By Simon Richards, CEO, Fonetic USA
Three years on from the Dodd-Frank act, passed by Congress in July 2010 in response to the 2008 financial crisis, banks are still coming to terms with its implications and how it impacts their day-to-day practice. Supporters hailing the emergence of Dodd-Frank as the answer to a corruption fed global banking crisis have become less vocal in the years since it came into law. As the economy moves from recovery to growth mode Dodd-Frank faces a wave of industry opposition against the restrictions it imposes.
In fact, according to law firm Davis Polk & Wardwell LLP, only 52 percent of the 398 rules mandated by the law have been completed as of April 1st this year. Fuelling the dissent is Senator Richard Shelby’s 50 page technical reform bill issued in March 2013 which highlights a raft of flaws in the legislation. Shelby’s bill serves as a reminder to the haste with which Dodd-Frank was first drafted. As any lawyer will testify, misleading or erroneous grammar, plus poor context and false references can alter the interpretive meaning irreparably.
However, despite these setbacks, the impact of Dodd-Frank is now beginning to be felt more keenly. At the end of April, Hugh ‘Skip’ McGee, the chief executive of Barclays Americas stepped down with the bank claiming McGee’s departure was a result of the regulatory burden it faces in the US. When the rules of the game change, some people just don’t fancy playing anymore.
Many banks have yet to understand that when it comes to compliance, technology can play a key role in making the transition as easy as possible. In recent years, banking technology has evolved to keep pace with the changing regulatory landscape. So by using technology to bring what is often experienced as a piecemeal governance programme into line, banks are ridding themselves of a lot of the resource intensive regulation and compliance monitoring and shift the focus back to making money once again, safeguarded against overzealous risk taking.
Putting compliance aside, when looking at some of the specific requirements of the Dodd-Frank Act, some of them make good business sense and will deliver real business value. In particular, trade record keeping enables greater transparency into any individual transaction, meaning banks can track anything out of the ordinary and call up trade history as and when required.
Trade analysis and reconstruction are two parts of the new regulatory environment that place huge demands on investment banks. The only way in which banks can reconstruct a trade right now is to physically listen to all of the phone recordings which could be months and many thousands of hours of recordings. To put this into perspective, one month’s worth of calls is 160 hours of listening, which would take one person eight hours every day for a month.
Aside from the financial burden of hiring individuals to listen to other people’s calls is the human error factor. Imagine listening to eight hours of almost identical, seemingly unrelated calls every day, even the best listener in the world would probably switch off. There is no way to tell if a call with the content “Hi how are you? Mine” relates to a trade without linking the meta data that links the voice recordings and linking that to the front office systems. The task of trade analysis and reconstruction is impossible to do manually with any degree of confidence in the accuracy, so the banks must turn to an automated transcription service.
Automated voice/speech recognition is not new, indeed, anyone who has ever used a customer support call centre will have come across basic voice recognition software. Its use has become mainstream driven largely by its adoption in mobile telephony. In most instances, voice recognition is designed for a specific purpose since each environment has different requirements. Trading environments are arguably among the most challenging of all for voice recognition software. Many traders speak a number of languages – transcription based alternatives cannot transcribe words that are not easily recognised – add a trader speaking his non-native language i.e. a German speaking French and add to that the noisy environment and low grade quality of recording and the data loss is significant.
If you’re looking for a voice recognition supplier for financial services you need to find a firm that has specialised in financial markets from the outset. Not only must the system be optimised for the physical conditions of a noisy environment with low-grade recordings, it must also be fluent and multi-lingual. You need a system that can understand myriad languages, accents and slang terms – specifically slang that is used in the trading environment. So, in short, you need a system that can not only listen, but also understand and alert.
Most compliance officers say that 99 percent of what happens on the trading floor is fine – it’s the one percent that is of concern, but they can’t find that one percent with their current systems. You don’t know what you don’t know – being able to set up automatic alerts which flag a potential problem is a massive revelation.
Trading floors are complex environments. Traders close deals directly on the phone, dealing with the documentation later. They use slang, short conversation, and they manage a significant amount of money. The current Dodd-Frank Act in the US and future MIFID II in Europe will implement significant changes and will create a new regulatory framework by increasing transparency in financial transactions. The challenge is massive. Recording and keeping mails is only the start, the real challenge is how you listen to all the conversations and make sense of them?