For all the recent talk of offshoring and outsourcing IT functions in financial services, as trading becomes more and more automated, the sector faces a shortage of technology talent.
This is one of the many findings of a huge, 184-page, new report by the Government Office for Science on the global future of computer trading in financial markets. In the next ten years “far fewer” human traders will be required and it’s not beyond the realms of possibility that some markets will require no people at all.
Welcome to the brave new world of ever-more complex trading robots that replace even the best humans. As Goldman Sachs’ CEO Lloyd Blankfein said this week, the bank is “much more focused” on technology, while Morgan Stanley is already replacing some bond traders with programmers and technology specialists as it looks to trade these instruments electronically.
By contrast, demand for designers and developers of hardware, automated trading systems and algorithms is set to increase, says the report, and there’s serious shortage in the UK: “Serious investment in coordinated programmes of research and development could help secure the future talent base (i.e. the pool of train scientists and engineers that have skills appropriate for work on advanced trading algorithms and hardware.”
Over the next ten years, these are the technological developments that are going to drive innovation in the financial sector, potentially reduce the need for human intervention and increase demand for the right technologists.
1. Cloud computing
Yes, any cynics please see through your fog of scepticism about the hype surrounding cloud computing for a second and focus on how it’s likely to be utilised by investment banks and hedge funds in the next decade. Already, firms have been tentatively adding cloud computing specialists, predominantly to develop in-house, internal clouds. In the future, however, they’re more likely to use the cloud to “provide cheap, easily scalable HPC [high performance computing]” in order to allow them the automated “design and optimisation of trading strategies and computer algorithms”.
“The ability to either extend existing in-house computing power by adding on externally provisioned cloud based resources, or to simply outsource all of the HPC provisioning to a cloud provider, opens up new possibilities that are only just being explored,” says the report.
2. Adaptive trading algorithms
Like (ahem) Skynet, trading algorithms of the future will not just be able to outperform humans, they will learn and adapt – primarily through interacts with other ‘traders’ on the market. This is not a new development (adaptive trading algorithms have been researched since the 1990s), but the “steep drop” in the costs of HPC is likely to accelerate their use. However, these algorithms may require little or no human design, so it’s unlikely to create too many jobs.
“A vast space of possible designs is automatically explored by a computer program: different designs are evaluated and the best-performing design found by the computerised search process is the final output. Thus, new trading algorithms can be designed without the involvement of a human designer,” says the report.
3. Easily programmable silicon chips
There’s already a move away from multiple core processing units in PCs towards using ultra high speed silicon chips that can be programmed by the user ‘in the field’. The most popular is the field-programmable gate array (FPGA), but programming these requires both a decent amount of time and specialist skills. It’s complex, but speed gains it offers means the move towards silicon chips is likely to accelerate in the next decade. Crucially, however, it will become a simpler process.
“FPGAs will be supplanted by a newer approach to custom silicon production, involving more readily field-programmable multi-core or many-core chips,” says the report. “Such chips will be programmable in a high level software language, much like current industry standard programming languages. This means that conventionally trained programmers can write algorithms that are then ‘compiled down’ onto the underlying silicon chip hardware.”
4. Forensic software
As less human intervention is required, security concerns inevitably increase, as the well-reported surveillance issues created by high-frequency trading highlight. And, the best way to detect market abuse by algorithms is to develop software to combat it.
“The development of software for automated forensic analysis of adverse/extreme market events would provide valuable assistance for regulators engaged in surveillance of markets. This would help to address the increasing difficulty that people have in investigating events,” says the report.