly and cleanly is paramount. For banks and companies in other highly regulated industries, advanced analytics can help them weed out crucial information faster and with greater accuracy.. New, intelligent processing of IM data, however, can turn this challenge on its head. After all, metadata is a double-edged sword: for whatever complexity it brings, it also offers valuable clues for fact investigations and useful data points for culling document volumes. EDisclosure platforms can embrace chat metadata, map it as closely as possible to the email fields we’ve all come to know, and make it available for effective filtering and targeted searching.

A stormy climate

The regulator’s appetite for creating a more controlled market is only increasing. When allegations of manipulation erupt, financial organisations can be on the front foot when it comes to dealing with an investigation by investing in platforms that enable automated, repeatable processes and provide fast access to the relevant data. This helps banks better understand what has happened and where the responsibilities lie, which ultimately reduces the impact of investigations on banks’ balance sheets all while keeping litigation costs down.

To Top