Automation of buy-side trading processes has come a long way in a short time. Are we reaching the law of diminishing returns or are there new advances around the corner?
New tools and technology have undoubtedly changed trading and execution for larger asset managers over the last decade, but what about the long tail?
You mean the smaller or perhaps more traditional asset managers that haven't embraced algorithms, smart-order routers and real-time transaction cost analysis etc. in the last 10 years or so? Certainly, there are plenty of them around. There are many long-only firms where the prevailing culture still dictates that the dealing desk implements the portfolio manager's decision with very little discretion, for example.
This might mean pushing the order down an algorithm with very strict controls, but a highly hierarchical PM-led investment philosophy might also require the dealer to pass on the execution risk in the traditional manner, i.e. by placing the order with a sales trader down the phone. No new tools and technology needed there. Similarly, there are those firms that, perhaps because of size and resource constraints, have not invested in new technology consistently over the past decade. In addition, no doubt some niche hedge funds are happy using the execution management system provided by their prime broker. Moreover, their volume of activity in the market simply does not warrant detailed analysis of which algorithms will achieve the best execution result.
So, you agree then. New tools and technology have only had a peripheral impact on institutional investment?
No, of course not! There might be the odd pocket of manual trading here and there, but the overwhelming majority of asset managers have devoted increasing levels of time and effort to deploying technology more effectively to retain the alpha in their investment ideas. What is more, the trend has only intensified in line with greater pressure on asset managers in the post-crisis environment to compete for business from end-investors.
In short, a highly skilled, but highly automated dealing operation is a pre-requisite for taking on more business. End-investors are pretty savvy on costs these days and high transaction costs are a barrier to winning mandates. Algorithms for example were initially marketed as a trader productivity tool that could automate the nuisance trades and allow the dealers to focus on the more challenging orders. For many, they have been the key to absorbing acquisitions and taking on new mandates without growing the overheads that a global, high-volume trading operation would otherwise attract. As OTC derivatives and fixed income trading gradually become more automated, buy-side desks are becoming more multi-asset in nature. They can only do this if the machines do the heavy lifting, leaving the humans to refine and adapt the overall trading process to circumstances and the variations of different geographies and asset classes.
OK, but have we plateaued in terms of returns on investment in trading technology?
Some would argue we've barely scratched the surface. As The TRADE’s 2014 Algorithmic Trading Survey (published in the Q1 2014 issue) reflects, the pace of growth both in number of algorithms used and overall algorithmic trading volumes has slowed. Behind these numbers, there are a lot of firms drilling down into the data to identify differences in performance and opportunities for improvement. The amount of volume put through the automated trading channels of larger institutional investors means that incremental gains all add up. Increasing transparency from the sell-side on order routing and greater message standardisation via FIX will only add to the potential advances. And that’s even before we start considering the potential opportunities for automating the trading and investment process presented by social media and big data.
Click here to vote …
Click here to find out more about The TRADE 100