Are Two C.E.O.s Better Than One?

Hey there, it’s Stephen Dubner, and I have some exciting news. If you consider yourself a superfan of Freakonomics Radio and you want even more than the weekly show — and if you’d like to hear the show without ads, there is now a way to do that. We’ve just launched a membership program called Freakonomics Radio Plus. As a member, you’ll get a weekly bonus episode of Freakonomics Radio every Friday. Plus, you can listen to Freakonomics Radio and every show in the Freakonomics Radio Network ad-free. To sign up for Freakonomics Radio Plus, visit the Freakonomics Radio show page on Apple Podcasts, or go to freakonomics.com/plus. There are already a couple member-only episodes waiting for you there. As for future bonus episodes: what do you want to hear? What kind of conversations do you think would be most valuable or exciting? Any and all feedback is welcome; send us an email, to radio@freakonomics.com. Meanwhile, if you want to keep hearing Freakonomics Radio exactly the way you’ve always heard it, all you have to do is nothing. Nothing is changing with this regular weekly show. But if you are the kind of person who wants even more — well, more is what we have. At Freakonomics Radio Plus. On Apple Podcasts, or at freakonomics.com/plus. As always, thank you for listening — however and whenever you listen. Now, as for this week’s episode …

*      *      *

People often ask me, “Where do you get the ideas for your show?” And I usually say something like, “Well, I don’t have a real job — this show is all I do — so I spend a lot of time reading, talking to people, wandering around, trying to sort out what’s interesting in the world, and what’s underexplored.” But occasionally a good idea just shows up in an email. Like the one we recently received from one Zach Levine of Tampa, Florida. “As an entrepreneur who has started and sold several successful tech startups,” he wrote, “I’ve always toyed with the idea of having a co-C.E.O. I shine at most C.E.O.-related things, but there are many I’m not great at. So, my question for you: do companies run by co-C.E.O.s perform better than those run by solo C.E.O.s?” Interesting question, Zack Levine!

Now this e-mail happened to land as we were producing last week’s episode, about how the U.S. leads the world in single-parent households, and why that’s a big problem. The economist we interviewed, Melissa Kearney, makes the argument that on average, married couples are able to give their kids significantly more opportunity than single parents are. So, if you think of parents as C.E.O.’s — in the business of raising a healthy, happy, successful kid — it looks like two are better than one. So, as Zach Levine asked: why wouldn’t this also be true in the business world?

Today on Freakonomics Radio, let’s find out! We will hear from one C.E.O. expert who thinks co-C.E.O.s are a great idea, and one who thinks it’s absurd. We’ll hear directly from some co-C.E.O.’s — one pair that is happily running a big company, and one from a huge company that went down in flames. We’ll also hear about some research which found that working in pairs makes people not only more productive but happier. And who doesn’t want to be happier?

*      *      *

Okay, let’s start with this man. Name and I.D., please.

Marc FEIGEN: Marc Feigen, I am a C.E.O. advisor.

And what does a C.E.O. advisor do?

FEIGEN: Good question. So, we have a firm that supports the C.E.O. with everything they need to be successful. They get a coach, they get governance advice, and investor-relations advice.

Fortune magazine has called Feigen the “C.E.O. whisperer.” People come to him with problems, with questions. Not long ago, there was a question about whether two C.E.O.s might be better than one.

FEIGEN: It was in a conversation with a client who was curious about it. And I had probably the same skepticism that anyone else would have, which is that, “Hey, it’s probably difficult to make work.” I didn’t know offhand of any co-C.E.O.s. And I said, “I’ll look at it.” We looked at 2,200 companies from 1996 to 2020 — the S&P 1200 and the Russell 1000 — and there were just 95. 

That’s right: in the case of these 2,200 large, publicly traded companies over more than 20 years — so that’s a lot more than just 2,200 C.E.O.’s — there were just 95 instances of co-C.E.O.’s. Some of them ran firms you may know: Chipotle, S.A.P., and Research in Motion, or RIM, the company that made the BlackBerry, the first big smartphone. But, again, co-C.E.O.s were unusual.

FEIGEN: Certainly this is rare in the public markets. Private companies, though, have co-C.E.O.s everywhere you look. If you go to a bakery or a pub, ask if there’s a co-C.E.O. in charge — “Well, no, we don’t have co-C.E.O.s. But yeah, my brother and I, we run the place.”

Feigen decided to focus his research not on private firms, but on big public companies — largely because there’s a lot of data that makes it possible to measure their performance. Once Feigen and his team had done this measurement, they wrote up their findings in a Harvard Business Review article called “Is It Time to Consider Co-C.E.O.s?” How did Feigen go from skeptic to believer? It turned out that the C.E.O. pairs in his data delivered annual shareholder returns that were nearly 40 percent higher than the returns of the thousands of firms run by solo C.E.O.’s.

FEIGEN: If you were a hedge fund, and I said, “Hey, I can give you a certainty of a 40 percent better total shareholder return,” you’d be the biggest hedge fund in the world.

