Motivations and observed behaviour: Evidence from ultimatum bargaining experiment

Motivations and observed behaviour: Evidence from ultimatum bargaining

experiment

Elena Tougareva

Laboratory of Social and Economic Psychology, Institute of Psychology of the Russian Academy of Sciences

Moscow, Russian Federation

selfishness axiom individuals seek to maximize their own material gains in economic interactions and expect others to do the same

Game theory: behaviour of economic agents is determined by selfishness they make their decisions from the individualistic point of view, they don’t consider the game situation from the point of view of the other

agents, they don’t take into account the interests of the others.

Deviations from game theoretic predictions in the economic experiments mostly are interpreted in terms of prosocial considerations and motivation

Question: Whether such interpretations are always correct and indubitable?

Contradictoriness of the data obtained in several economic experiments conducted in Russia contradictoriness from the point of view of their interpretation as evidences of prosocial versus individualistic decisions

Three- and Four-person ultimatum bargaining games (Linz – 1990, in Moscow – 1991) data supported our predictions, which were based on equity theory revealed effects were much stronger in Russian sample, than in western sample in typical two-person situation the decisions of western proposers were much closer to the

“economic rational” predictions, Russian proposers tended to split the “pie” equally (50/50)

Gift exchange market experiments (Moscow – 1993) strong reciprocal fairness in subject’s behaviour in a competitive environment reciprocal fairness was not undermined even by high stakes

Consistency of behaviour across experimental games (Moscow – 1994) There were used four types of games with two players (allocators and recipients) in them: trust game,

decomposed game, variant of ultimatum bargaining game, and dictator game. The games included individualistic and altruistic choices where altruistic behaviour could not result in higher payoffs than individualistic behaviour. Possibility of symmetric distribution was avoided.

overwhelming majority of the subjects demonstrated consistent individualistic behaviour only small part of them demonstrated consistent altruistic behaviour

Questionscaused by predominance of individualistic choices in the experiment where symmetric distribution was avoided and predominance of prosocial choices in all other experiments

Question: Whether when we observe prosocial behaviour we can be sure that it is indubitable evidence of prosocial motivation?

even when people behave in economic games cooperatively they tend to be conditionally cooperative (see Fischbacher, Gächter & Fehr, 2000)

Question: Whether when we observe prosocial behaviour we can be sure that it is indubitable evidence of prosocial considerations?

exactly the same decision could be caused by absolutely opposite kinds of considerationsprosocial considerations: - reciprocity as adherence to the social reciprocity norm

- “willingness to pay” as “I want to pay to the other” or “I should pay to the other” - equal splitting as satisfying of both side’s interests, equity, equality, fairness etc.

selfish considerations: - reciprocity as a reaction to punishment or rewarding from bargaining partner- “willingness to pay” as “I have to pay to the other” - equal splitting as satisfying own interests, get maximum gain with reduction the risk of losses

when we measure in the experiment only final decisions of our subjects and don’t measure real considerations that bring the subjects to these decisions the interpretations in terms of their considerations mostly remain hypothetical

Questionscaused by predominance of individualistic choices in the experiment where symmetric distribution was avoided and predominance of prosocial choices in all other experiments

Question: Whether stronger selfishness undermine prosocial behaviour or in fact prosocial decisions could be determined by pure “economic rational” considerations?

Exact replication of the four-person UB experiment (Moscow – 1991) in 1999 (Moscow).Subjects: at the same age, also university students.The subjects in 1999 were considered as more selfish: the dominating values in the society altered

profoundly from cooperative to selfish during the period 1991-1999. Question under research: whether decisions of obviously more selfish subjects will become

more close to “economic rational” predictions?

Results: the equity seeking effect was even stronger than in 1991 both proposers and responders tended to make “fair” decisions

Possible interpretations: doubtful – subjects in 1999 are much more prosocial and have stronger equity considerations more believable – subjects cares only about own interest to get as much money as possible in

given situation and they expect other player to do the sameIt looks as if it was rather competition than cooperation.

