I’ve been thinking a lot about this question: if AI enables a 30%-3x uplift in investment research productivity (my base case…range depends on duration & coverage approach), what do analysts & PMs do with their excess time?
The knee jerk reaction is “I can process more ideas”, but I’m not so sure how I feel about that answer. If a PM of a 12-person shop asks each analyst for 1 great idea per month, and each analyst can now produce 2 great ideas per month - does the portfolio double in position count / does sizing get cut in half? More idea flow is better than less, but at some point there can be *too many* ideas, particularly on the long side (where the fight for capital is typically more fierce than the short book). Or does the holding period decrease / turnover of the portfolio increase 2x? For a PM/team with a successful track record, that seems ill advised and an unlikely response.
“Well, why does the team need to be 12, can’t it now be 6 if the analysts are 2x as productive?” Certainly, AI enables small teams to do much more with much less, and the productivity benefits of AI are transformational (TODAY!) for sub-scale and startup managers. But I doubt that successful, institutional teams will cleave large parts of their team. For a successful fund with institutional scale, i.e. >$1bn, the collective wisdom & judgment of the senior investment team is the core economic engine of the firm. Most often, these senior investors are specialist by asset type or sector. And, in high functioning firms, there is a camaraderie and friendship amongst this group. “Well, let’s shrink the analyst team”. That is certainly possible. At a long duration firm, an analyst could certainly more easily cover 80-100 names vs. 40-50 names. I hear anecdotes of that happening today. However, the ROI of cutting junior investment talent at a scaled firm isn’t clear…a few hundred grand is really a rounding error for a 2 & 20 firm running real scale.
So, to what end do these 30%-3x productivity benefits serve?
Well, as all investment research innovation has done for decades, whether it was the fax machine, Bloomberg, excel or alternative data, AI will, most likely, allow institutional investors to go even deeper by reallocating their time to more high impact due diligence.
I spent a lot of my first year in a hedge fund seat calling Burger King franchises and Cabela’s gun counters, meticulously spreading sell-side models into detailed consensus, and updating Nielsen scanner data every 4 weeks. As tools and software emerged to handle those workflows, I certainly didn’t work fewer hours! But I had time in my week to deploy to other research ends.
What are these workflows? My belief is that many teams will start to spend more time off the desk, as capturing the first 80% of the envelope of information on a company becomes much more efficient with AI. The multi-manager investment approach has become, in many ways, quite mechanical and regimented. Often by mandate, the process of updating models, regimented sell-side & IR calls, building thoughful earnings previews, navigating earnings, and trading high velocity data points requires investment professionals to be on the desk the majority of the time. As a multi-PM, other than the marquee conferences, I was on the desk…call it 85% of my working hours in a given year. If these workflows can be augmented/automated (say 60 earnings previews take me 1 hour each instead of 6 hours each), I free up a lot of time to attack the remaining 20% of the envelope which isn’t accessible sitting behind a desk. As a small cap generalist, I was off the desk much more…going to many more industry conferences, trade shows and bespoke HQ visits. As the first 80% of the envelope of information about a company becomes easier to access, the last 20%, which requires a lot more effort, a lot more travel, a lot more “connecting the dots” will become, in my opinion, the last frontier of informational edge in markets.
Brett