Chicago Booth Review Podcast Why AI Might Not Make You More Productive
- May 29, 2024
- CBR Podcast
Artificial intelligence promises to make us all more productive, so what role does old-fashioned management play? Chicago Booth’s Chad Syverson says management still matters, and even the best technology won’t necessarily improve productivity if the right managers aren’t in place. But what does good management even mean in the age of hybrid work and remote teams?
Chad Syverson: And the answer is the managers turn out to be hugely important. So just overall, on average within both of these companies, about a quarter of the variation in productivity across stores, and there's a lot of variation across stores, about a quarter of it is explainable by the manager of that store.
Hal Weitzman: One of the promises of AI is that we can all be more productive, meaning organizations can do more with less. Bill Gates has even predicted that AI might enable us to move to a three-day working week. So what role does old-fashioned management play in a world of artificial intelligence? Welcome to the Chicago Booth Review Podcast, where we bring you groundbreaking academic research in a clear and straightforward way. I'm Hal Weitzman, and today, I'm talking with Chicago Booth's Chad Syverson, an expert in productivity who isn't so sure that AI in itself will spell a big productivity boom. "Management still matters," says Syverson. "And even the best technology won't necessarily improve productivity if the right managers aren't in place." But what does good management even mean in the age of hybrid work and remote teams? Chad Syverson, welcome back to the Chicago Booth Review Podcast.
Chad Syverson: Thank you. It's nice to be here.
Hal Weitzman: I'm delighted you're back because last time we had a great conversation about productivity and you mentioned something last time about intangibles. So intriguing, and I know you have some research about management, good old-fashioned management, even in the age of AI and robotics and automation, a good old-fashioned manager can make a huge difference. Tell us about what you found.
Chad Syverson: Sure. Yeah, just generally, you could think of managers are like conductors of input orchestra, and that orchestra can sound great if they're conducted well or it can be another coffinous mess if they're not conducted well. That's what the manager does and that's how they affect productivity and give them the same inputs. A good manager will get more output out of those inputs than a bad manager. So I looked at this issue with some co-authors, Rob Metcalf and Alex Solacci inside two big retail companies, and both of these retailers are huge companies with hundreds of stores, and we had data on the productivity of every one of their stores as well as the identity of the managers at any given time.
Hal Weitzman: Productivity of stores, meaning how much stuff they sell?
Chad Syverson: Yeah. So sales per employee, you can adjust it for sales per employee square foot or whatever you want. You're going to get the same results I'll talk about, which is the core results are about sales per worker is the way to think about it. And what we did is we econometrically decomposed that productivity into two effects. One was something tied to the store itself, the permanent element of the store. So that could be, it's in a really good location, for example, probably a big factor. And then we separated that from the effect of the manager. And to do that, you actually have to have managers move across stores or else, the same managers that was in the same store, you can't do the decomposition. Fortunately, for both these companies, there's a lot of movements over time. We had a long enough data period where we could see movements and so we could do this decomposition.
And the point was to basically ask how important managers actually are. Is it all about the stores, the location? Etc, etc. Or do managers actually have some effect on the productivity of the stores themselves? And the answer is the managers turned out to be hugely important. So just overall, on average, within both of these companies, about a quarter of the variation in productivity across doors and there's a lot of variation across stores, about a quarter of it is explainable by the manager of that store. How much does that variation imply? One way to think about it is going from the bottom end, not the bottom bottom, but towards the bottom end of manager productivity effect to near the top end. That's equivalent to adding a fifth worker to a team of four. So it's like the same sales-
Hal Weitzman: Quartet to a quintet?
Chad Syverson: Yeah, sure. Exactly. To stick with the music analogy. That's right. Yeah. So they really have an effect on the bottom line. That's the way to think about it, that you can take a bad manager, replace them with a good manager, and you're going to see a substantial increase in sales for a given store. And then we went to try to figure out, okay, now that we know managers matter, one, how do the companies seem to deal with the fact that they have managers of different types, different talents? And two, what is it about these managers that seem to explain the variation in whether they're good or bad? What we found is that the companies tend to put their best managers in their lowest performing stores.
