I recently participated in a conference on the growth in the use of Exchange Traded Funds (ETFs). A representative from one of the large mutual fund companies described ETFs as “just another fund class”, which may have been an attempt to reassure the portion of the audience that made a living selling mutual funds that ETFs could be simply added to their shelf of products, and life would go on much the same. My presentation took a different direction.
A key difference between ETFs and active mutual funds is that the traditional fund manager is replaced by an algorithm. For example, the iShares S&P/TSX 60 Index ETF tracks a basket of the top 60 Canadian companies ranked by their size, as measured by their stock market value. This is a simple algorithm; it can get more complicated by changing the weighting so, for example, smaller companies have a greater presence. Fund managers (and advisors) who previously made a living picking stocks (in this case, small company stocks) with the promise of outperformance have increasingly hit a wall of data that suggests that algorithms can not only produce better after-fee performance but that they have much lower tracking error. What is tracking error? Tracking error is a measure of the variability, or random noise, that arises when the actual result differs from a benchmark. ETFs that track an index have much lower tracking errors than actively managed mutual funds, which is a complicated way of saying that when you buy an index ETF it does what it says on the box.
Nobel prize winning psychologist, Daniel Kahneman, co-authored an article on the impact of noise in a variety of professions. Distinguishing noise from professional judgement is not just a problem for investors and their advisors, but also for software engineers estimating project costs, doctors making diagnoses and judges handing out sentences. Kahneman’s next book, Noise (due 2020), expands on this theme. The article suggests that simple algorithms can outperform human judgement because “professionals often make decisions that deviate significantly from those of their peers, from their own prior decisions, and from rules that they themselves claim to follow”.
Returning to the investment world, much of the value of investment advice is shifting from security selection to wealth management (from portfolio alpha to advisor alpha, using the industry jargon). Kahneman’s article suggests that the problem of tracking error, or noise, applies not just to portfolio alpha but also to advisor alpha. That is not to suggest that everything can, or should, be reduced to algorithms. Retirement planning done well, for example, involves a nuanced understanding of the client’s objectives and preferences for income stability versus the risk of income shortfall. In this situation, advisor appreciation of the client’s preferences needs to married with goal-based forecasts.
Vanguard’s Personal Advisor Services (PAS) provides one glimpse of the trends in the investment industry. As described in a recent article on applying Artificial Intelligence (AI), the PAS splits responsibilities between algorithms and advisors as summarised in the table below.
| Cognitive Technology (Algorithms) | Adviser |
|---|---|
| Generates a financial plan | Understands investment goals |
| Provides goals-based forecasting in real time | Customizes an implementation plan |
| Rebalances portfolio to target mix | Provides investment analysis and retirement planning |
| Minimizes taxes | Develops retirement income and Social Security drawdown strategies |
| Tracks aggregated assets in one place | Serves as a behavioral couch |
| Engages clients virtually | Monitors spending to encourage accountability |
As the article makes clear, success comes from not just replacing parts of existing processes with algorithms but re-designing the work flow to accommodate the best of what both algorithms and advisors have to offer.
This is good news for investors. Progress with algorithmic approaches in wealth management, coupled with advisor’s professional judgement, offers similar potential benefits to clients as the replacement of actively managed funds with ETFs and index funds: lower costs, more predictable outcomes (less noise) and greater accessibility.
