As though Plan sponsors didn’t have enough to be concerned with when overseeing retirement plan assets, Artificial Intelligence is creeping into the investment function of qualified plans whether plan fiduciaries realize it or not. In the interest of getting to that Next Big Investing Idea before everyone else is using it, plan sponsors are tacitly accepting investment engines which use Big-data, Data Mining, Analysis of credit-card transactions (to identify purchasing trends), satellite imagery (to spot both auto and foot traffic), or filtering devices to spot trends on Twitter before the masses are aware etc.
Neither the math nor the actions initiated by the results are new. What is new are the devices used to collect such data, the methodologies used to collect the data, the logarithms used to analyze the data and the speed at which the captured data is converted into a buy or sell signal.
Is this Acceptable to Plan Fiduciaries?
Hedge fund managers use the latest technologies to analyze the benefits or drawbacks associated by incorporating artificial intelligence and machine learning strategies. Retirement plan asset managers compete with hedge funds, so is it wrong to act like one? That depends upon the industry’s acceptance of the technology and each retirement plan’s Statement of Investment Policy (SIP).
In many ways we have machines teaching other machines how to react when a certain set of circumstances is foreseen or may have already occurred. This is clearly the direction of the future and acceptable if your SIP has addressed the investing style and trading activities.
This is Where it Becomes Confusing for Everyone
A paper written by Max M. Schanzenbach, from Northwestern University, School of Law, and Robert H. Sitkoff, from Harvard Law School, notes that fiduciary status under ERISA imposes not only a duty of loyalty but also a duty of care. The DOL pending new Fiduciary rule has acknowledged, a financial adviser to a retirement saver will be subject to “trust law standards of care” in addition to “undivided loyalty.”
The fiduciary standard of care can be measured by the “prudent investor rule,” which is deeply based upon Modern Portfolio Theory (MPT). (MPT requires an investment strategy which employs risk and return objectives which are reasonably suited to the purpose of the investment account. Under the prudent investor rule, no investment is categorically permissible or impermissible).
The authors take the position that compliance with the rule is attainable by utilizing tools already available – which, not surprisingly, do not include artificial intelligence. Bank trust departments have long been subject to the prudent investor rule. The centerpiece of every Bank Trust compliance department is the use of an “investment policy statement.
So the question remains, is the use of artificial intelligence permitted within your existing Statement of Investment Policy?