So that’s why Marc Feigen became a true believer in the co-C.E.O. Now, to be clear, Feigen is not an academic economist, and his study didn’t have anywhere near the sort of robustness checks that a peer-reviewed research paper would have. He didn’t control for a variety of other factors that may have explained the big difference in performance. It could be that the kind of company that has two C.E.O.’s has some other characteristics that produce better returns. Still, Feigen was impressed by this finding. As he wrote in that Harvard Business Review piece, “Today the job of running a company has become so complex and multifaceted, and the scope of responsibilities so great, that the co-C.E.O. model deserves a fresh and close look.” Some readers, however, were less impressed.

FEIGEN: People say, “Well if co-C.E.O.s are such a good idea, why don’t they exist in countries? Why aren’t there co-presidents?” Well, guess what? Rome, which was pretty successful, you have to agree, had co-consuls. The Senate said, “We’re not giving this job to one person. We’re going to have co-consuls.” And those co-consuls were responsible for administration, for the military. But they had arrangements. Each could veto the other’s decision, so they had to agree. So, in fact, they were partners.

DUBNER: I appreciate this reference, but the fact that you have to go back a couple thousand years to point to a prominent example makes me think that there’s something in either human nature or in the setup of large organizations that conspires against pairs. What is that? Why is that? 

FEIGEN: Well, I think that as countries became more militaristic — command-and-control, the top-down leader, the male figure is very much in charge. That’s how countries grew. And that’s how companies grew. That’s the history of management. But don’t you think that’s a bit dated? We talk about team-based organizations. Why not have a team at the top? I don’t argue it should be used everywhere. By no means. It’s a choice, as opposed to a prescription.

DUBNER: If we were to ask, Marc, the C.E.O.s of, let’s say, the top 100 public firms in the U.S., how they would feel about having a co-C.E.O., how many of those 100 do you think would say yes, or at least maybe? 

FEIGEN: I would expect most would say no. I think the better question is, “Would you imagine handing your company to successors who are co-C.E.O.s?” Because I’ve asked them. Far more than you think have listened to what we’re saying and have said, “I’m open to it. I am skeptical.” And by the way, I think it’s important to be skeptical because, I think there are reasons why this shouldn’t work, and you have to do it carefully. But more and more companies have said, “I think this makes sense. Let’s discuss it. And what are the best practices?” And remember, we’re not dividing the role in half. We’re doubling the capacity — in the most important job in a company.

DUBNER: You said that co-C.E.O.s are much more common among private firms than public firms. Why is that? Do boards behave differently with public firms? Is it the media scrutiny? Is it shareholders?

FEIGEN: No. It’s because most private companies are small. And when they’re founded, two people get together and found a company. My mother and her best friend from grade school, Lois and Ellen, founded a travel business, and they were partners until Ellen sadly passed. That’s the norm.

DUBNER: What is it about two people collaborating that is special? 

FEIGEN: I think the most important thing is they coach each other. Co-C.E.O.s — they have a partner, they have a coach, they have a collaborator they can bounce an idea off. You know, when you go rogue and come up with a kooky idea, like we all do, you have someone who can ground you. One of the co-C.E.O.s said to me in our interviews, “The best thing about this is we can call bulls*** on the other.” That’s important. And I think when sole C.E.O.s fail — there’s a set of reasons why C.E.O.s fail — but often, they get too aggressive with an idea. 

DUBNER: “Aggressive” or “invested” because of ego? 

FEIGEN: Oh, I think that’s well said. You’re right, it becomes wrapped up in their identity. They’re certain they’re right. After all, they got to where they got to because they were right more often. And they were often right when other people told them, “Don’t do that.” And then they did it, and it worked. And now they’re promoted, and they’re in the big job, and so, you know what, “I know how to do this, and I’m going to tell you what to do.” I’m thinking of Ramani Ayer at The Hartford, which is a company that was founded 150 years ago — Hartford Fire Insurance Company. It was a great company. Ramani built it up in the early 2000s by selling variable annuities, which is a type of financial product where Hartford took interest-rate risk. And then came the Great Recession, and the stock went from $70 to $3. Ramani became in love with an idea. But Ramani had no one to check him. And I can think of many examples like that. And so, when you have a partner, you might take a bit less risk. So, the model may be more cautious, but when you do take risk, both are aligned. You’ve really thought it through.

Now, to be fair, nearly every C.E.O. has at least a few allies, either internally or on the board, to help them think through every big move. But Feigen’s argument is that unless you are technically, officially in a power-sharing role, the power really doesn’t get shared. To think about power-sharing, consider how a lot of families work:

FEIGEN: We think of the strong father figure. Well, tell that to my mother or your mother, right? Most of us, if you’re lucky, came from a home with a mother and a father. So, you had co-C.E.O.s. So, we’re used to it in human nature. I think co-C.E.O.s suffer by disproof by example because anyone can point to a disaster. And you can imagine when co-C.E.O.s unwind, I mean, they can unwind horribly. They become distrusting of each other, camps develop. They are indecisive. They spread rumors about the other. I mean, it can get ugly. 

