A key principle in the book is that responsible business is not only about splitting the pie fairly (paying generous wages to employees, or offering affordable prices to customers) but also growing the pie through innovation and excellence.
That sounds all well and good, but how do you motivate innovation? Chapter 8 discusses cultural factors, such as empowerment, actively encouraging disagreement, and tolerating failure. Chapter 6 presents research showing that, contrary to common belief, hedge fund activism actually improves innovation, and Chapter 5 demonstrates that long-term CEO incentives improve the quantity, quality, and novelty of innovation.
But what about financial incentives for employees below the CEO? Do they encourage innovation or instead crowd it out? And if they do encourage innovation, what’s the optimal structure of incentives? This question is typically very tricky to answer, because correlation doesn’t employ causation. If firms that offer financial incentives to its employees are more innovative, it might be because firms that train and mentor their employees also give them financial incentives (as part of a general employee-centric approach), and it’s training and mentorship, rather than incentives, that cause innovation.
The key problem is that financial incentives are almost never randomly assigned – so you can’t do the typical treatment-control approach used in medical trials. Firms choose to implement financial incentives, and those that do so could be more innovative for other reasons.
But I said “almost never”. A new paper, by Joshua Graff Zivin and Elizabeth Lyons of UC San Diego, uses the “gold standard” of causation – the randomised control trial. They partnered with Thermo Fisher Scientific, a life sciences company, to run an innovation competition in Mexico. Participants were invited to design software solutions to increase access to high-quality medical equipment among small health care providers.
Those invited to participate were randomly assigned to one of two incentive structures. The first, a “winner-takes-all” structure, gave $15,000 to the winners. This prize is substantial – 79% of the 2018 average annual salary for Mexican software developers. The other, “egalitarian” structure, split the $15,000 pot across the top ten submissions (e.g. $6,000 went to the top-ranked submission and $600 to the tenth-ranked). The submissions were judged by an expert panel on five different dimensions. The dimension that the researchers were most interested in was “novelty”, since this is a particularly critical feature of innovation.
Joshua and Elizabeth found that the incentive structure had no effect on the quantity of innovation – in both cases, one-third of those invited to participate submitted a proposal. The overall quality of the submissions also didn’t differ significantly. But the winner-takes-all scheme led to a markedly higher increase in novelty.
This result is particularly interesting because the “winner-takes-all” scheme also leads to the greatest inequality, so it seems unjust. Critics of CEO pay typically use the CEO-to-worker pay ratio as a measure of fairness Here, the ratio of the winning team’s prize to everyone else’s is infinity – the highest possible level of inequality.
But, as stressed in the book, growing the pie is even more important than splitting the pie evenly. An unequal distribution of something, created through innovation, is almost always better than an equal distribution of nothing. No-one is worse off, and some are better off. Moreover, as covered in Chapter 5, fairness is not the same as equality. Fairness is reward that’s proportional to your contribution, and here the winning team contributed the most. (Unfairness in executive pay arises if high pay is achieved despite poor performance, or due to favourable market conditions). Moreover, innovation is often a “winner-takes-all” activity. A drug that cures a disease is vastly better than a drug that nearly cures a disease, so the winner of an innovation competition may well have contributed substantially more than the second-place entry.
But don’t winner-takes-all-schemes discourage the collaboration and diversity of perspectives that’s necessary for innovation? No – because the prize structure and the degree of collaboration are two separate concepts. Giving a winner-takes-all structure doesn’t prevent teams from forming to try for the prize. In fact, it encourages it. Joshua and Elizabeth’s competition allowed either individuals or teams of up to three to enter. Importantly, participants were asked to form teams or choose to go it alone before they were told the prize structure – so team composition wasn’t affected by the structure. The researchers found that the novelty of individual entries didn’t depend on prize structure. But teams submitted more novel entries under the “winner-takes-all” structure.
So the returns to teamwork are higher with “winner-takes-all” – unequal pay structures encourage team formation. Why might this be? An innovator might prefer to go it alone – she doensn’t need to share the prize with her team-mates, and enter what she views as the best solution rather than having to compromise to accommodate others’ views. But, if you know that only the winner gets the prize, you can’t afford to be pig-headed. You have to collaborate and learn from the perspectives of others.
As with all Randomised Control Trials, the study is on a specific setting and the results may not be widely generalisable. (See this article for the trade-off between generalisability and causality.) It took place in Mexico, was for a specific type of innovation, and innovation quality was judged subjectively by a panel rather than studying the future actual effectiveness of the ideas. The latter would have been infeasible as it would have been extremely costly to bring dozens of ideas to market. Despite these caveats, the study provides surprising results on how incentive structures not only affect innovation, but also the returns to collaboration.