Dodd-Frankenstein? How the US Election could change Wall Street forever
US regulators have shown unusual reticence on mandating the phone recording of traders, wasting time and money for companies facing litigation
Today, US voters will be worrying about unemployment, taxes and healthcare... Except if they are working on Wall Street. The Dodd-Frank Act takes a large thumb, and presses it firmly on US financial institutions, and foreign firms who deal in the US. And if Romney wins, he has said he will scrap it.To put it in perspective, Standard & Poor's have said that the cost of Dodd-Frank to just the top 8 US lenders is likely to be $22 billion annually, which is the GDP of El Salvador.
Dodd-Frank is designed to protect the US consumer from the excesses that caused the 2008 crash, by amongst other things limiting exotic financial instruments, creating a Financial Stability Oversight Council, and increasing the size and scope of recordkeeping. It is this last part where some of the real long-term issues might come, especially for corporate counsel and their advisers.
Recording phone calls
To date, there has been a reticence by US regulators to follow the rest of the world and mandate the recording of telephone calls made by traders, which seems counter-intuitive as that is one of the key sources of information for an investigation, and where the meat of deal-making takes place.
The reasons are pretty simple. It is not really the cost of implementation: With modern IP based phones and trading turrets in use in all major financial institutions, capture and storage of calls is a solved problem. It is not the cost of storage: These days, you can easily store several years of one person's calls on a disk no bigger than the one in the PC under their desk.
Litigation is the issue
Call storage is easy: Call indexing is hard. Yes, you can quickly retrieve a call if you know who the caller was and when the call took place. But searching for what was said is a difficult technical problem. And made much more difficult by the call recording systems that most banks have in place.
E-discovery encompasses all electronic communications, and that includes voice files. And if suddenly you insist that all trader calls are recorded, as is already happening for Swap Trades under Dodd-Frank, you have to know what is said in those calls to allow you to properly comply with a discovery request.
The challenges of listening
If you have ever listened to call recording, you can begin to understand the challenge. Most call recording systems are designed to use the least amount of storage possible. Which means taking a stereo phone call, merging it to mono, and then compressing the file yet further. Frankly, it is often difficult for a human being to understand the call, let alone get a computer to try to help you do so.
This means if you are hit with a discovery request you have to have all relevant phone calls manually transcribed: Very expensive and very time-consuming. There are some short cuts, using semi-automated processes, but it is still slow and again, expensive.
Questions you should be asking
So what should you do? Well, you could vote Romney, and hope that he fulfils his campaign promise. Or you could take a good hard look at your infrastructure, and see if it is fit for purpose.
The questions you should be asking are:
- Do we record all voice channels separately? Bear in mind that trading platforms regularly support conference calls, so you want each channel to have its own file. Even a regular phone call should have an inbound and a separate outbound channel.
- Do we compress the stored call data, and if so, do we lose any quality? When a phone call hits your network, it has already been compressed, usually to one of the recognised phone-quality standards (u-law in the US and japan, and a-law in the rest of the world), and you don't want that degraded further.
- Could we capture our audio at higher quality? The outbound stream of a call (you to the outside world) could be transmitted over your network using a "wideband" compression algorithm, and if both sides of the conversation support it, the inbound could be "high definition" as well.
Why are you asking these questions? Simple. When you do get that discovery request, or a nasty letter from your regulator, you want to be able to use automated tools to search your voice files, and retrieve the relevant ones. This saves you time, money and potentially reputation. Better still, if you have better quality recordings in-house, you can start to apply analytical tools to your voice, e-mail and Instant Messages, and try to find that smoking gun before it's been fired.
If you do have a vote, you really might just want to go out and use it...