Will AI Adoption Actually Pay Off?

On this episode of The Long View, Michael Gates, lead portfolio manager for target-allocation models at BlackRock, discusses the current macroeconomic and market environment, the role of productivity in growth and profitability, and the potential impact of artificial intelligence on the labor market. Here are a few excerpts from Gates’ conversation with Morningstar’s Ben Johnson. AI‑Driven Productivity and the Market Outlook Ben Johnson: What are some of the key macroeconomic and market themes that you and your team are keeping close tabs on today? Michael Gates: We’re entering into a period of sustained productivity growth in the US especially and globally. A lot of that precedes AI’s penetration throughout the economy, but that kind of ability for AI to have a meaningful impact on productivity is starting to appear, and we think it’s going to become more important in coming years. So the economic backdrop has that positive aspect to it, which is a productivity-led growth impulse. The interesting thing to me about productivity-led growth is that when you get a productivity shock, it’s positive to real GDP, and it’s negative to inflation. So it’s a supply-side shock and has that dual benefit. So if we extend that forward, that’s a positive environment for stocks. And right now, it’s a time when, even with that, it’s not easy to invest. So, for instance, in fixed income at the moment, credit spreads are very tight. And for a fixed-income investment, you’re not going to get more than the yield to maturity in most instances. So the fact that the compensation over Treasuries is very low for credit-risk-bearing assets is something we’re taking note of recently. Still have a pretty positive outlook for stocks. I think the other thing right now to pay attention to, and this will come as no surprise, is what’s happening geopolitically, and that’s changing week to week. Why Results Vary Across Companies Adopting AI Johnson: I’m curious just to get your overall AI thesis and how it might be either different from or more nuanced than just what seems to be a prevailing “AI is going to eat everything and everyone and take all of our jobs” thesis and how that’s beginning to manifest in the portfolios. Gates: Why don’t we start at the end? Let’s just go right for that concern. And the answer to that is we don’t know, but let’s kind of build a mental construct here quickly. Imagine that you have an elasticity, a sensitivity of the labor market to GDP growth. So in the numerator, you have the monthly change in nonfarm payrolls, right? How many jobs net are created in the economy this month or in any month? And the denominator, you have the GDP growth rate for that month or every quarter. So, you have a per unit of GDP growth: How many jobs are created? So that’s the elasticity of the labor market to real growth. And imagine that that number, that elasticity, is reducing, right? So for every unit of GDP growth, the job intensity of that, the number of jobs created, is less than you would have expected if you were looking at the average over the last five years. If the productivity boom drives GDP up substantially, though, you can make up for, or more than make up for, that drop in the elasticity. So you can see a labor intensity in the economy reduced, but an overall level of jobs creation sustained under such a scenario. You can imagine such a breakeven. And I think that’s going to be a part of this. If you listen to leaders like Jensen Huang, who’s spoken about, at his company, how profits are recycled into new and exciting projects, you can imagine that writ large across corporate America, firms that are succeeding, adopting AI, and infusing AI into their operations, have higher profitability, and continue to see a need for labor, but the labor intensity per unit of revenue goes down. So the amount of revenue that we see per employee in the S&P 500 is rising, and I think that’s going to continue, but the amount of growth in the economy is going to determine the net of this in terms of the job market. Putting the AI Thesis to Work in Model Portfolios Johnson: Michael, on behalf of the entire global labor force, I want to thank you for one of the rosiest pictures of the future for employment and productivity, certainly that I’ve heard in quite some time. And I think in many ways what you’ve described is just congruent with our long experience with all forms of productivity growth more broadly. When you think about how specifically you make investments against this narrative in the AI sector, how that manifests specifically in the context of the model portfolios. Gates: A couple ways, just taking a look at models, one of the decisions we’ve made is to carve allocations away from the benchmark asset into thematic assets. So what do I mean by that? So if I’ve replicated my performance benchmark within US equities, my advisors would recover something like the Russell 1000 or the S&P 500. They’d get back a cap-weighted benchmark asset and nothing else. Instead, what they get is an equity portfolio that’s very similar to that large-cap benchmark, but differentiated. And one of the ways we’ve differentiated has been to have a carve-out position to technology. We had that up until, I think, 2022, then we got out for a while, and then came back in late 2023 or sometime in 2023. So that’s one level of active risk taking is to carve away from the core asset and overweight a specific sector in the market. Something we did last year was then to move that allocation to a specialist team. We’re using an active exposure for the tech position, and that active exposure is focused strictly on the AI theme. And that’s been a very accretive decision that actively managed AI-specific exposure has substantially outperformed that capitalization-weighted tech index, and both have, in turn, outperformed the funding asset, which is that US equity market benchmark. In the most recent repositioning that we did, I mentioned that we cut off the top in terms of taking out the top 100 stock exposure and moving to a broader exposure. We could have just done that with an index. And some of the money that came out of the largest stocks did go into index exposures in the US, but a portion of the proceeds went into an active US core manager that’s got, in the Morningstar data, the top one percentile of performance—not top decile—but top percentile performance in the trailing 12 months. And his investment process, his and his team’s investment process is very focused on earnings. And they also have a thematic overlay looking for firms within each sector. So we’re talking in addition to the IT and communication services sector, within each sector, looking to identify firms that are effectively embedding AI into their operations. And they’ve had a lot of success doing that. And there’s a number of anecdotes in the portfolio, and it’s a concentrated portfolio, less than 50 names, where the firms they’re identifying, competing in a given space, are growing their earnings and having very strong stock performance relative to firms that are not being very effective at adopting AI into their practices, and their earnings are not growing as quickly and their stock prices are languishing. So this is a dynamic that’s happening inside of the equity market, inside of all the sectors of the market, in addition to tech. And so that’s part of the broadening out that we’re seeking to capitalize upon.