Monetary and non-monetary gift exchange.
Article Type: Report
Subject: Employee incentives (Analysis)
Ceremonial exchange (Analysis)
Social exchange (Analysis)
Authors: Mahmood, Saima
Zaman, Asad
Pub Date: 12/22/2010
Publication: Name: Pakistan Development Review Publisher: Pakistan Institute of Development Economics Audience: Academic Format: Magazine/Journal Subject: Business, international; Social sciences Copyright: COPYRIGHT 2010 Reproduced with permission of the Publications Division, Pakistan Institute of Development Economies, Islamabad, Pakistan. ISSN: 0030-9729
Issue: Date: Winter, 2010 Source Volume: 49 Source Issue: 4
Geographic: Geographic Scope: Pakistan Geographic Code: 9PAKI Pakistan
Accession Number: 302769875
Full Text: We study the role of reciprocity in a labour market field experiment. Recently, many experiments have been conducted to test the extent to which cash and nonmonetary gifts affect workers' productivity through social exchange. Our experiment concluded that non-monetary gifts are more effective at invoking reciprocal behaviour. Despite a higher preference for money, a non-monetary incentive results in a significant increase in productivity as compared to an equivalent cash gift. Extending the experiment for negative reciprocity, the results highlight the asymmetry of positive and negative reciprocity that exists in the field. Discontinuation of a monetary gift results in a stronger negative effect than the discontinuation of non-monetary gifts.

JEL classification: C93, J30

Keywords: Gift Exchange, Non-monetary Perks, Social Exchange, Asymmetry of Reciprocity

1. INTRODUCTION

A standard labour contract has two important components, agreed upon wage from principal and efforts that in return is provided by agent. On one hand both principal and agent have full knowledge of wage, while information on provided effort level is always incomplete due to its abstract nature. Principal can only observe output of agent, which is joint function of effort, skill level and work environment [Green (1992)]. Assuming economic agents strictly follow their material gain, the game theoretic model predicts that agent will utilise minimum possible effort level. Similarly, the principal will pay minimum wages, since additional wages cannot extract additional effort.

In contrast, the gift exchange model (GEM) is based on the critical assumption that reciprocal behaviour creates a positive relationship between wages and workers' effort levels [Akerlof (1982, 1984)]. Workers are assumed to reciprocate higher wage levels from firms by increasing their effort (positive reciprocity) and/or by decreasing their effort in retaliation for low wage (negative reciprocity). In labour market as partial gift exchange, the loyalty of workers is exchanged for higher wage, and this loyalty then can be translated to higher productivity through effective management. Experimental evidence has supported the reciprocity hypothesis both in laboratory [Fehr and Falk (2008); Fehr, et al. (1993); Fehr and Tougareva (1995); Fehr and Falk (1999); Fehr, et al. (1998); Fehr, Gachter, and Kirchsteiger (1997)] and in the field [Falk (2007); Henning-Schmidt, et al. (2005); Bellemare and Shearer (2007)].

In the real world, we find widespread use of both monetary and non-monetary incentives for labourers. Monetary incentives include provision of reward in terms of money i.e. such as commissions and bonuses, while non-monetary incentives involve non-cash payments (in-kind perk, small gifts, tickets to restaurants, picnics, social event organised at work place, encouraging employees by providing them job autonomy, involvement in decision-making, recognition certificates, assigning challenging duties, etc). It is traditional in economic theory to convert non-monetary incentives into money equivalents, and deal with only one type of incentive for labour. However, emerging research shows that the two types have different effects. For example, "According to a March 1998 survey by The Gallup Organisation Inc. and Carlson Marketing Group Inc., almost 70 percent of the 2,000 IT employees polled said nonmonetary benefits provide the best motivation for sticking around. The study also revealed that employees favour recognition from managers and supervisors by a margin of almost 2 to 1 over recognition such as large cash bonuses or salary raises tied to productivity". [Villano (1999), para. 5].

