On July 19, U.S. authorities arrested HSBC Holdings PLC’s global head of foreign exchange cash trading for his involvement in a series of trades effected in December 2011. The trades in question are nothing more than front-running. Here’s the gist of what happened:
Cairn Energy, a UK-listed oil and gas producer, approached HSBC in 2011 to convert $3.5 billion in proceeds from the sale of an Indian subsidiary to pounds. With the knowledge of a large buyer entering the market for sterling, the HSBC foreign exchange desk engaged in aggressive buying in the days and hours leading up to the conversion. On the day of the conversion, the HSBC desk liquidated their position generating $3 million in profits for the bank on top of the $5 million in fees tied to the client’s transaction.
I have no idea whether or not these trades were flagged by internal surveillance teams and simply swept under the rug, but I can assure you that this type of activity, combined with such a serious breach of client confidentiality, is very detectable with the right tool set.
So what is the right tool set? One that doesn’t focus on zero-day analysis.
Zero-day refers to the number of days between the time a trade is effected and some external event (like a client’s pending currency conversion). The post-trade analysis on the day of the currency conversion likely didn’t raise any questions. You have an FX desk selling out of pounds and a client buying into pounds. This is exactly what you would expect. But what did the days leading up to the conversion look like?
Probably something like this:
This is an example of a 3-month rolling volume profile analysis where the light blue line represents the FX client trades and the dark blue line represents HSBC trades in the cash market for GBP. It is evident that a pattern of escalating bank trades emerges shortly after Cairn Energy approached HSBC. If rolling volume profiles were put in front of a trading surveillance analyst in real-time each day, this most certainly would have raise a red flag.
Of course, this is a very simplified example. The complexity of the bank’s FX transactions surrounding this event would require a more sophisticated analysis, and potentially, the analysis of trades in other markets. But in an event like this, one thing is certain – it is impossible to rely on systems that ignore the tremendous amount of data available to us in the transaction history. Post-trade analysis that begins and ends with the days trading activity is simply ineffective.