The term risk culture refers to how things are seen, done and justified.
What Culture Isn’t
It’s not the explanation of last resort to be used only when other explanations—economic, demographic, organizational—are inadequate, i.e. “Oh well, it must be cultural then.”
It’s not a veto on comparison that assumes each culture is unique and can only be understood in its own terms.
It’s not the uncaused cause that explains “why did he do that?” with “because his culture told him to.”
Though often dressed up in elaborate reasoning, and bolstered by statistics, these are not explanations—just ways of saying “I don’t know.”
What Culture Is
A previous WillisWire post describes a view of risk culture based on work originated by Mary Douglas, now called Plural Rationality Theory.
This framework relies on two independent, measurable dimensions—hierarchy and attachment—producing four fundamental types of risk cultures:
- Low hierarchy, high attachment: Conservator culture sees the world as dangerously risky, requiring a careful risk-minimizing approach
- High hierarchy, high attachment: Manager culture sees the world as moderately risky; risk can be “tamed.” This culture closely aligns with the ideas usually put forward as ERM
- Low hierarchy, low attachment: Maximizer culture expects losses to be recoverable from subsequent gains, and accepts risk when compensation (price) is right
- High hierarchy, low attachment: Pragmatist culture sees unpredictability in the world, often choosing to avoid commitments and over-concentration in any one type of risk
Traditional ERM tacitly assumes that risk management is “best” when a single risk culture—the Manager culture—is universally adopted. But each of the four risk cultures can be found within most companies and management teams.
Survey of Risk Attitudes
To date, about 200 insurance executives from eleven companies have completed our survey. Each person’s answers yield a score between -10 and +10 for each risk culture, with scores 5 and above indicating a preference for that risk culture, and -5 or below indicating disagreement.
About half of respondents showed clear preference for one of the four cultures, while the other half indicated agreement with two of the four cultures.
ERM in its “purest” form would only appeal to the 17% of respondents who prefer the pure manager culture. However, since another 45% showed some affinity for manager culture, a blended maximizer/manager ERM would get the widest support, aligning directly with 29.6% and at least partly with approximately 75%.
We also investigated who in the company favored each culture.
Chart 2: Risk Preference by Type of Position
Top management contains more maximizers and fewer conservators as compared to underwriters, other staff, and board members.
When it was pointed out to one management team that they had no conservators, they responded: “That would be Stuart; he retired last year and our meetings have had many fewer arguments since then.”
Board members surveyed had slightly fewer maximizers and more conservators. A lower appetite for risk than top management is probably appropriate given their respective roles, but the lack of dramatic differences makes sense given board members are usually top managers elsewhere.
Underwriters and staff showed a much higher concentration of conservators, suggesting that a significant slice of middle management may see top management’s ideology as too aggressive.
From this evidence, we tentatively conclude that a “pure” (manager) ERM approach will probably not suit most insurers’ leadership or match up with most risk management strategies currently in place.
The plural rationality framework offers a useful way to think about risk culture. Almost all companies will find each of the four risk preferences within their staff and management team. This multifaceted view can help in developing a successful risk culture for the firm.
This blog is excerpted from a paper named as one of the Outstanding Prize Papers in the 2013 Call for Papers by the North American Chief Risk Officer Council. Alice Underwood and I wrote the paper in conjunction with Michael Thompson of the International Institute for Applied Systems Analysis. The full paper can be found at www.crocouncil.org.