The Role of the Boss

Document Sample
The Role of the Boss Powered By Docstoc
					                                    Perceiving the Boss
                         Martin Leroch (University of Hamburg) and
                               Wesley Pech (Wofford College)

                                       - Project Outline -

The Problem
       The main goal of this project is to investigate experimentally how different ways of
organizing team production affect their productivity. In the standard Labor Economics
literature, it has been argued, for instance, that workers subordination to the hierarchy of a
firm, usually called “discipline”, has succeeded as a way of organizing work because it
increased the workers’ effort levels, and therefore avoiding the standard free-rider problem
present in this type of interaction (see, e.g., Alchian and Demsetz 1972 and Clark 1994).
According to this line of argument, letting the capital owner be the “manager”, i.e. the
monitor, reduces the incentive problems in team production. This, in turn, has been argued to
have increased the productivity of workers and firms. Consequently, firms adopting discipline
as an organizational feature would out-compete other forms of organization.
       At the core of this argument is the view that workers respond to hierarchies, based
purely on extrinsic incentives, such as monetary compensation, fines, threats of dismissal,
etc.. In other words, it is assumed that people increase their effort levels under supervision,
combined with the threat of being fired, than when they work for themselves. From a
sociological point of view one could add that people do indeed respond to hierarchies. For
instance, the now famous Milgram experiments (see Milgram 1974) suggest that merely
telling a person that someone is an “authority” over her or someone else already triggers
severe changes in behavior. This is even the case if the change of behavior is accompanied by
massive psychological distress. However, what causes this response to hierarchies is not clear.
It could be argued that the fear of retaliation from the authority causes these changes. There is
also a more “positive” story. According to Akerlof (1982), workers increase their effort while
conducting a certain task as a response to a positively evaluated behavior of the firm, a motive
that came to be called “gift exchange.”

Related Work

       The validity of Akerlof’s argument that positive behavior is responded in the same
manner is now well supported on experimental grounds, e.g. by Fehr and Gächter (2002) and
Fehr and List (2004). The main idea behind these results is that, instead of having self-
regarding preferences, humans behave reciprocally. This means that they will be kind towards
those who were kind to them (positive reciprocity), and will be hostile towards those who
were unkind to them (negative reciprocity), even when no future material benefits are to be
expected from these responses.
       In general, these findings of reciprocal behavior should be applicable to experimental
settings of team production. Accordingly, introducing a “boss”, i.e. someone who can in some
way or another exercise authority over the members of the team, should alter the behavior of
the team members by shifting their reference point. Despite this theoretical reasoning,
experiments concerning the role a boss are rare. Exceptions are given by van der Heijden et
al. (2006) and Pech (2007). Van der Heijden et al. (2006) tackle the question of how the boss
alters behavior in contrasting two settings. In the first setting, a non-hierarchical team
produces some output which is then distributed equally at the end of the period among the
team members. Consistent with the experimental evidence, the free-riding problem of team
production occurs in this setting when the game is repeated. In the second setting a leader is
introduced. He can observe the effort levels of the individual team members and condition the
distribution of team output on this effort level, ex post. This introduces the possibility for the
leader to keep the whole production for himself. What van der Heijden et al. (2006) find is
that the introduction of a boss is in most cases efficiency-enhancing, that is, free-riding was
reduced by the introduction of the leader. Most leaders do not take advantage of the
possibility to capture all gains. The authors also find that “What characterizes … poorly-
performing teams is that their leaders fail to send unambiguous signals of fairness. As a
consequence, cooperation by the other team members breaks down” (p. 17). However, their
experiment does not introduce differences in the way the leader is introduced, and does not
analyze the role of hierarchy when the division of the proceeds are decided in advance. That
is, there is only one “type” of leader who decides the distribution only after observing the
levels of contribution.
       This is changed in Pech (2007), where four different settings are introduced; an
unproductive boss with partly fixed endowment, an unproductive boss without such an
endowment, a productive boss and no boss at all. In all cases in which there is a boss, she has
to decide how to divide total output between her and the workers before the workers decide
how much to contribute. Moreover, this study also introduced the possibility of punish free-

