Why Firms Struggle to Incentivize Innovation

Reflection

What makes this paper interesting is that it challenges a deeply rooted management assumption: that better incentives simply mean stronger accountability and closer links between pay and performance. Manso shows that this logic breaks down when the goal is innovation. In that context, the problem is not how to push people to perform harder within known routines, but how to make them willing to explore uncertain and unproven paths. Those are not the same thing, and the paper is strongest when it insists on that distinction.

The core argument is that innovation requires a different incentive structure from ordinary productivity. Standard pay-for-performance contracts work well when tasks are repetitive and outcomes are relatively predictable. But innovation is different because it involves experimentation, delayed returns, and the real possibility of failure. If people are judged too quickly or punished too harshly for weak early results, they are pushed toward safe exploitation of what already works instead of exploration of what might work better. That seems to be the paper’s central insight: incentives are not neutral tools. They shape whether an organization becomes cautious or genuinely innovative.

I found the discussion of failure especially useful because it reframes failure as something more ambiguous than bad performance. In routine work, failure often signals weak execution. In innovation, early failure may be part of the process of discovery. That does not mean all failure is valuable, of course. But it does mean organizations can misread experimentation if they use the same evaluative logic for innovation that they use for stable operations. This is where the article becomes more than a compensation argument. It is really about how firms interpret outcomes. A company that claims to value innovation but treats failed experiments as proof of incompetence is structurally discouraging the very behavior it says it wants.

Another strong point is that the paper connects this logic to governance more broadly. Measures often treated as unquestionably good — strict termination threats, intense board oversight, shareholder litigation pressure — may improve discipline while also narrowing managerial time horizons and increasing risk aversion. That tension is what I found most compelling. The article is not arguing against accountability as such. It is arguing that the form accountability takes matters. A firm can be well governed in a narrow sense and still be badly designed for innovation.

The further question this raises is how organizations can distinguish productive failure from ordinary poor performance. The paper persuasively defends tolerance for early failure, but in practice that is probably the hardest part. Every failed project can be reframed as experimentation after the fact. So the real managerial challenge may be less about whether to tolerate failure and more about developing criteria for recognizing when failure is part of serious exploration rather than just weak judgment or weak execution.

Q&A

Q1. What is the conventional theoretical model for understanding the design of organizational incentives? To what extent does this fit the context of innovation?

The conventional theoretical model is the principal-agent model. In that framework, the main problem is how a principal, such as an owner or employer, can design incentives so that an agent, such as a manager or employee, works hard rather than shirks. The standard answer is pay-for-performance: compensation is tied to output, productivity, or other measurable results. This model fits routine and repetitive work quite well, especially when effort is the main determinant of performance and outcomes can be observed relatively clearly. Manso gives examples from earlier research showing that piece rates, bonuses, and similar schemes can increase productivity in settings like installation work, agricultural labor, or other standardized tasks.

The fit becomes much weaker in the context of innovation. The problem is that innovation is not just a matter of exerting more effort within known routines. It involves experimentation, uncertainty, and delayed returns. A manager may try something genuinely promising and still fail in the short run. In that context, conventional pay-for-performance can become counterproductive, because it pushes people toward safe, predictable methods rather than exploratory ones. So the article does not reject the principal-agent framework altogether, but it shows that the standard version of it is too narrow for innovation. It works better for exploitation than for exploration.

Q2. What are different incentive mechanisms available for managers to stimulate innovation? What are their advantages and limitations?

The article discusses several incentive mechanisms managers can use to stimulate innovation. A central one is long-term compensation, especially arrangements that reward results over a longer horizon rather than immediate outcomes. In executive settings, Manso points to tools such as stock options with long vesting periods, option repricing, golden parachutes, and forms of managerial entrenchment that provide enough security and time for experimentation. Their advantage is that they reduce pressure for immediate results and make it more rational for managers to take risks. Their limitation, however, is that these tools can also be abused and are often criticized in the governance literature because they may reward managers despite weak short-term performance.

A second mechanism is tolerance for early failure, which can be built into both formal contracts and organizational culture. Its advantage is obvious: it lowers fear and makes experimentation possible. But the limitation is that tolerance can easily become vague in practice. Organizations may struggle to distinguish between productive failure that comes from exploration and simple poor performance.

A third mechanism is job security, or at least a reduced threat of termination. Manso argues that fear of dismissal may encourage safe behavior and discourage exploration. The advantage of job security is that it gives managers room to pursue uncertain projects. The limitation is that too much protection may weaken discipline and accountability if not balanced carefully.

A fourth mechanism is regular and informative feedback. The paper argues that feedback is especially important for exploration because it helps agents adjust their experiments and learn more efficiently. Its advantage is that it improves learning without necessarily turning every evaluation into punishment. Its limitation is that feedback only works well if it is actually developmental rather than just another layer of control.

More broadly, Manso also points to corporate culture, research tenure-like arrangements, and even wider institutional conditions like bankruptcy laws. Their advantage is that they shape the broader environment for innovation. Their limitation is that they are harder to design and harder to control than formal compensation contracts.

Q3. According to this article, what principles underlie the design of incentives for exploration? Explain the logic behind these principles.

The article identifies three central principles underlying incentives for exploration: tolerance for early failure, reward for long-term success, and support for learning through feedback and sufficient time. These principles follow from the paper’s distinction between exploration and exploitation. Exploration means trying new and uncertain approaches whose value is not known in advance. Because of that uncertainty, good exploratory efforts may look unsuccessful at first. If people are punished too quickly, they will rationally avoid exploration and stick to proven methods instead. That is why tolerance for early failure is essential.

The second principle is long-term orientation. The logic here is that experimentation produces information that becomes more valuable over time. A failed first step may still teach something that leads to a better method later. So if incentives focus only on short-term outcomes, they systematically undervalue exploration. Long-term rewards give managers reason to absorb early setbacks in exchange for the possibility of larger later gains.

The third principle is that exploration needs security and adjustment, not just pressure. Job security matters because the threat of termination pushes agents toward safe exploitation. Feedback matters because exploratory work often requires interim correction and refinement. In other words, exploration is not a single gamble but an adaptive process. The underlying logic of the paper is that innovation depends less on maximizing immediate output and more on creating conditions in which people can search, learn, revise, and persist long enough for better ideas to emerge.