The Control Premium: A Preference for Payoff Autonomy

Abstract

We document individuals' willingness to pay to control their own payoff. Experiment participants choose whether to bet on themselves or on a partner answering a quiz question correctly. Given participants' beliefs, which we elicit separately, expected-money maximizers would bet on themselves in 56.4 percent of the decisions. However, participants actually bet on themselves in 64.9 percent of their opportunities, reflecting an aggregate control premium. The average participant is willing to sacrifice 8 percent to 15 percent of expected asset earnings to retain control. Thus, agents may incur costs to avoid delegating, and studies inferring beliefs from choices may overestimate their results on overconfidence. (JEL C91, D12, D82, D83)

The Intrinsic Value of Decision Rights

“…Non-contractible intrinsic benefits of control are likely to bias principals towards keeping decision rights, possibly to the detriment of other parties within the organization and of organizational efficiency.4 For example, Fehr, Herz, and Wilkening (2013) found significant underdelegation of decision rights from principals to agents in experimentally controlled situations in which delegation would clearly be preferable for both parties in terms of expected monetary payoffs. Similarly, Owens, Grossman, and Fackler (2014) found that individuals are willing to sacrifice expected earnings to retain control.5 Falk and Kösfeld (2006) provided a related experimental finding by showing that limiting agents' choice sets can reduce positive reciprocity. However, none of these papers identifies the intrinsic value of decision rights.…”

“…4 For example, Fehr, Herz, and Wilkening (2013) found significant underdelegation of decision rights from principals to agents in experimentally controlled situations in which delegation would clearly be preferable for both parties in terms of expected monetary payoffs. Similarly, Owens, Grossman, and Fackler (2014) found that individuals are willing to sacrifice expected earnings to retain control. 5 Falk and Kosfeld (2006) provided a related experimental finding by showing that limiting agents' choice sets can reduce positive reciprocity.…”

“…However, none of these papers identifies the intrinsic value of decision rights. In Fehr, Herz, and Wilkening (2013) as well as in Owens, Grossman, and Fackler (2014), the authors cannot rule out that the reluctance to delegate is driven by other factors such as regret or ambiguity aversion because of the uncertainty about the agents' choices after delegation. 6 Finally, our findings have a bearing on the implications of high-performance work systems (HPWS) (Ichniowsky, Shaw, and Prennushi (1997), Ichniowsky and Shaw (2003), Bartling, Fehr, and Schmidt (2012)).…”

“…However, none of these papers identifies the intrinsic value of decision rights. In Fehr, Herz, and Wilkening (2013) as well as in Owens, Grossman, and Fackler (2014), the authors cannot rule out that the reluctance to delegate is driven by other factors such as regret or ambiguity aversion because of the uncertainty about the agents' choices after delegation.6…”

Decision Initiation, Decision Implementation, and the Allocation of Decision Rights

“…How decision authority is employed is then only secondary. Frey, Benz & Stutzer (2004) indeed argue that some people enjoy decision autonomy, and Benz & Frey (2008), Fuchs-Schündeln (2009), Bartling, Fehr & Herz (2013), and Owens, Grossman & Fackler (2014) provide empirical evidence for such preferences for decision autonomy. To isolate the importance of social preferences -no matter whether distributional or reciprocal concerns -for the allocation of authority, we conduct an experiment in which potential implementation problems never occur and are thus irrelevant if subjects are selfish in the sense that they maximize their own monetary payo↵s.…”