Question for this research: Whether when players make their “fair” decisions they really have any care about interests of the bargaining partner (have prosocial motivation) or they study only their own interests (are selfish)?

Ultimatum bargaining experiment (Moscow – 2000) Tougareva Elena, Grishin Vyacheslav

Method

Subjects: young people 15-18 years old (50% male and 50% female). Total N of subjects – 108, 18 pair of players in each experimental condition.

Experimental procedure: 1 step – subjects completed a self-assessment of conflict style (Thomas Conflict Mode Inventory) 2 step – two-person ultimatum bargaining game: anonymity of bargaining partners random allocation to the role of proposer and responder “pie” – 12 roubles (as in 1991 and in 1999) proposers and responders make their decisions separately only one game is played decision of proposer: distribution of money – for me and for the other decisions of responder: on all possible distributions – accept or reject this offer

Before the game the players were informed that pairs are already randomly matched (plus got additional information about bargaining partner in two experimental conditions) and answered two questions:

Proposer Responder

1. Whose position is more powerful in this bargaining situation – position of proposer or position of responder?

1. Whose position is more powerful in this bargaining situation – position of proposer or position of responder?

2. How do you think, how much money the responder will accept as minimum?

2. How do you think, how much money the proposer will keep for him/herself?

Ultimatum bargaining experiment (Moscow – 2000)

Method

Experimental conditions:

Control condition – without any additional information for the subjects in ultimatum bargaining game.

Competitive condition – before ultimatum bargaining game each subject got the information that according to the results of previous questionnaire his/her partner prefers to use competitive strategy of behaviour in conflict situations, tends to pursue his/her own ends in prejudice of others.

Cooperative condition – before ultimatum bargaining game each subject got the information that according to the results of previous questionnaire his/her partner prefers to use cooperative strategy of behaviour in conflict situations, tends to take into account interests of both sides and try to solve conflict in such way where interests of both sides are satisfied.

Ultimatum bargaining experiment (Moscow – 2000)

Method

The data were analysed from the point of view of what players expect to get under given circumstances and what final decision on their own outcome they make.

Dependent variables

Proposer Responder

Perceived distribution of power in the bargaining situation: position of the proposer is more powerful or position of the responder is more powerful.(from the answer to the first question before bargaining)

Perceived distribution of power in the bargaining situation: position of the proposer is more powerful or position of the responder is more powerful.(from the answer to the first question before bargaining)

Expected maximum allowed to keep for him/herself (from the answer to the second question before bargaining)

Expected to get for him/herself (from the answer to the second question before bargaining)

Keeps for him/herself Minimum accepts to get for him/herself Maximum accepts to get for him/herself

Deviation of the decision to keep from the expectation of maximum allowed to keep

Deviation of the decision to accept as minimum from the expectation to get

Ultimatum bargaining experiment (Moscow – 2000)

Hypotheses

Basic hypothesis: Prevalence of “fair” decisions of our subjects in ultimatum bargaining game is determined by selfish motivation and by economic rational considerations and there should be difference in data between control and experimental conditions. Effects are expected especially in the cooperative condition, because such situational factor as bigger compliance of bargaining partner could lead to more self-profitable decisions of the players.

Alternative hypothesis: Prevalence of “fair” decisions of our subjects in ultimatum bargaining game is determined by prosocial motivation and by prosocial considerations and there should not be any difference in data between control and experimental conditions. Effects are not expected, because such considerations and motivation brings to rather stable prosocial decisions, which should not be influenced sufficiently by such situational factor as partner’s personality trait (tendency to cooperate or tendency to compete).

Ultimatum bargaining experiment (Moscow – 2000)

Results

Perceived distribution of power in the bargaining situation

Proposer – proposer is more powerful control condition 50.00 % partner is competitive 66.67 % partner is cooperative 83.33 %

all conditions 66.67%

Responder – responder is more powerful control condition 61.11 % partner is competitive 61.11 % partner is cooperative 77.78 %

all conditions 66.67%

Both proposers and responders tend to consider own position as more powerful especially in cooperative condition.

