Artificial Intelligence in the context of Alternative Asset management will manifest itself across a number of dimensions...
Artificial Intelligence in the context of Alternative Asset management will manifest itself across a number of dimensions; the need to understand the vast quantities of data being generated by economic activity moving online, the ubiquity of networks for conducting these online transactions, and the emergence of algorithms for seeking investment opportunities and making investment decisions. There has already been a rise in the number of Quantitative Funds – such as Aidyia- relying on machine intelligence, neural networks and the ability to mine Big Data to compete alongside the likes of Renaissance Technologies, DE Shaw, Two Sigma and a host of other brand name funds such as Bridgewater Associates and Point72 Asset Management that investigating the scope for AI to generate Alpha.
Active vs. Passive AI
The rise of this machine intelligence can be broken down into two categories: “Active” AI uses technology to seek out opportunities for hedge fund alpha by analyzing data in unconventional ways, uncovering hidden correlations in market data or interpreting “signals” in the marketplace which then generate trading ideas. “Passive” AI takes place in the post-trade arena, opening up time-series of data, connecting flows between markets, counter parties and assets to support Portfolio Managers and Traders in their decision making. By using AI to open up post trade data more aggressively, funds could find their optimal financing mix or most credit-worthy counter parties by searching the market for news that may impact their credit rating or actively size/shape trading portfolio positions in real-time.
The old distinctions between front office, middle office and back office are largely defunct in the world of AI. The back office functions of clearing and settlement will generate data to be mined, the middle office will do the same and the risk, pricing and transaction data from the front office will be combined with this to uncover opportunities for maximizing returns and reducing investment risk. These AI techniques will go beyond the relatively simplistic notions of VWAP and traditional Transaction Cost Analytics (TCA), it will encompass a wider array of metrics (e.g. order book depth and liquidity fragmentation). Another major consequence of the use of AI in the manner will be a greatly reduced need for headcount to support the overall fund management business.
You may also like to watch
In the post trade functions of middle and back office the emergence of distributed ledger technology such as Blockchain is being actively looked at by banks/exchanges to eliminate clearing/settlement cycles or to limit counter party credit risk and overall systemic risk in the market. Under a Blockchain model “smart” contracts could reduce counter party credit risk by enforcing re-hypothecation limits or “tagging” collateral for easy location/recall in the prospect of default. An intriguing possibility opened up by AI and machine learning is that of an Operations function that is a machine/human “hybrid platform”. Incoming flows/tasks/processes are analyzed by an AI “engine” which can assign “jobs” and learn as humans process tasks and eventually perform those tasks itself, learning the decision making steps performed by human workers and organizing accordingly.
Big Data, Big Need for Talent
Previously Portfolio Managers at hedge funds relied largely on their own – analog – information sources to do their research and generate ideas (e.g. expert networks or investment banks) but using AI they will be able to sit in the nexus of information flows from the market, their own fund and other disparate information sources and using “deep learning” technology that can recognize images and speech, look for novel investment ideas. The virtual feedback loop of more economic activity moving online requiring the ability to mine, analyze and interpret that data will become a key table stake for all funds, not just those reliant on AI for generating their trading alpha.
This phenomenon has already started and is causing a surge in demand for personnel with “big data ” backgrounds such as data scientists in fields such as bio/life sciences, genetics, particle physics, astronomy, nanotechnology etc. The ability to recruit and train this specific talent pool is becoming a competitive differentiator for many alternative asset managers as well as the entire financial services sector.
Dangers Lurk Ahead.
The rise of AI also brings with it the prospect – remote at present – of “rogue” Algos that could cause damage the functioning of the financial markets. High Frequency Trading (HFT) requires that exchanges closely monitor trading flows and spikes in volume to enforce circuit breakers so that “flash crashes” do not occur. The is a pre-cursor to some of the market-wide safety measures that will need to be taken to ensure that AI, Algos and the like do not become a new source of systemic risk to the markets.
Similarly hacking or industrial/crypto espionage and cyber terrorism are all credible dangers in the world today and the scope for the same sources to manipulate prices or market liquidity will need to be closely monitored with appropriate counter-measures deployed.
Despite these dangers the rise of AI in Alternative Asset Management and the wider investment management space appears inevitable. Investing itself may become more fundamentally “imaginative” in its use of emerging technology, combining AI with drone technology to check a company’s supply chain logistics or the number of customer cars in a car park.
The future is here. IT is called AI.
The author of the article is Bijesh Amin, Co-Founder, Indus Valley Partners