DUBNER: Can you give an example of that? 

FEIGEN: Yes, Chipotle. The founder, Steve Ells, brought his best friend, Monty Moran. And when they went public in 2006, they became co-C.E.O.s. And in 2015, E. coli was found in their food, and people were sick. It was a crisis, and they fought, and they never spoke to each other again. At BlackBerry, as the iPhone and Google came out, they were on the defense. Mike Lazaridis and the co-C.E.O. Jim Balsillie were very close. But when the company was up against the wall, they didn’t perform, and their partnership fell apart. 

Feigen’s story about the Blackberry got my attention. Some of you may not remember just how big the BlackBerry smartphone was, but — it was big. It was made by a company called Research in Motion, or RIM, which was founded in 1984 in Waterloo, Canada. They developed a variety of wireless technologies, but it was the BlackBerry — with its addictive raised keyboard — that became the breakout superstar. In 2007, RIM became the most valuable company in Canada, with a market capitalization of $67 billion. Its co-C.E.O.s — Mike Lazaridis and Jim Balsillie — were hailed as not only visionaries and tech wizards but as the ideal corporate couple. We reached out to both of them for an interview; one declined, but the other did take our call.

Jim BALSILLIE: Sure. My name is Jim Balsillie. I’m the retired chairman and co-C.E.O. of Research In Motion. It’s a pleasure to be with you today.

I asked Balsillie to describe how he and Lazaridis got together in the first place.

BALSILLIE: Mike and I, we got to know each other because he was a supplier in 1989 to a company I was working for, and I was blown away with his engineering and his innovation-computing skills. So we became partners in 1992. We agreed to share the C.E.O. because we had distinctive skills. I did the commercialization and the financing side, and Mike did the engineering and product side. But we talked every day, multiple times a day. I can’t think of a day in 20 years we didn’t talk, and exchange messages. 

DUBNER: Including weekends? 

BALSILLIE: Every day of the year. Tech is — there are no holidays in tech. And so, we were that typical ten-year overnight success story where we beavered away for a while and then we had this convergence of mobility, and messaging, and the internet protocol. We took it from an idea and 11 years later, we were the fastest-growing company in the world.

DUBNER: How did the relationship evolve as the success got so intense? 

BALSILLIE: Well, Mike and I shared an office for the first couple of years. We literally talked before meetings, debriefed after meetings, we had the password for each other’s voicemail. When we traveled, we shared hotel rooms. It was incredibly intimate — and a true partnership. We actually got our own office at one point, but they were beside each other with an assistant, Lisa, and we had an angry C.E.O. that was wanting something we couldn’t provide at that time. And he would call, and when I heard his name, I crawled under my desk and said to Lisa, “I’m not here, but Mike’s in his office.” And Mike crawled under his desk, giggling, saying, “There’s no one in here but us chickens.” And Lisa’s got her hands on her hips saying, “Boys!” So we had to flip a coin to decide who was going to take the difficult call, which shows that you got to work together, you got to have some fun. It’s not always clear who gets the good stuff, and who gets the bad. 

But technology, as we all know, moves fast. Being on top is no guarantee you’ll stay there.

BALSILLIE: What happened was the company was facing two shifts in the competitive terrain. And, if you know the old adage, “Who wins a battle between an alligator and a bear? It depends on the terrain.”

Okay, I’ll be honest, I did not know that old adage. Maybe it’s a Canadian thing? Anyway …

BALSILLIE: And in parallel — and it’s very important to understand that these happened in parallel — Apple came out with a high-end phone accompanied by a rich e-commerce ecosystem, but also Google came up with a subsidized business model that enabled much cheaper handsets in exchange for their data. We lived in the mid-market of phones, and so it plummeted the price structure. And the carriers were telling me that. They’re saying, “Sell your stuff way below cost or, I’m sorry, it’s too expensive now.” And so, I believed that the phone business was dead, but we had this amazing services business that provided the rich messaging that everybody liked. Mike wanted to double down on hardware. I called it a suicide march, and he considered my view heresy. And so, it was a strategic difference. It was an impasse. It went to the board. The board picked Mike’s path to go hardware, and jettison the services business — and that was the end of the relationship.

DUBNER: So what was your relationship like in those last months? Was it you and Mike cursing and screaming at each other? Was it more silent treatment?

BALSILLIE: Well, Mike and I never raised our voices to one another, ever, in 20 years. But he thought I was, like, off my rockers. He thought, like, “Are you crazy? I mean, that’ll kill hardware.” And I’m like, “Mike, the baby’s dead.”

DUBNER: So if I’m a skeptic of the co-C.E.O. model, I might say, “Well, sure, it works fine in happy days, in growth mode, but once you’re under serious threat or chaos, it’s a problem. You need a captain of the ship.” Would you agree with that? 

