Status Quo Bias under Uncertainty: An Experimental Study
∗
Amnon Maltz
†
Giorgia Romagnoli
‡
February 8, 2017
Abstract
Individuals’ tendency to stick to the current state of affairs, known as the status
quo bias, has been widely documented over the past 30 years. Yet, the determinants of
this phenomenon remain elusive. Following the intuition suggested by Bewley (1986),
we conduct a systematic experiment exploring the role played by different types of un-
certainty on the emergence of the bias. We find no bias when the status quo option and
the alternative are both risky (gambles with known probabilities) or both ambiguous
(gambles with unknown probabilities). The bias emerges under asymmetric presence of
ambiguity, i.e., when the status quo option is risky and the alternative ambiguous, or
vice versa. These findings are not predicted by existing models based on loss aversion
(Kahneman and Tversky, 1979) or incomplete preferences (Bewley, 1986) and suggest
a novel determinant of the status quo bias: the dissimilarity between the status quo
option and the alternative.
Keywords: Status Quo Bias, Risk, Ambiguity, Reference Effects, Experiment.
JEL Codes: C91, D11, D81.
∗
We are deeply grateful to Andrew Schotter, Efe Ok, Pietro Ortoleva and Debraj Ray for invaluable
advice. We also thank Andrew Caplin, Guillaume Frechette, Fabio Maccheroni, Jacopo Perego, Ariel Ru-
binstein, Severine Toussaeart, Sevgi Yuksel and seminar participants at NYU, BRIC Workshop at Barcelona
GSE and ESA 2014 for helpful comments. We thank the Center for Experimental Social Science at NYU
for financial support.
†
University of Haifa, amaltz@econ.haifa.ac.il
‡
University of Amsterdam, G.Romagnoli@uva.nl
1
1
Introduction
In their seminal work, Samuelson and Zeckhauser (1988) coined the term status quo bias
to describe the tendency of decision makers to stick to the current state of affairs. In
a series of field and lab experiments they showed that individuals choose to keep their
status quo option more frequently than would be predicted by the classic model of choice.
Following these original findings, the status quo bias has been established as a widespread
phenomenon whose effects have been observed in many contexts, such as the markets for
retirement plans (Madrian and Shea, 2001), electric services (Hartman et al., 1991), organ
donations (Johnson and Goldstein, 2003), financial assets (Kempf and Ruenzi, 2006) and
Medicare (Ericson, 2012). However, while heavily documented, the status quo bias is not
universally present and its magnitude varies across domains and depending on decision
makers’ characteristics.
1
One possible determinant of status quo bias is the ambiguity associated with the alter-
natives in the choice set. It stands to reason that when the environment is well known to
the decision maker, he might not hesitate to trade one option for another. However, when
there is uncertainty regarding the relative ranking of the options, the decision maker may
prefer to act “cautiously” and stick to the current state of affairs.
This hypothesis was originally advanced by Bewley (1986). He introduces the “inertia”
assumption (which is, in fact, a status quo bias assumption
2
) according to which the
agent would not move away from his current position unless one of the alternative options
dominates it, i.e., is unambiguously better. Bewley also suggests to experimentally test the
inertia assumption in relation to the presence of ambiguity in the choice set. Surprisingly
1
For example, List (2003, 2004) finds that individuals with more trading experience are less prone to
the bias. The size of the choice set, on the other hand, has a positive effect on the bias (Samuelson and
Zeckhauser, 1988; Dean, 2008; Kempf and Ruenzi, 2006; Redelmeier and Shafir, 1995).
2
The terms status quo bias and inertia will be used interchangeably in this paper.
2
enough, in the 30 years that have passed since, no experiment has investigated how different
types of uncertainty affect status quo bias. While numerous experiments have examined the
endowment effect (Thaler, 1980) in the uncertainty domain
3
, the status quo bias literature
has mostly focused on everyday ordinary goods as in the well-known study by Knetsch
(1989).
4
In this paper we attempt to fill this gap by examining the relationship between inertia
and uncertainty. For this purpose, we make use of ambiguous prospects, i.e., gambles
for which the probabilities are not objectively specified, and non-ambiguous (but risky)
prospects, i.e., gambles with specified probabilities (which we will also refer to as risky
lotteries or, simply, lotteries). We construct a 2-by-2 design where the status quo option
and the alternative may be either risky or ambiguous and test whether inertia arises in
each case. In every treatment we perform an identical within-subject design: Subjects’
preferences over gambles are initially elicited through pairwise comparisons without a status
quo option. Subsequently, subjects receive an endowment and face the same choices again.
This time around, questions are presented in the form of a switch from the endowment to
the alternative. Comparing choices with and without an endowment allows us to assess
the presence and magnitude of the status quo bias for each uncertainty structure.
5
Our results are summarized in Table 1. We find no bias when the status quo option
and the alternative are both risky or both ambiguous. The bias emerges under asymmetric
presence of ambiguity, i.e., when the status quo option is risky and the alternative ambigu-
3
Notice the difference between the status quo bias, a phenomenon reflected through choices between
goods, and the endowment effect (Thaler, 1980), defined as the gap in the evaluations of a good when it is
in the possession of the decision maker compared to when it is not.
4
In Knetsch (1989), subjects were divided into 3 groups: group 1 receive an endowment of a mug, group
2 receive a candy and group 3 (control) do not receive any endowment. Subjects in all groups are asked to
choose between the mug and the candy. Results show a significant increase in choices of the endowed good
compared to the control.
5
When referring to the experiment we use the term endowment to describe the prospect which was
given to the participants. For all purposes of this paper, it is equivalent to the term status quo option.
3
Dostları ilə paylaş: |