Now, that's a little surprising to us for the reason of, if good stores and good managers put together, create more productive stores, what you actually want to do, if you just want to maximize sales in the company, is you take your best managers and put them in the best stores, the best locations. That's the way to maximize sales. So they're giving, and we do calculation, they're giving up three to 6% in revenue by doing that. On the other hand, they are getting something else by doing that. If store closures are costly for some other reason other than you just lose the revenue, it looks bad, whatever. There's a number of things you could imagine why retailers wouldn't want to be closing stores. Then you might try to rescue marginal stores with better managers, and that seems to be what they're doing. So that's one thing about what the companies were doing vis-a-vis this variation in the quality of managers and-
Hal Weitzman: So just to understand that, so the managers who are in the better performing stores are not the best managers?
Chad Syverson: On average, that's true, yes.
Hal Weitzman: But are they lifted by the fact that the stores themselves are really good or in a good location or have good teams or whatever?
Chad Syverson: Well, the overall store performance doesn't look bad because they have a good location. Yeah. Now, their contribution to that overall performance is not so hot. But you wouldn't look at that store and say, "That's a failing store," because they have a bad manager. And that's why we have to do this econometrics. You can't just look at the total sales per employee of the store, no, that's because of the manager. There are other things going on and we have to separate that, separate that out.
Hal Weitzman: Okay. So this whole thing is one of the intangibles that we talked about, right? Management really makes a difference.
Chad Syverson: That's right. Management is an intangible. It's one of the things that affect the productivity of all the inputs you do measure. Okay? So it's an intangible, and as we would say as economists, it's a complementary intangible to tangible inputs.
Hal Weitzman: And just remind us what is tangible and what is intangible? What does that actually mean?
Chad Syverson: Intangible just means when we go measure the inputs and outputs of a company, we're not measuring those things. And we're not measuring because we don't think they're important. We're not measuring them because they're really hard to measure. So we can see that a manager is there, but unless we have a massive amount of data like we did for these two companies here usually can't see how good that manager is. So the manager quality is an intangible. Or if you think about something like the [foreign language 00:08:16] among employees or the knowledge base of those employees or the customer list that the sales department has, knowing that these people really like this thing and that client really doesn't like that. Those are all intangible inputs that help the company make sales and make stuff, but they're not in the typical inputs that we measure that they use to make this stuff-
Hal Weitzman: And you just measure on sale? So would a good manager have people more likely to stay? To go back to our earlier conversation, is the churn rate lower for good managers?
Chad Syverson: It depends on the setting. We define productivity as about output. Sometimes those other things are means to that end. So the amount of churn you have, turnover, et cetera, affects your output to input ratio. So it can affect productivity. Often those things tend to be correlated. So an example here too is even though we define productivity as sales per worker, we found that managers who are really good at raising sales per worker are also really good at running energy efficient stores, for example. So that's like a correlated outcome measure. It's not about output per se, but managers who are good at getting productivity traditionally measured, often are really good at getting other things that you might think are positive for the company.
Hal Weitzman: Now, so you looked at retailers. Do you think that the same dynamic applies in let's say manufacturing where there's a lot of automation and technology?
Chad Syverson: Yes, we actually know that. So there's been some of the earliest work was done by Nick Bloom, John Van Reenen, and then the team they've created to run what's called the World Management Survey. They started collecting systematic data on managers and outcomes and correlating management practices, at least with outcomes. So they definitely establish management matters. I think an issue that researchers have been trying to deal with lately is, okay, we know these practices matter. Is it about the people, the managers themselves who implement the practices? Or do you have to have a special manager to implement those good practices? Or is it just, as long as you do these practices? It doesn't matter who the body actually is in the management chair? I think our results are pushing towards the manager, the person, also matters because we're holding constant, since we're looking inside these companies and just inside these companies, we're holding constant a whole bunch of corporate practices.
So we are focusing more on the people than the practices, and we're still finding the people matter a lot. But that aside, it's certainly generally true that management matters well beyond retail. We definitely know in all sorts of sectors, healthcare, manufacturing you mentioned, education, even in schools, the principles matter a lot, for example. Probably in construction too, as far as we know, and of course, as we were talking about last time, construction's got its own productivity issues. But everywhere folks have looked and really, at any level of development of economy, whether you're talking about very wealthy economies or still developing economies, management seems to matter and matter a lot for productivity.
Hal Weitzman: Yeah. We did talk about construction a lot last time and how construction's actually declined in its productivity over the past 50 years. Is it possible that management played a big... I know you don't have the evidence, but could you speculate that management played a big role in that?