Jeffrey and Shaffer (2007) give distinct features of tangible gifts like justifiability, social reinforcement, separability and evaluability. In his influential paper, Jeffery (2002) analysing the use of monetary and non-monetary motivational strategies, argue that cash doesn't have the trophy value, does not have long life--it comes and goes. Non-monetary incentives, on contrary, have higher trophy value thus higher utility value is attached to it. Monetary rewards are treated as compensation (for doing hard work), while non-monetary rewards are treated as recognition [Pfister (2007)], and these are treated differently by employees. Any monetary incentive merged with the salary is considered as salary and not as rewards. Experiments based on concepts from social exchange theory have clearly shown that the economic implication for both kinds of incentives differ significantly [Heyman and Ariely (2004); Kube, et al. (2008)]. Even though nonmonetary incentives are commonly used in real world, economic literature implicitly or explicitly assumes that non-monetary incentives can be translated to monetary equivalents, which introduces bias in research. (1) The objective of this paper is to capture gift exchange mechanism in labour market specifically using monetary and non-monetary gifts. Under standard assumptions of economic theory, non-monetary gift (diary here) should result in similar productivity gain to its equivalent cash alternative. Many labour market studies, using this argument, pool up all money and non-money incentives into their monetary equivalent by ignoring the practical implication of using different kinds of incentives.

The exchanges in labour market are assumed to operate through market norms, where employees exchange their effort and time for monetary rewards. Heyman and Ariely (2004) argue that we are living in two markets simultaneously: social and money market. There is strong conflict between the two markets; money market operates strongly on the basis of payment and material gain. Exchanges in money market are on-spot, sharp and short term in nature. Exchanges in social setup are long term, coordinated, consistent and independent of magnitude of payment. Any occurring exchange operates either in money market or through social network, so if one is used other one is driven out. Heyman and Ariely (2004) in series of experiment using monetary and non-monetary incentives for real effort task show that reciprocal behaviour for monetary gift was lower than non-monetary gift and was sensitive to magnitude of payment. This perspective can shed light on the well-established observation that people sometimes expend more effort in exchange for no payment (a social, market) than they expend when they receive low payment (a monetary market). [Heyman and Ariely (2004), p. 787]. They also show mixed markets of both social and monetary norms more closely resemble monetary than social markets.

This paper borrows heavily from Heyman and Ariely (2004) and Kube, et al. (2008). Both of these studies find that non-monetary gift provide stronger incentives than equivalent monetary gift. However, Heyman and Ariely (2004) attribute this higher efficiency to gift exchange mechanisms [GEM], while Kube, et al. (2008) attribute it to kind intentions signalled through non-monetary gift.

We have conducted this comparative experimental study in Jun-Sep, 2008 to evaluate the use of two different incentives in gift exchange framework. We have chosen two kinds of gifts for testing the GEM: non-monetary gift (Diaries) and cash equivalents. Results are similar to that of Heyman and Ariely (2004) and Kube, et al. (2008): non-monetary gifts result in significantly higher productivity gain compared to cash equivalents. Further, this experiment was extended to investigate the asymmetry of reciprocity using both kind of gift. We have invited few of subjects (due to budget constraint) for second round and paid them originally announced wage. Results supported the asymmetry of reciprocity, a stronger negative behaviour was observed in monetary gift group. Discontinuing non-monetary gift also resulted into productivity loss, however, significantly less than cash gift group.

This paper contributes to existing literature in many ways. First of all it provides additional experimental evidence of reciprocal and social exchange theory which is mainly tested in lab environment, except Kube, et al. (2008). While there are many lab studies proving gift exchange mechanism, field evidence is not always convincing [Kube, et al. (2006, 2008); Gneezy and List (2006) and Henning-Schmidt, et al. (2009)].

Secondly, experiment also provides alternative explanation for use of non-monetary gifts. As discussed earlier, Kube, et al. (2008) attribute the efficiency of non-monetary incentives to kind intentions signalled by 'the gift' aspect of non-monetary gift. In addition to kind intentions, we also found greater pleasure and trophy value of non-monetary gift which yields higher utility gain from non monetary gift. Separability of non-monetary gift also inflates personal value attached to it. People do not evaluate assets collectively; they rather make separate mental accounts for each type of incentive. In such case "the neutral reference point for evaluating the cash bonus will be the employee's base salary, and will make the award more subject to the value-reducing effects of diminishing marginal utility" [Jeffery (2002)].

Thirdly, while there are many studies on asymmetry of reciprocal behaviour, we are not aware of studies comparing asymmetry of reciprocity using monetary and non-monetary gift. Our study has supported the asymmetry of reciprocity for both kinds of gifts. The patterns were however very different. Cash gift created stronger response to act of taking back the incentive, while positive reciprocity dominated in non-monetary gift. The explanation came within our data set that cash award is soon mixed with higher wage, while non-monetary gift is treated as 'the gift'. Intuitively, a wage cut should have stronger response than discontinued gift.