BALSILLIE: No. I think the board should have said, “We got to make you guys duke this out. And, you know what? We’re not going to pick a side.”

But instead of having the two C.E.O.s duke it out, the board did side with Lazaridis. Balsillie left the firm in 2012; Lazaridis wound up leaving a year later. Between 2011 and 2016 — the years you may remember for the rise of Android phones and the iPhone — the company behind the Blackberry saw its sales fall from $20 billion a year to $2 billion.

BALSILLIE: And so, I think had we had the co-C.E.O.s and the board said, “You guys brought us here and you’re going to figure out going forward and we’re not going to let you do anything but invest in it unless one of you says ‘uncle.’” 

DUBNER: Did Mike ever come to you later and say, “Hey, Jim, I think you were right”?

BALSILLIE: Um, no. No, no, no. That didn’t happen.

DUBNER: What has your relationship been since you left the firm? 

BALSILLIE: Oh, I’ve seen him socially a couple of times at events and all that. But we mix in very different circles, so it’s all just pleasantries.

DUBNER: So we recently spoke with this C.E.O. consultant, Marc Feigen, whose research found that public companies with co-C.E.O.s perform better than companies with more traditional, single leadership. Empirically, it’s not a super-robust finding, but it’s not nothing, either. So, let’s say we trust the data and we trust the finding. Why would you — I mean, you’ve got better experience in this realm than just about anybody could have — why would you expect that would be?

BALSILLIE: Oh, I think for the reasons I’ve told you and many more. I read Marc’s piece recently. I thought he’s spot on. I’m reading it, and I’m thinking, “Yes, yes, yes.”

DUBNER: Let me run a few by you and maybe you can just give me examples from your partnership. He said that among the many benefits of two C.E.O.s — I’ve got a list of maybe seven or eight here — one is that, “They can be in two places at once, literally.”

BALSILLIE: Gold. Gold. Like, I lived on planes. I was tired, chronically tired. I loved it, it was exhilarating — but I had no chance to be in the places Mike had to be. And I don’t know how we could have done it without the two of us, because there was just too many times we had to be in two places at the same time. When you’re growing any company, people want the C.E.O. So, when we had all these suppliers, and these were big partnerships, they got Mike as a C.E.O. And when those in the commercialization and the carriers, they want to see the C.E.O. because if you show up as a C.E.O., it’s respect. If you don’t show up, it’s disrespect.

DUBNER: Here’s another argument from Marc. Retention. Co-C.E.O.s can ensure the retention of two C.E.O. candidates. So again, in your case, it happened very, very early. It wasn’t like they had to offer you the job so that you would stay. But I’m curious if you have seen that dynamic elsewhere. 

BALSILLIE: Well, I would extend that view a little bit to key executives — people want a relationship with the C.E.O. And the kind of magnetism that Mike brought for a lot of the people — he was their leader, and their leader was the C.E.O. And in the financing and the commercialization part of the business, I was their leader. So I would say the retention is much more on the whole executive suite, right underneath the C.E.O.s, because you have two force fields of supporting executive relationships where they all say, “I report to the C.E.O., and I believe in the leadership of the C.E.O.”

DUBNER: So to me, the ultimate pair partnership is marriage. And if the institution of marriage has generally succeeded for so long, do you think there’s anything we have to learn from operating in pair partnership that might be applied to business?

BALSILLIE: Many of our senior executives used to say, about Mike and I, they’d say, “I wish my spouse and I had as good a relationship as Mike and Jim.” So I think it’s — you feel you’re better because of the other. I think good relationships happen because each side thinks they got the better side of the deal. But together it’s one and one is three, and you’re aware enough to know it, and invest in it, and nurture it.

DUBNER: I have to say, considering how things ended with you and Mike and with RIM, it sounds to me, at least in this conversation, that there is zero bitterness on your behalf. I sometimes hear that same lack of bitterness in certain divorced couples, although many are not that way. Do you feel frustration, regret, hurt — any of those things and you’re good at covering it? Or do you really not feel those negative emotions?

BALSILLIE: No, I don’t feel the negative emotions. I mean, I prospered mightily. I changed the world. You know, being a C.E.O. is so demanding. You eat well and you prosper, but it owns you. I was very, very tired when I stepped down. I just couldn’t believe how much I slept for almost a year. Certain times, I think what would have happened had they done what I thought was the right thing. But then I think, you know, the privilege would have been flying around the world for another 10 years. And obviously making a lot more money than you’ve already got — though you’ve already got more than you can ever imagine. So, no, no, no. It’s not, not bitter at all.

Coming up: a current C.E.O. pair tells us what works for them, and why. And then: let’s throw some cold water on the whole co-C.E.O. idea, with help from The Beatles.

*      *      *

The C.E.O. consultant Marc Feigen has become an evangelist for co-C.E.O.s.