Chad Syverson: I'm going to step away from the data that's in the paper because we didn't have this kind of data. But I think there's a couple things going on that is special about construction, and one of which I think is management related, and that is, in a lot of settings in that sector, disruptions to the process of production aren't viewed as costly or a sign of things gone wrong. They're looked at as profit-taking opportunities. This is when you do a change order and we can mark up the change order or that sort of thing. Or we're going to have to have a redesign and we can redesign things in a way that the original bit releases us from the original bit, which might've been very competitive and now, we don't want it to be so much. That sort of, I guess I'd call that a lack of an incentive for well-managed projects I think, is a plausible explanation for part of the sector's broader issues. Yeah.
Hal Weitzman: Have you ever wondered what goes on inside a black hole or why time only moves in one direction or what is really so weird about quantum mechanics? You should listen to Why This Universe? On this podcast, you'll hear about the strangest and most interesting ideas in physics broken down by physicists, Dan Hooper and Shalma Wegsman. If you want to learn about our universe from the quantum to the cosmic, you won't want to miss Why This Universe? part of the University of Chicago Podcast network.
You're bringing a lot of trauma into my remembering previous construction projects. So I'm going to move on for that. So last time we talked a lot about AI, so I'm just thinking like a company that's thinking about where to invest. First of all, I guess the question is, if I invest in technology, is it going to happen if the managers can't make it happen? I can invest a lot in automation or other AI-type technology, machine learning, but if I haven't got the managers to run it, is it not really going to make any difference to my productivity?
Chad Syverson: I think that's right. I think that goes back to the conductor. Is it going to be a cacophony or a symphony? You can give the orchestra the tools, but without someone or a set of folks who know what they're doing to coordinate the use of those tools, you're not going to get the gains from them. I think that's definitely, definitely true. AI is, I think, no different than any other technology. Without some thoughtfulness and talent in implementing its application, you're not going to get the gains from it.
Hal Weitzman: So take the example of a conductor. A conductor's great with an orchestra. I don't know what a conductor's like with a construction project, but a good manager, even a good general contractor, should be able to take of those skills, execution skills, project management skills, and translate them to a quite different environment. Would the same managers who are in place now and have been learning their management education at places like Chicago Booth is that without wanting to make an advert for Chicago Booth, but are they going to be able to manage an AI revolution where work looks very different?
Chad Syverson: Well, booth managers are always going to do well.
Hal Weitzman: Of course, of course. I knew you were going to say that.
Chad Syverson: That goes without saying, yes, of course. I don't know if AI is so different that the old things that made a manager good no longer apply. I think a lot of what makes a manager good is just systematic and systemic thinking. Being good at picking talent, being good at managing talent. I think that's going to be true for AI as much as any other past technology. Sometimes, technical acumen and technical know-how helps a manager apply technology, so I wouldn't dismiss that entirely, but I think that's just part of the story. I think you can still manage the application of new technologies pretty well, even if you're not at the frontier. In terms of your own understanding of the nuts and bolts underneath it.
Hal Weitzman: And I talked about some of the execution skills. Do you have a sense of, I don't know if this was beyond the evidence that you collected, but what actually makes managers efficient and effective?
Chad Syverson: So in our retail setting, I got to be honest, we had a hard time finding anything in our data that predicted the good managers from the poor ones. Now, we didn't have a lot. We knew their names. We could infer their gender. We knew how long they were with the company and some basic stuff like that. We didn't have their work histories before the company, their education histories. We didn't have a personality profile. I would bet if we had some of that, it might predict more of the variation than we were able to with the few things we did, but we couldn't. I can't tell you anything definitive from our setting for, ah, that's the real secret. But those things I talked about earlier, picking talent, managing talent, those are probably in metrics, we just didn't have access to. I would love to, but we didn't have it in the-
Hal Weitzman: But it does suggest that if you have a team and you know that some people are higher performing, then they will be typically be able to translate those skills across other teams and other units.
Chad Syverson: Yeah. No, I think what you're... Is management a general technology? If I'm a really good manager in higher education, does that mean I'd be a terrible manager in healthcare or in manufacturing? No, I think there is an element of general skill to it because picking talent, managing people, coordinating, systemic thinking, that kind of thing. I think that's broadly applicable.
Hal Weitzman: Sorry. I was going to say, an example I think of is like football sports managers who have a career after their sports career of motivational speaking or whatever. But you're suggesting actually, it does make a difference. They might be effective.
Chad Syverson: Sure.