riding behavior after a sequence of rounds without punishment. The main results show that: 1.
Intentions matter. An exogenous egalitarian distribution, as it happens in the No Boss case, is
not as successful as the case in which a Productive Boss has the power to get any fraction she
wants, but instead decides to divide it equally. The material incentives were exactly the same,
but in the Productive Boss situation the workers perceived an equal split as a fair actions, and
then reciprocated accordingly; 2. Even though the introduction of an unproductive boss
decreased the marginal benefit of workers, contributions were higher when the boss is very
generous when compared to the “no boss” case, which suggests that workers heavily
reciprocate the generosity of the boss; 3. Although punishment increased average
contributions in all treatments (particularly when there was no boss), it was the punishment of
a productive boss that caused free-riders to increase contributions more in the next round,
compared to the punishment of other types of bosses/co-workers; and 4. The possibility of
punishment shifted the reference point of workers, making them more sensitive to negative
reciprocity than positive reciprocity.
       These findings suggest a similar conclusion to the one from van der Heijden et al.
(2006), namely that the perception of the boss is crucial for his effect on the team members’
contributions. If the boss is perceived as fair, team members will try to reciprocate this
fairness by higher contributions. If the boss is perceived as unfriendly, contributions will be
lower, respectively.
       To the best of our knowledge, there is no experiment which deals explicitly with
differences in the perception of the boss in case of different kinds of bosses. Both Fehr and
List (2004) and van der Heijden et al. (2006) focus on one type of boss. This implies that the
“role of the boss” is the same, no matter what “kind of a boss” she is. Pech (2007), however,
has shown that it is crucial for the outcome what kind of a boss is observed. This, in turn,
implies that the way the boss comes about, his relation to the workers, already induces a
specific perception of him on the side of the workers. It could well be that non-productive
bosses are regarded as being “different” and the situation is framed “negatively”. Therefore a
significant amount of trust could be necessary in order for the bosses to overcome the
negative frame and to induce positive reciprocity on the side of the workers.
       This argument is in line with experimental evidence in psychology. For instance, van
Vugt and De Cremer (1999: 595) found in an experiment that “in choosing between different
leaders, individuals consistently preferred to adopt those with a legitimate power base. That
is, group members preferred a democratic over an autocratic leader, an elected over an
appointed leader, and a leader from the inside over a leader from the outside.” The reason for

the latter finding, i.e. that leaders from the inside are preferred over those from the outside,
lies in the fact that group members “perceive some similarity between the leader and group in
terms of shared attributes and values” (p. 588) if they face an inside leader. However, the
focus of van Vugt and De Cremer (1999) is on how group members will choose their boss,
and how much power the boss should have, and not on how they will react on different types
of bosses, as we suggest.

Experimental Outline
       The core hypothesis to be tested is that the differences in “additional” contributions
from the side of the workers vary between the two stages. If this were the case, the different
roles of the bosses could be interpreted as altering the expectations about their behaviour and
hence the adequate reaction to a certain kind of behaviour. Put differently, the role of the boss
is not neutral but influences the expectations people have about his behaviour and hence how
they perceive him. We hypothesize that the less “involved” the boss is (i.e. the more features
distinguish him from the workers), the more positive kind behaviour will be evaluated and
hence reciprocated. Thus, we assume a negative “starting effect” of difference, i.e. the more
different people are, the more negative they perceive each other. If it turns out that they are
kind, one might tend to evaluate the kindness to a larger degree (i.e. more positive) because
we do not expect kind behaviour from people we have actually perceived as “negative”.
       In order to test our hypothesis, we will analyze two different settings of a standard
public goods game. More specifically, there is a team of 3 people. All three members of the
team are given an amount of 20 experimental units (EU). They may choose to contribute part
of their initial endowments to a public good. Each EU contributed will be multiplied by a
specific factor which will depend on the specific setting the team members will face. In order
to exclude differences in monetary incentives, the marginal per capita return (MPCR) of each
EU contributed to the public good will be the same, namely 0.5. Thus, each team member will
receive 0.5 EU for each EU he contributes to the public good. (A table with the exact
multipliers is given below.)
       In the first setting one of the team members will (randomly) be assigned the role of a
“boss”. By being assigned this role, he will be able to choose between different distributions
of the public good, either to his own advantage or disadvantage, or with an equal share for all
members. In order to simplify matters, the boss may choose between a “kind distribution”,
which will yield him a share of 0.2 of the total output and distribute the remaining share of 0.8
equally among the other two team members. In order to yield an MPCR of 0.5, the multiplier

thus has to be equal to 5/2. In case of the “egalitarian distribution” each player will receive the
same share of total output, i.e. 1/3 (multiplier 1.5). In case of the “unkind distribution” the
boss may keep half of the total output and distribute the remaining half equally between the
two team members (multiplier 2). Consequently, the boss in this setting has two different
ways of increasing his earnings in the game. On the one hand, he may increase his own
contributions to the public good. On the other hand, he may choose a different distribution of
the final outcome.
       The second setting is similar to the first one. Again there will be a team of 3 people
contributing to a public good. Again the MPCR of doing so will be 0.5. However, in this
setting there will be assigned an “unproductive boss”. That is, the boss will be a 4th member,
not have the possibility to contribute to the public good. She may only choose between the
different distributions of the public good. Again she will have the choice between 3 different
distributions, securing her either a share of 0.2 in the kind (multiplier 15/8), 0.25 in the equal
(multiplier 2), and 0.5 in the unkind distribution (multiplier 3). Again, the MPCR of
contributing to the public good will constantly be 0.5 for the team members.
       Besides these distinctions in settings, there will be one stage without punishment and
one stage with punishment in each setting, each stage lasting 10 rounds. That is to say that the
boss in the first 10 rounds will not have the possibility to punish members contributing “too
little” according to her expectations. In the second stage (again lasting 10 rounds), the boss
will at the beginning of the stage have the option of deliberately excluding one of the
alternatives and exclude the possibility of punishing the workers. His choice whether he
excludes an option is made public in the group before contributions are made. We hypothesize
that contributions increase if the boss in the second stage announces that he will not punish
because this will be interpreted as showing trust in the workers. This setting thus corresponds
to the one in Fehr and List (2004), where it was found that the mere availability of incentives
(as opposed to their use) gives rise to “hidden returns”.