FEIGEN: If you look back at the early years of Goldman, John Whitehead and John Weinberg decided to become co-C.E.O.s and run the firm together. They went to McKinsey, and Marvin Bower — who ran McKinsey and built McKinsey — said, “It’ll never work, but try it, because when you screw it up, you’ll need my firm even more.” And guess what? It worked for 25 years. So, the skepticism is natural. I understand it. It would be silly to say, “Oh, this always works. It always makes sense.” Of course, it doesn’t. It’s a choice. And it’s worth having a serious look at, and not discounting because of skepticism. It’s worth discounting because you don’t think it’s right for your company. Fine. 

In the Harvard Business Review article he wrote promoting the idea, Feigen lists a number of conditions that help predict whether co-C.E.O.s will succeed. The first, and probably the most important, is that they are both willing participants in the power-sharing.

FEIGEN: This is why founder co-C.E.O.s work so well. They’re super willing. You need to sit down and say, “Okay, we’re friends now. We’re willing. But we’re still going to disagree. And disagreement is healthy. But how are we going to resolve those disagreements?” When mom and dad are fighting, it’s not fun.

Mike CANNON-BROOKES: Our legal shareholders’ agreement actually came down to, if we cannot agree, we will end up in a game of scissors, paper, rock, and that will make the decision.

That is Mike Cannon-Brookes. He’s a co-founder and co-C.E.O. of Atlassian, a publicly traded Australian firm with a market cap of around $50 billion. They make collaboration and productivity software, and their customers include over 75 percent of the companies in the Fortune 500. Cannon-Brookes started Atlassian in 2002 with this man, Scott Farquhar:

Scott FARQUHAR: Mike and I met in a co-op program, which meant that we were with a whole bunch of students who, during their university career, got to do different internships. That was a sort of an uninspiring experience to work in technology in Australia for very large, three-letter acronym companies — banks and consulting companies, and so forth — and when we graduated out of that, Mike sent an email around to many different people asking, “How about we start something ourselves?”

In the beginning, they didn’t call themselves co-C.E.O.s.

CANNON-BROOKES: There were just two of us doing every single job, and you just did whatever needed to be done, right? Whether it was writing some code, helping a customer, trying to write some marketing copy, filling the coffee machine, whatever it was. And I think it was a fair few years in when we actually probably even thought, “Oh, we should put a title down on a piece of paper as to what we are.” And at that point, it was pretty obvious that we were co-C.E.O.s.

And how do they divide their work today?

FARQUHAR: Mike does most of the work and I just try and sit on a beach and drink cocktails and do podcasts. The way we split the job up at the moment is, Mike runs all of product and engineering and design, and I run the go-to-market function, so sort of sales and marketing.

CANNON-BROOKES: At different times over the 20 years, we’ve both done the whole job. So there’s a lot of benefits of the co-C.E.O. model that we see. But certainly, one of them is being able to take a break, or change your responsibilities, or get some sort of a refresh. 

FARQUHAR: I got to do my honeymoon and get married and take some time off because Mike could run the business. There are people I know, our peers in publicly traded companies, who have never taken more than a week’s holiday in a decade. There are people I know who’ve never taken more than 24 hours away from being in contact with the business. And that’s because when you have tens of thousands of people, millions of customers, things can go wrong at any point in time. And ultimately, you’re the custodian of that company for shareholders, for staff — everyone looks up to you, and you can’t really abdicate that responsibility. And it’s always come down to, if I can’t convince Mike that something’s a good idea, then there’s probably really good reason for that. And I’d like to think vice versa.

CANNON-BROOKES: I do think it’s one of our superpowers as a company. Maybe each of us wouldn’t make a very good C.E.O. by ourselves. Maybe we make three-quarters of a C.E.O. But if you have two three-quarter C.E.O.s, you’re well ahead of a singular C.E.O. I’ve often said that Scott and I are compatible, we are overlapping — but we are not identical. And this is really critical. We have a compatibility of skill sets. We can communicate well. We have a similar set of values around openness, honesty, respect, et cetera. But we do not have identical skills. If we were exactly the same person, you’d kind of be like, “Well, one of us is redundant.” If we have vastly divergent people, then we’d be fighting all the time.

Cannon-Brookes and Farquhar make it sound as if having a co-C.E.O. is a no-brainer, as long as the fit is right. So why do so few big public companies try it?

CANNON-BROOKES: It’s not for everybody. I’m often asked, “Oh, do you think every business should have a co-C.E.O.?” And I say, “No.” I think it’s a very good structure. That doesn’t mean every business can make it work. Nor do I believe every individual can make it work.

FARQUHAR: And the second thing I think is, you know, ego runs against it as well. It’s very appealing to be the sole decision-maker and the sole boss who gets to make all the decisions.

CANNON-BROOKES: 2008/2009 in the global financial crisis was the first really rocky period we hit. And you know, there was a lot of panic around. That is exactly the time when you see how good a co-C.E.O. relationship is. There was no sniping at each other. There was no upsetness. There was another moment where we get in a room, we work out what we need to do. We close the door, we try to honestly assess the situation. We made some pretty strong decisions. You know, we froze pay rises. We stopped hiring for a while. We did a whole lot of very quick decision-ing together. But that’s where you want two people in the boat to help navigate you out of tricky waters, actually. If you have the right two people. It was not tense or friction-filled. It was quite the opposite. 

