One of the reasons I enjoy our conferences so much is the capacity of the high quality speakers with whom we are fortunate to engage to raise a point that just makes you think, “why have we not done this before?”
This week we held our inaugural Frankfurt conference and during our Technology Futures session the discussion turned to the role of technology in making markets safer. It was one of those great sessions where I turned up to moderate armed with a bunch of themes and questions following (ahem) “extensive research” and got to ask one of them! As with all great conference sessions the speakers drove the agenda and I was able to very much stand back and enjoy with the rest of the audience.
But out of the blue John Ashworth, CEO of Caplin, decided to use an interesting analogy to describe the importance of man and machine working with each other. He described the famous events that led to Chesley Sullenberger landing his stricken airliner on the Hudson River in New York in 2009.
John pointed out that the computer was still making most of the decisions while the pilot dealt with the big issue – what was he going to do to get his plane and passengers safely on the ground?
So far, so normal we were all thinking, but then it was pointed out that in interviews afterwards (probably for the rest of his life) Sullenberger observed how he had been trained incessantly for just such an event. Obviously, as history shows, the pilot made some great decisions and saved his passengers thanks largely to the fact that the computer allowed him precious seconds for that training to kick in.
John Ashworth then observed that Sullenberger didn’t really practice for take offs or landings, but he did practice, as all airline pilots do, for emergencies, adding the pay off that in FX markets we don’t really provide ongoing training for staff to deal with market events, the way the airline industry does.
Obviously one is a matter of life and death and the other is about saving people on the plane but it does raise an interesting question – the industry has the computers to manage matters when markets are “normal”, but as has been proven in fairly recent events, they can’t handle events as they either switch off or exacerbate the move.
I do have to share the observation of a senior banker in the audience who is heavily involved in the e-business and told me later that he believes that the computers can handle flash events now, thanks to the lessons learned from SNB day and the Cable flash crash. Not being one to argue(!) I am less convinced and still believe we will see flash events – they may not be on the scale of the aforementioned, but they will still occur. At some stage in the future, the computers may handle the issue, but what happened in the skies over New York only serves to highlight the challenges involved.
Which raises the question, should we – as an industry – engage in formal training exercises for the human oversight function? I am talking both e-trading managers and human traders. As a fellow speaker pointed out, we run disaster recovery exercises for technology and infrastructure and algos are tested and tested again, but we don’t test the ability of our humans to cope with flash events – or even train them to cope.
I accept that many institutions have vastly experienced people in key positions and that counts a lot because traders of a certain generation have shown during the flash events that they can make some serious money for their institutions, but could a training programme actually help avert some of the worst excesses?
There is a sense that we could be asking a trader to stand in front of a runaway truck – and there will inevitably be time taken to check that nothing serious has actually happened on the news front – but given how machines operate on data, if a human trader has the backing of their institution and can take some risk, their bids or offers will help fuel the machines with data, and they in turn will put in liquidity and suddenly our flash event is limited to a few hundred points at worst, rather than 10 big figures or more. Equally, the e-trading manager can override the platform and post interest electronically.
Regular readers will know that I help run the ACI Australia Dealing Simulation Course globally and we have been simulating market events for years (given how we try to slow things down for people to learn and still use a lot of voice trading I can’t really call them flash crashes) with scenario-based moves in the 10s of big figures. That course, however, is squarely aimed at newcomers to the industry and staff in the support functions such as compliance, risk management etc, it’s not there for more senior traders.
I suppose the problem is that such a training programme might be seen as tantamount to an insult by some experienced traders and senor managers, some of whom have proven their ability to cope with flash events, but that shouldn’t stop the debate occurring over the relevance of such a programme. Flash events are a risk, potentially to institutions, clients and individuals, therefore they should be mitigated against as comprehensively as possible – and training is one such aspect.
I happen to believe that if we had a few more experienced human traders in the FX market then it would be a safer place, but those traders have to have experience and be backed up with risk limits – as do the e-trading books. This could be seen as suggesting that I don’t see the need for training but I think that is too narrow and, perhaps, arrogant a view.
Of course the computers are going to handle more and more flow, that is fine and how it should be. But as the world found out in New York in 2009 having a human function that can step in and work with the technology can be a life saver. That should override any damage to pride – after all, airline pilots are professionals but they submit themselves to a rigorous training programme, why shouldn’t traders?