Strategy research has long spoken in two voices.
In one voice, the firm is a unitary actor: it maximizes, chooses, searches, competes.
In the other, the firm is a plurality: people disagree, coalitions form, conflict matters.
Cyert and March captured the tension well: “the idea of an organization goal and the conception of an organization as a coalition are implicitly contradictory.”
This paper, coauthored with the fantastic John Eklund, grew out of that tension. It started before Covid and took years because we didn’t want to offer only another critique of the unitary-actor assumption. We wanted a way to move beyond it that was both realistic and tractable.
In our new paper in Strategy Science, “Revisiting the Unitary Actor Assumption: Toward Realistic Aggregation of Individual Preferences in Strategy Research,” we ask:
If organizations are made up of people with different preferences, what is the organization’s utility function—in other words, what does the organization behave as if it values?
Our answer is not to average individual utilities.
Instead, we:
- map each individual’s utility into a choice probability,
- aggregate those probabilities using the organization’s actual decision structure,
- recover the organizational utility function that rationalizes the resulting collective behavior.
The central result is that decision structure doesn’t just shape process. It reshapes preferences.
- Under unanimity, the organizational utility function behaves roughly like the minimum of underlying utilities. It becomes more cautious and more risk averse.
- Under polyarchy, it behaves roughly like the maximum. It becomes more willing to pursue upside and accept risk.
So the same people, arranged differently, can behave like different firms.
For me, this paper also pulls together a line of work I have been pursuing for years on organizational structures and their relationship to innovation, reliability, and exploration vs. exploitation. In that sense, it feels less like a standalone paper and more like a foundation underneath several earlier ideas.
I think this matters for both scholars and managers.
For scholars, it means you can take a standard competition or incentive model, swap in a more realistic organizational utility function, and get different, testable predictions.
For managers, it suggests that governance design quietly rewrites the firm’s appetite for risk. Changing who is in the room matters. But changing how the room aggregates judgments—whether a team requires consensus or lets any champion greenlight a project—may matter just as much.
And as organizations increasingly combine human and AI judgment, the same logic applies: an AI system is another member with its own utility function, and whether it serves as a gatekeeper or a spotter reshapes the organization’s risk profile.
Structure doesn’t just shape what organizations do. It shapes what they want.
Open-access paper: https://doi.org/10.1287/stsc.2024.0257
Have you seen a team’s decision rule change what it was willing to pursue?
—