But what happens when co-C.E.O.’s can’t work their way out of a jam? It happened with BlackBerry. More recently, it happened with Allbirds, the eco-friendly sneaker brand. Like Atlassian, Allbirds had two co-founders who were also co-C.E.O.s. They went public in late 2021, and for a time they were phenomenally successful. But it didn’t last: in an article headlined “How Allbirds Lost Its Way,” the Wall Street Journal reported that as sales began to fall, the two C.E.O.s had strategic differences. Their stock price fell by 95 percent; and in May one of the C.E.O.s was downgraded to chief innovation officer. A co-C.E.O. booster like Marc Feigen might argue that the fit simply wasn’t right. But a co-C.E.O. skeptic might argue that it’s the very model that is wrong.

Jeffrey SONNENFELD: The trouble is, there’s a lot of ambiguity as to who’s in control.

That is Jeffrey Sonnenfeld.

SONNENFELD: Senior associate dean, and Lester Crown Professor of Management Studies at the Yale School of Management.

DUBNER: You also run something called the Chief Executive Leadership Institute. Is that correct? 

SONNENFELD: Yeah. The world’s first school for incumbent C.E.O.s.

DUBNER: What’s your mission and how does it work? 

SONNENFELD: Our goal is to take highly accomplished people who can atrophy in high office because they’re very busy. They’re not always the most reflective people. They’re very action-oriented people — often brilliant, and many times just as smart, if not smarter than many of my scholarly friends, but they have short attention spans. So we create a learning format that matches their wide portfolio of pragmatic interests without drilling down too long on any one topic. And the goal is to make sure that people at the top stay fresh. Because that old Greek adage about how the fish rots from the head — we want to make sure that these people stay current, or else their misinformation goes cascading down the firm to disastrous consequences for the rest of us.

Okay, that old adage I have heard of. So, rotting fish, one; alligator versus bear, zero. I asked Jeff Sonnenfeld about another old adage, one that might apply to any C.E.O.: is it truly lonely at the top?

SONNENFELD: It’s very lonely at the top. One of the world’s largest consumer-goods companies told me that, they basically have no friends. Similarly, the C.E.O. of one of the very largest, top-three banks, said they have no friends. And they’re pretty affable people. There are people they meet all the time, and they have an awful lot of business associates, but their lieutenants — as we see this baby boom generation is getting pushed out by anxious millennials that want the job already — their lieutenants are often eager aspirants that they can’t always completely confide in, where their vulnerabilities are. And with the elevated and needed board vigilance, and because of regulation, financial disclosure, they are very cautious about what they even share with family members. So, you know, intimate friends and family may be able to give them emotional support, but just to talk through business problems is hard. And consultants sometimes can become, you know, corporate intestinal tapeworms that when they come in, they never leave. So when they sometimes confide in consultants, it’s hard to get one that doesn’t have a commercial agenda to lengthen their stay and build a dependance.

DUBNER: So it sounds to me like you’re making our argument for us, which is co-C.E.O. is a model that’s at least well worth considering for a lot of firms. So what is your position on that? 

SONNENFELD: Not to disappoint you, but the right answer is that these C.E.O.s need to find community, like our programs — where they can talk off-the-record and go for candid feedback — and that’s what we and others provide that’s so critical. But a co-C.E.O. — there’s a lot of trouble with it. There’s role confusion as to who should be the lead spokesperson. A unity of command makes a lot of sense.

DUBNER: So Jeff, we recently spoke with another C.E.O. consultant, Marc Feigen, who as I believe you know has written an article in favor of co-C.E.O.s. This was based primarily on the fact that big, public firms with co-C.E.O.s — although there aren’t many of them — that they outperform firms with a single, traditional C.E.O. by nearly 40 percent. So first of all, do you believe that research?

SONNENFELD: No, not at all.

DUBNER: Say what you really think, Jeff. Come on now. 

SONNENFELD: What’s actually there in practice? Say, at Netflix or at Salesforce, where they have kept people as a retention strategy, it’s very clear who is actually in charge, and what they did is often held on to a very talented executive that was going to be stolen away by a competitor.

DUBNER: So it’s co-C.E.O. in name only, you’re saying?

SONNENFELD: Yeah. Where it’s been authentic — say, at Nordstrom, it was disastrous. And Microsoft makes the case how the shared leadership can be disastrous. It was very unclear at times — and I saw this firsthand, repeatedly on the premises — is Bill Gates calling the shots or is Steve Ballmer calling the shots? By contrast, there’s no doubt that Satya Nadella is the unitary C.E.O. And he’s not an autocrat, a bully. He’s a wise and courageous, bold leader. But it’s the antithesis of co-leadership.