Akerlof, G. (1982), “Labor Contracts as Partial Gift Exchange”, The Quarterly Journal of
Economics 97: 543-569.
Alchian, A., and H. Demsetz (1972), ”Production, Information Costs, and Economic
Organization”, The American Economic Review 62 (5): 777-795.

Clark, C. (1994), “Factory Discipline”, The Journal of Economic History 54: 128-163.

Fehr, E. and S. Gächter (2002), “Do Incentive Contracts Crowd Out Voluntary
Cooperation?”, Zürich, Institute of Empirical Research in Economics.

Fehr, E. and J. List (2004), “The Hidden Costs and Returns of Incentives: Trust and
Trustworthiness among CEOs”, Journal of the European Economic Association 2: 1-30.

Milgram, S. (1974), Obedience to Authority: An Experimental View, London: Tavistock.

Pech, W. (2007), “The Role of the Boss in Team Production”, Paper presented at the Graduate
Workshop in Economics 2007, New School, New York, and University of Massachusetts,

Van der Heijden, E., J. Potters, M. Sefton (2006), „Hierarchy and Opportunism in Teams“,
CeDEx Discussion Paper No. 2006-15.

Van Vugt, M. and D. De Cremer (1999), “Leadership in Social Dilemmas: The Effects of
Group Identification on Collective Actions to Provide Public Goods.” Journal of Personality
and Social Psychology 76: 587–99.

                                Summary of the Experiment
Basic setup
   -   The basic setup is that of team production or public goods game. The members of the
       team initially receive 20 experimental units (EU). They may then choose to contribute
       any share of their EU to a joint project; the rest may be kept as private return. Every
       EU contributed to this project will be multiplied by a factor. The full produce will be
       distributed among the players after each round.
   -   2 different kinds of team production:
           1. Productive boss with endowment (PB): A team of workers consisting of 3
              members, one of them being the “boss”. The boss may in advance choose
              which additional share of team production she wants to keep for herself, and
              how much to distribute among the other members of the team. She may also
              contribute to the team production. Each of the other team members will receive
              the same share.
           2. Unproductive boss with endowment (UPB): A group consisting of 3 workers
              and an additional “boss”. The boss receives a fixed (external) endowment of 20
              EU. The boss may in advance also choose which additional share of total
              output she wants to keep for herself, and how much to distribute among the

               other members of the team. Each of these other team members will receive the
               same share.
   -   Contributions to the joint or team product will lead to a marginal per capita return
       (MPCR), which is defined as follows:
       MPCR = multiplier * (share the boss leaves for the team) / (nr. of team members)
   -   Each of the above kinds of team production will offer the same marginal per capita
       (MPCR) for workers of contributing to the team production in each of the following 3
       treatments (teams in the NB setting will be randomly put in each of the 3 treatments):

                                Kind                         Egalitarian                         Unkind
Setting                 MPCR              Multiplier      MPCR         Multiplier          MPCR             Multiplier

Productive       (1 - 0.2)/2 * x = 0.5   x = 5/2       1/3 * x = 0.5   x = 1.5      (1 - 0.5)/2 * x = 0.5   x=2

Unproductive     (1 - 0.2)/3 * x = 0.5   x = 15/8      1/4 * x = 0.5   x=2          (1 - 0.5)/3 * x = 0.5   x=3

Stage 1:
   -   The bosses precommit to one of the three possible divisions introduced above (Kind,
       egalitarian, unkind).
   -   Workers make contributions.
   -   Sum of contributions is multiplied with the respective factor. The boss receives his
       share according to his initial commitment. Each worker receives 1/3 of the share the
       boss leaves them.
   -   Bosses have to accept any contributions made – they do not have the possibility of
       punishing or influencing workers in the sequence of rounds.

Stage 2:
   -   The bosses again precommit to one of the possible divisions.
   -   They also decide whether or not they will choose to punish individual members of the
       group if these contributed “too little” – leaving open what “too little” means in exact
       monetary terms.
   -   The choice of the bosses between punishing or not punishing is communicated to the
       workers, also explicitly mentioning the alternative the bosses would have had.
   -   Workers make contributions

-   Sum of contributions is multiplied with the respective factor. The boss receives his
    share according to his initial commitment. Each worker receives 1/3 of the share the
    boss leaves them.
-   Bosses either punish or not punish individual workers according to their initial
    announcement if these workers contribute “too little” (again leaving open what this
    means in exact terms). Punishment reduces the amount of money the bosses receive by
    1 EU and reduces the amount the punished worker receives by 3 EU. It leaves
    unaffected the amount the unpunished workers receive.


Shared By: