Buy-side Perspectives Issue 3 - Page 6

Deutsche Asset Management is building a buy-side block trading desk with dedicated traders looking after outsize positions. The idea is that traders should focus on the largest blocks possible, while other flow can be automated. “We all have electronic trading in our DNA, we can all slice and dice, but block trading is a lost art. Put the most complicated trades in front of a human. That’s where they add the most value.” Deutsche Asset Management’s plans go beyond execution, however. The next step is to automate the broker selection process. This is done on past performance statistics, market share per instrument and other relevant data-driven factors. The idea is that when a trade instruction comes in from the PM, the automation software presents the trading desk with three potential brokers to choose from. Data is a key part of that informed process. The trading desk has statistics on block trading with each broker going back five years. According to Bellaro, Deutsche Asset Management runs the only buy-side trading desk that has access to such information. The data is doubly important, he says, because a lot of blocks don’t exist on the OMS – so working with the sell-side in an informed way is crucial. Beyond that, automation can be pushed further still – to have the software inform the trader about which algorithmic strategy would work best given the type of order. This would be based on analysis of previous best results. The idea is to help the trader to make an informed decision about which strategy to use – however the final decision would still rest with a human, at least for the more important trades. 6 The Buy-side Perspectives | Issue 3 | April 2016 “The buy-side and sellside utilising technology is the solution” said Bellaro. “We want to automate as much as we can and as intelligently as we can. Do I really need a sales trader to negotiate that sales trade for me? Can we automate some of that process? Is the next generation of algo design going to be on block trading?” One of the more controversial implications of the automation process is the potential automation of the sales trader role altogether. In the long-run, as more and more of the chain is automated, the possibility emerges to create some kind of ‘superalgorithm’ that sits on top of the entire process and takes into account all of the relevant information across the industry, using Big Data techniques to read the signals and warn of expected changes. Such a system would need to take into account real-time movements, as well as other feeds such as social media. While no such system yet exists, Bellaro is in favour of pushing the technology as far as it c