Deviation of the player’s decision from the expectation (allowed to keep / expect to get)

Proposer – % of the cases without deviation control condition 77.78 % partner is competitive 55.56 % partner is cooperative 94.44 %

Responder – % of the cases without deviation control condition 55.56 % partner is competitive 55.56 % partner is cooperative 72.22 %

Both proposers and responders tend to make their decisions reasoning from the perceived boundaries of possible actions, that is perceived maximum gain (maximization of gain).

Decisions mostly coincide with the expectations especially in cooperative condition.

Deviations – 1 (or 2) roubles less than perceived maximum possible gain – intended to decrease the risk of bargaining failure (minimization of losses)

Ultimatum bargaining experiment (Moscow – 2000)

Resultsexpectations and decisions of players

Proposermedians are presented

control condition expected maximum allowed to keep 6 Keeps 6

partner is competitive expected maximum allowed to keep 6 Keeps 6

partner is cooperative expected maximum allowed to keep 7 (p < 0.0016) Keeps 7 (p < 0.0006)

Amount of money that proposers expect to be allowed to keep and keep for selves under cooperative condition is higher.

Majority of proposers under cooperative condition proposed distribution 2/3 for self and 1/3 for the other.

Respondermedians are presented

control condition expected to get 6 minimum accepts to get 5 maximum accepts to get 12

partner is competitive expected to get 5 (p < 0.135) minimum accepts to get 5 maximum accepts to get 12

partner is cooperative expected to get 6 minimum accepts to get 6 (p < 0.061) maximum accepts to get 9.5 (p < 0.134)

The level of minimum accepted offer for the majority of responders under cooperative condition is higher.

Ultimatum bargaining experiment (Moscow – 2000)

Concluding remarks on the results Basic research hypothesis that prevalence of “fair” decisions of our subjects in ultimatum

bargaining game is determined by selfish motivation and by economic rational considerations was confirmed.

Players in both roles fill that they have control under the situation and are strong enough to influence the bargaining outcomes, such confidence is the greatest under cooperative condition.

Players in both roles tried to get maximum gain with minimum losses in the framework of perceived boundaries of possible actions imposed on them by counter-partner. They cared about own selfish interests and didn’t care about interests of the counter-partner. They just expected that other player would pursue his/her own ends and would strongly resist against more selfish distribution.

When proposers and responders considered the counter-partner as more compliant (cooperative condition) they behaved more selfish and tended to get maximum gain in this situation.

The player’s dominating motivation in presented here experiment could be defined as socially constrained selfishness.

Ultimatum bargaining experiment (Moscow – 2000)

Concluding remarks on the results

Prosocial considerations were activated mostly when the players perceived possible damage for their own interests and decided that it would be “fair” to split equally (but they defended here only their own interests).

Prosocial considerations were activated also as slight reciprocity of responders under cooperative condition, as “a fine gesture” of not absolute egoists (they decrease a little – 1 or 2 roubles less – the level of maximum accepted offer for them).

Only few responders (from 3 till 5 persons – 16.7 % - 27,8 % – in each experimental condition) really had prosocial considerations and cared about balance of interests. Their decisions could be described as “equity or nothing”.

Conclusions This experiment corroborates that prosocial decisions in bargaining games are not doubtless

proof of prosocial motivation, cooperation, fairness and equity considerations. Observed “fair”, “equitable” decisions of the bargaining players don’t refute the selfishness axiom.

The interpretations of observed in the economic experiments decisions could be wrong until the researchers will try to elicit real considerations and motivation of economic agents by more sensitive and direct than observed behavioural acts techniques.

There should be assumed plural considerations and polymotivation for the same behavioural acts and a possibility of the same considerations and motivation for different behavioural acts, therefore a distribution of the different kinds of considerations and motivation should be analysed in economic experiments.

For better understanding of economic agent’s behaviour we should go deeper into the “black box” of human psychology and study not only observed economic decisions but mostly real ulterior motives and prerequisites of economic agents.

Thank you for your attention!