DUBNER: When we look at partnership pairs, pairs of people working together on a common goal, they’re pretty common in many other domains. We see in scientific research, a lot of academic research, and in music, elsewhere in the arts, and so on. Why does it seem to work relatively well in certain domains, but not in business?

SONNENFELD: Well, if you take a look at the backstage view of that great film that’s, like, six hours long that came out on the Beatles, you realize what we thought was collaborative, in fact, wasn’t all along. Anybody who watched that can’t help but walk away feeling really sorry for George, and that even John Lennon didn’t have the force of personality we thought they did. Paul McCartney was driving the show, and Ringo was trying to placate — so what we thought as a combo wasn’t. 

DUBNER: I once spent a lot of time with Paul Simon when he was writing this Broadway musical called Capeman — which turned out not to be a great success, even though it had a lot of great music in it. And he was talking about how this was the first time he’d ever done a truly collaborative creation, because Broadway is totally different — you’ve got the book writer, choreographer, da da da. And I said, “Yeah, but what about Simon and Garfunkel?” He said, “Oh, please, that wasn’t a collaboration. Like, Artie just showed up and sang.”

SONNENFELD: Oh, Stephen, you made my point so well. Of course. Simon and Garfunkel were not Simon and Garfunkel. It was Simon and Garfunkel.

DUBNER: So not only are you not a believer in the pair as a viable business-leadership construct, but you believe that most pairs are probably much less pair-like than we think, that it’s more of a presentation and not a reality? 

SONNENFELD: Yeah. I mean other than the tribute to Iwo Jima and the tribute to the public safety workers at 9/11, I don’t think your listeners could come up with two or three more examples of public monuments to committees, task forces. Wherever public park you go to, common square, anywhere in the world, any continent — it’s a tribute to a bold individual. Entrepreneurs, they can make mistakes, but they do have courage. And that boldness matters. And you lose that boldness and that courage and that prudent risk-taking. You don’t want reckless people, which is why you want a board to help backstop them. But they should have a decision-maker who ultimately has a vision, has authority, and can take command, and don’t have to keep saying, “Oh, no, after you.”

What Sonnenfeld is talking about here reminds me of the “great man theory of history” promoted by Thomas Carlyle, the 19^th-century Scottish philosopher. Carlyle believed that every generation, if they’re lucky, is blessed by a divinely-inspired hero who’s capable of leading the rest of us. That idea has fallen out of favor with historians, but it seems to live on in business schools. Jeff Sonnenfeld, who strongly prefers the solo C.E.O. to the co-C.E.O., has been called “the C.E.O. whisperer” by Business Insider. And that, you may recall, is exactly the same name that Fortune magazine gave to Marc Feigen, who likes the idea of co-C.E.O.s. So which whisperer is right? And: what other jobs might be better done by two people instead of one?

Laurie WILLIAMS: There’s a lot of naysayers saying, “Why would you ever have two people doing something one person could do? That must be twice as expensive.”

That’s coming up.

*      *      *

Laurie Williams is a professor of computer science at North Carolina State University. But back in the 1990s, she was writing software at IBM. And this usually meant working alone.

WILLIAMS: I could classify some of my time at IBM as being in my office suffering in silence. I mean, you could just sit there and be like, “I don’t know why this isn’t working.” And you just sit there and sit there and sit there, and not figure out why it’s not working.

So she quit, and decided to get a Ph.D., at the University of Utah.

WILLIAMS: Back in 1998, when pair programming came out, my advisor’s eyes lit up. And those of us who have been Ph.D. students, we know if our advisor’s eyes lit up, then that’s the topic we’re interested in all of a sudden. 

Pair programming is just what it sounds like: two software engineers working together on a single program, often using a single computer.

WILLIAMS: Typically one person is called the “driver” and one’s called the “navigator.” And the driver is typing on the keyboard, moving the mouse like getting something recorded. And the navigator is watching the work of the driver. But, like, if you looked at a video of pair programmers, it’s chatter, chatter, chatter. Like, it’s a constant communication. So it’s not like the navigator is just watching and falling asleep.

For her dissertation, Williams wanted to test how pair programming compared to the kind of solo programming she’d done at IBM.

WILLIAMS: There’s a lot of naysayers saying, “No, why would you ever have two people doing something one person could do? That must be twice as expensive.”

Williams designed a simple experiment for a computer-science course.

WILLIAMS: There was 41 people in the class. Twenty-eight were formed into 14 pairs, and 13 students worked alone for a whole 15-week semester. They had a number of programs that they did, and each one was about two weeks. So it was pretty sizable. They weren’t trivial programs. In a particular two-week period, the people who worked in pairs had to do two programs, and the people who worked alone had to do one. Then we measured the results of the work, on a number of dimensions. One of them was quality. They had to record the amount of time they spent. And then we also continuously recorded their satisfaction.

Williams found that the pairs worked a bit more slowly than the solo programmers. If, for instance, a solo programmer spent 10 hours on an assignment, a programming pair spent about five-and-a-half hours working together, or 11 total hours. But they made way fewer mistakes.

