OPTrust, the $25 billion Canadian pension fund, began using artificial intelligence (AI) five years ago, consistent with its investment philosophy to be data driven and evidence based, and hopes to continue to build on this approach, said director and head of total fund completion portfolio strategies Romain Rouillon.
The fund’s mission is paying pensions today, preserving pensions for tomorrow, so Rouillon told Investment Magazine’s Fiduciary Investors Symposium the aim is to earn returns to pay those pensions and to reduce risk.
OPTrust embraced AI to improve investment outcomes with two goals in mind: increasing returns and reducing risk, using both deep and reinforcement learning.
“We’re a mature plan so, for us, drawdowns are very impactful. A big focus for the team has been to develop models that help us in cutting the tail [risk] and over the long term, achieving better returns.’’
He said equity risk was the main risk factor for the fund and it has trained “neural nets,” or artificial neural networks to get a sense of the uncertainty price in the markets and whether it is different from expectations.
Rouillon said reinforcement learning gives greater insight into the factors driving price action with AI, allowing detection of the non-linear impacts on prices that humans struggle to track. “Then we can, based on that signal, dial [the risk in the portfolio] up and down,’’ he said.
When developing new and novel strategies, it is important to get buy-in from internal stakeholders and leverage research and partners in the space, Rouillon noted
Further, the drivers of why machine learning will work in the future are not going to slow down, Rouillon said, highlighting the use of more data and that computing power is getting cheaper.
“What surprised me the most is the number of new techniques and the evolution we see in this space.
“Because it’s a tool, it’s used by a lot of different industries, then you have to adapt it to finance. But there are new techniques that are developed everywhere that I think are going to have a very strong impact on our industry and our jobs.”