Hal Weitzman: I wanted to ask you about hybrid work because obviously, that's been very challenging for manager. The workforce has changed the way they want to work has changed. What do you think, what difference has that made to your effective managers?
Chad Syverson: I have a very Chicago view of hybrid work in the most traditional Chicago sense. So Ronald Coase, famous economist here, he had this theorem. The Coase Theorem says, well, if parties in a transaction can bargain, what we should expect to happen, regardless of who has the decision rights or the power, or however you want to think about it, there's an incentive to make the most efficient thing happen, that you can bargain over that and one party can compensate the other for the gains from that negotiation. So here's how I think that applies to remote work. One big question is ultimately, are people more productive at home or in an office? And folks are split on that. So some people say, "No, people are more efficient at home, or at least there's no sign they're less efficient. Plus then they get to save all this commuting costs. They like the flexibility," etc, etc.
Then there's the other side. "No, you lose control over everything. You lose one of the intangibles," which would be like the spirit of the company or the culture, something like that, or it's fine. Work at home is fine for doing routine things, but it's bad for innovation or it's good for older workers, but bad for newer... Okay. And I get these arguments on both sides and there's evidence for each part of that. The way I view it is, well, regardless of what it is, the Coase Theorem says ultimately, we've got this thing to bargain with, which we call the wage. And either side can use that to make the most efficient thing happen, happen.
Okay, so let's suppose that for some particular company, it's just the fact that the company is more productive if people are in the office. Okay? Well, now that there's an expectation in a lot of industries of work from home time, at least partially if not fully, what are you going to have to do to get people in the office? You're going to have to pay them more. You're going to have to compensate them for coming in now, to lose their flexibility to have to commute. But you would be willing to, if you're more productive, when they're in, right? You've got that extra gains from the higher productivity, you'll share some of that with the workers to compensate them for leaving hybrid work.
On the other hand, if some other company that's not true, that the productivity gains of being in the office aren't very high, then the wage doesn't go up because it's not worth the company to pay the workers more to come in. The workers get to stay at home, which they like. They're not paid more to stay at home. They're paid the same as before, but they get what's ultimately best for them and really, ultimately, best for the pair of the company and the workers because the company wouldn't be better off enough had they been in to compensate the workers for.
Hal Weitzman: So this is an at work premium?
Chad Syverson: Yeah. So what I imagine will happen in the long run is the companies where being in the office does actually show up in terms of real productivity boost, they're going to have people in the office and those workers are going to be compensated more for that.
Hal Weitzman: But to go back to the issue of management, a good manager, an effective manager will be effective hybrid, completely remote in person?
Chad Syverson: Probably yes. There might be some workers who are a little better in person than remote and vice versa. But again, I think it's largely general skill. But if you talk about zooming out to what I was talking about before, one of management's jobs is to figure out which is true. Are we actually more productive if we're in the office? We got to figure out whether that's true or not. And if so, what's the productivity gain? And then how much do we have to share with the workers to get them to come in and yield that productivity gain to the company? That's actually one of the things managers are going to have to do in this new world of hybrid work.
Hal Weitzman: Except for conductors, because I don't think you can really...
Chad Syverson: It's hard. I think the lag in Zoom is just not good for music.
Hal Weitzman: Right. They become video editors. It's right. Chad Syverson, thank you so much for coming back to the Chicago Booth Review Podcast.
Chad Syverson: You're welcome. It's a lot of fun.
Hal Weitzman: That's it for this episode of the Chicago Booth Review Podcast, part of the University of Chicago Podcast Network. For more research, analysis and insights, visit our website at chicagobooth.edu/review. When you're there, sign up for our weekly newsletter so you never miss the latest in business-focused academic research. This episode was produced by Josh Stunkel. If you enjoyed it, please subscribe and please do leave us a five-star review. Until next time, I'm Hal Weitzman. Thanks for listening.
The cost is likely minimal to achieve a fairer outcome.
Algorithms and AI Can Make Hiring More DiverseAn expert panel discusses the yet-unknown and steps policy makers should take.
How to Design Policy in the Age of AIConventional economic analysis of slavery overlooks the costs imposed upon the people who were enslaved.
Emancipation May Have Generated the Largest Economic Gains in US HistoryYour Privacy
We want to demonstrate our commitment to your privacy. Please review Chicago Booth's privacy notice, which provides information explaining how and why we collect particular information when you visit our website.