Buddy can you spare an algo?
I was catching up on some reading last night and came accross this editorial by Eugene Grygo at DWT entitled ‘Will an algorithm take your job?’
He comments on an article from Waters magazine about Knight Capital Group:
[Knight Capital] freely acknowledge that algorithms that manage the firm’s retail order flow replaced 210 retail traders. From 2003 onward, the firm’s roster went from 250 traders to 40 as the firm has increased its use of algorithms – now up to 90% of the firm’s retail order flow. Knight officials say that trading via algorithms in this space cannot be ignored because it is so much faster than anything humans can achieve. In addition, the care and feeding of algorithms is much less expensive than that of their human counterparts. The algorithms are kept updated by Knight’s quant traders on the West Coast of the US. It has led to a new business model for the firm.
This is indeed the business model of the future. And yes there obviously is some truth to the old say that ‘humans will (always*) be needed, only their role and jobs will evolve over time (for instance to writing, managing and adapting algorithms.) But Eugene flags something that is hardly ever uttered out loud in polite company: that the jobs of many of today’s traders and professionals will not survive in the future, however redefined the roles may be. On the one hand there will not be a need for the same numbers, and on the other many will be unwilling and/or unable to adapt. A good (rough and ready) comparison could be made with the transition of European futures trading from floor to electronic. Many floor traders moved on and adapted and have continued to be extremely successful. But many others were forced to reinvent themselves completely and often outside the industry.
But the resistance to change goes deeper than just the reluctance of professionals to embrace their own obsolescence. The economic and management models of most large financial institutions make a fundamental and deeply entrenched distinction between human and technological capital (in budgeting, performance measurement, etc.) – the two are very rarely interchangeable below the very highest executive level. This usually results in making iterative and tactical choices between human and IT resource impossible. Reducing sales and trading front office staff by 20 to hire 5 developers and to buy 100 servers is not a choice availble to most line management in most large banks or securities firms. At least not outside annual (or longer) budgeting cycles, and even there you often bump up against natural divisional and power politics protecting or promoting a status quo.
Large organisations have many advantages however preserving business model flexibility is an enormous challenge and one that – with the economics of the internet age – is most likely going to become more rather than less important. And it will just possibly create the opportunities needed for a new breed of financial institution to emerge on the crest of the changing face of markets.
(* I know that ‘always’ and ‘never’ are often used in a ‘forseeable future’ sense, not in their absolute sense, however at the risk of being pedantic, they really make me uncomfortable in a world that is changing as fast as ours; check out this great article on quantum computing from a recent edition of Fortune if you want a better flavour of where I am coming from…)
