180 7: Decision Th eory
of economic cost and benefi ts, but easy to explain in the likely overestimation of
the possible risks of being caught undercomplying. Overcompliance can also be
normative. Many universities, for example, essentially overcomply with affi rma-
tive action requirements simply because of ideological commitments to diversity.
Institutional undercompliance is more oft en explained by ignorance rather than
by risk-prone decisionmakers (DeHart-Davis and Bozeman 2001).
Th e propensity toward risk taking is associated with goals or targets. Less risk
will be taken if goals are met or nearly met, whereas more risk will be taken if
the individual or institution expects to fall well below expected goals. Goals and
targets tend to be adjusted to adapt to risk. Successful risk taking opens the way
for higher goals, and unsuccessful risk taking leads to lower aspirations (March
1994).
Success-inclined risk has to do with the prosperity of key decisionmakers to
attribute success to their abilities and failure to their bad fortune. Persistent exec-
utive successes lead to an underestimation of risk because experiences have been
based on successes. Successful executives are promoted and tend to have high
confi dence in their abilities. Because they know the secrets of success, they have
such confi dence that they can beat the odds that they may guess wrong or fail to
anticipate changing circumstances. Th e remarkable success of the dot-coms in
the 1980s and 1990s and the sharp reversal of dot-com fortunes at the turn of the
century illustrate the success-induced underestimation of risk.
Risk underestimation based on experience is, at one level, rational, because
most decisionmakers have not directly experienced unlikely events. On the other
hand, experience with a particular decisionmaking context can cause decision-
makers to treat that decision as the “anchor” from which all future decisions are
made (Ariely 2009). Th is tendency, coupled with a bias for the status quo and the
overwhelming tendency toward loss aversion (Kahneman, Knetsch, and Th aler
1991), can create conservative institutions.
Because rare events are unlikely, the decision process is biased—fi rst, in the
direction of overlooking the very substantial consequences of some rare events,
and second, in assuming that, if one rare event is unlikely (a fl ood, an earthquake,
a depression), all possible rare events are unlikely. It is correct to assume that one
rare event is highly unlikely, but it is incorrect to assume that all possible rare
events are all highly unlikely. Organizational decisionmakers and planners tend
to base their plans and decisions on essentially linear extrapolations of their direct
experiences (Tuchman 1984; Roberts 1999; Conquest 2000).
High-reliability organizations (nuclear power systems, air traffi c security sys-
tems, nuclear ships, space travel, etc.) are especially structured to reduce risk.
High-reliability decision protocols include a much diff erent logic when com-
pared with trial-and-error, failure-tolerant systems. Th e incremental, mixed
scanning, loose coupling, resource scarcity, and bounded rationality theories—
theories that explain much of standard organizational behavior—are replaced
in high-reliability theory, described in Chapter 4. Most persons associated with