WILLIAMS: If you took the typical time it would take to fix the defects, then the pairs would have done better. So from an economic standpoint, overall, it was a positive picture. 

The programming pairs also reported higher satisfaction.

WILLIAMS: This is both in the study as well as in industry — people were happier. They’re social, humans are social. And if you’re working alone on probably anything, you’re like, “Okay, this is how I’m going to do it.” And you have in your head how you’re going to do it. But if, in that process, someone says, like, “Oh, why are you doing that? I would have done this,” now you have a conversation, and maybe you end up sticking with what you said. But maybe what the other person said will make you think again.

Like Marc Feigen’s research on co-C.E.O.s, Laurie Williams’s research had a small sample set. But her findings seem to have had a real impact. In a 2018 survey of 100,000 software developers, nearly 30 percent said they now sometimes use pair programming.

WILLIAMS: People were very appreciative. I got a lot of email and whatnot saying, like, “Thank you for giving us permission to do it.” So it really allowed, — it, like, unlocked the ability of companies to actually use the practice because it took away the manager’s, like, “It’s going to cost twice as much.” The permission that it didn’t cost 2x, I think that made a big difference, honestly.

So, will Marc Feigen’s work on co-C.E.O.s have the same kind of impact? Will co-C.E.O.s someday be as common as programming pairs?

FEIGEN: I counsel my C.E.O. clients to be riveted in the detail. And that means they really have to understand how human resource and finance and investor relations ticks. And they have to understand how R&D is working, and they have to understand how the capital management process is working at a granular level. And so, co-C.E.O.s can double the ability to penetrate and really get underneath what’s making this company perform and what the obstacles are that are keeping us from performing at a higher level. 

DUBNER: So, Marc, assuming that the trend of sole C.E.O.s being dominant doesn’t change — let’s say over the next 10 or 20 years, there’s no big rise in co-C.E.O.s. — why will the idea not have taken root? What would you ascribe that failure to? 

FEIGEN: I think in smaller companies we will see, and do see, lots of co-C.E.O.s. So, in the very largest companies, I think that boards will be fearful of taking a risk of the blow-up. And that’s understandable. So, I don’t expect that the Fortune 100 is going to have 20 or 30 co-C.E.O.s in the next 10 years. But I bet it’ll have 5 or 10, and I bet that those companies will do well.

DUBNER: And then, of course, if that trend catches on, copycats being copycats —

FEIGEN: Yeah. I think 30 years. I’m willing to make a bet — 25 percent might have. And I think if you peeled the onion — now and in five years you’ll see many partnerships. So, you’ll see a C.E.O. and a president, you’ll see a C.E.O. and a C.F.O. who work together, are in each other’s offices eight times a day. I think another secret sauce in getting co-C.E.O.s to work is to give them time to work together as leaders before they become co-C.E.O. K.K.R., a huge private-equity firm, did this very well. In 2017, they appointed Joe Bae and Scott Nuttall as co-presidents, and four years later they became co-C.E.O.s. By that time, they had been working together so well and had such a partnership and had succeeded — the market cap tripled, the assets under management doubled, so it was a success — but not putting co-C.E.O.s into it cold. My mother and her best friend had known each other since they were six. Having experience together is helpful if you have the luxury to do that.

DUBNER: So, as someone who’s running a professional C.E.O. advisory service, is it possible, Marc, that you advocate for co-C.E.O.s simply because that means there will be more C.E.O.s, which is good for your business?

FEIGEN: It’s terrible for my business. I was thinking of not going on your podcast. Because the lonely C.E.O.s like me, and I like them. A pair of C.E.O.s have each other. So, no. I have zero commercial interest in this. I’m fascinated by this. The research surprised me as much as it has others, and so, I want to share it. 

And what do you think? Did this research surprise you? Do you find the research believable? I’m guessing a lot of you have strong opinions about whether co-C.E.O.s are a good idea, and we would love to hear them. Send us an email: radio@freakonomics.com. And remember: Freakonomics Radio Plus is our new membership program, with bonus episodes and ad-free listening. Sign up to hear from a married couple whose attempt to be co-C.E.O.s didn’t work out. Just go to the Freakonomics Radio show page on Apple Podcasts, or go to freakonomics.com/plus.

*      *      *

Freakonomics Radio is produced by Stitcher and Renbud Radio. This episode was produced by Ryan Kelley and mixed by Greg Rippin, with help from Jeremy Johnston. Our staff also includes Alina Kulman, Daria Klenert, Eleanor Osborne, Elsa Hernandez, Gabriel Roth, Jasmin Klinger, Julie Kanfer, Lyric Bowditch, Morgan Levey, Neal Carruth, Rebecca Lee Douglas, Sarah Lilley, and Zack Lapinski. Our theme song is “Mr. Fortune,” by the Hitchhikers; all the other music was composed by Luis Guerra.