Finally, experiment also provided additional evidence for preference-decision conflict observed by Hsee, et al. (1999): the preference and labour supply decision for gifts were not consistent. Hsee, et al. (1999) attributed this to pseudo-value attached to money due to high fungibility in joint valuation. An additional explanation comes from Heyman and Ariely (2004) they argue that mixed market will resemble more to money market. Additional psychological features attached to non-monetary incentives that seem working in non-monetary gift treatment may not be able to alter predicted utility of reward in joint evaluation

The remainder of this study is organised as follows. In Section 2, we describe our experimental methodology. In Section 3, we present our results and discussion. Section 4 gives the extension of experiment for negative reciprocity. Section 5 concludes the study.

2. THE EXPERIMENT

A set of 120 students were selected randomly from different universities through an advertisement displayed on notice boards. (2) Wage was announced as rupees 120 per hour in advertisement. The subjects were hired without any knowledge of being part of any experiment. There were total 179 calls from interested student of which a group of 120 students was selected randomly. (3)

The pool of selected subjects was randomly divided into four groups (i.e., control group, monetary gift treatment, non-monetary gift treatment and choice treatment groups) of thirty each. The students were informed to report on separate days. The experiment was organised in four hourly sessions with ten minutes break between each session.

They were assigned a task to grade multiple choice question (MCQ) answer sheets with given answer keys. (4) Each MCQ answer sheet had 100 questions with five possible answer options (A, B, C, D and E).The subjects have to match answer sheets with the answer key. Answer sheets were of three different types, coded as versions A, B and C along with their respective answer keys. The description of the questions were omitted, and set of answer sheets were mixed randomly just to minimise the chance of memorising answers with practice that can create a confounding factor. The subjects have to calculate the number of correct questions and write it on the session record sheet (Appendix A2). At the end of each session, recorded sheets were collected. The output is define as

Output = Total checked questions - mistakes (5)

The experiment was conducted at separate office in university. Few popular magazines (sports and fashion) and newspapers were deliberately placed in the room. A half hour paid training session was also conducted before experiment to make them familiar. Both monetary and non-monetary gifts were announced, immediately after the training session. The subjects were not monitored directly; however, coordinator was available for help if needed. Furthermore, all subjects interacted with the same project coordinators to eliminate experimenter effects.

The control group was paid the advertised wage i.e. 120 rupees per hour; in the monetary treatment group an unexpected 31 percent wage increase was announced after training session(additional 37.25 rupees per hour). In non-monetary group, New Year diaries (of worth rupees 150) were presented as gift along with their announced wage. For the choice treatment, after training session they were given a choice to select a diary or additional 150 rupees apart from announced wage. For comparison purpose, the monetary worth of two gifts was kept same.

After experiment, all subjects were asked to fill the post experiment questionnaire. Question had information on gender, major, preference for monetary and non-monetary incentives and Big 5 personality test.

To analyse the asymmetry of reciprocity, last ten subjects in monetary and non-monetary gift groups were given option to participate in the same job for second day as well. All of them opted to continue for the next day. Before start of second day's session, wage for that day was announced as per originally advertised wage i.e. rupees 120 per hour. No one refused to complete the job, rest of the experimental procedure remained same.

3. RESULTS AND DISCUSSIONS

Main findings are

* Non-monetary incentives performed significantly better than monetary incentives, temporal dimension shows productivity gain remained significant over experiment.

* Despite stated higher preference for money, higher effort level was provided in non-monetary gift treatment. Productivity gain due to non-monetary gift is attributed to both kind intentions and greater trophy value of non-monetary gift.

* Higher job satisfaction was also reported by non-monetary gift group.

* Significant asymmetric behaviour was observed after discontinuing both monetary and non-monetary gift.

* The negative response to discontinuing non-monetary gift was significantly less than the negative response to discontinuing monetary increase in wage.

3.1. Non-monetary Gift Performed Well

In simple non-parametric analysis, (6) the gift exchange mechanism is working for both kinds of gifts at 10 percent. By introducing monetary gift, i.e., an unexpected 31 percent wage increase resulted in 5.5 percent productivity gain that is statistically significant at 10 percent. Many earlier studies also reported positive wage-effort relationship [Fehr and Falk (2008); Fehr, et al. (1993); Fehr and Tougareva (1995); Fehr and Falk (1999); Fehr, et al. (1998); Fehr, Gachter, and Kirchsteiger (1997)].

Non-monetary gift treatment resulted in significant 15 percent increase in average output with an equivalent 31 percent increased labour cost. Productivity gain due to non-monetary gift is, however, less than reported by Kube, et al. (2008); they stated a 31 percent increase in average output with 20 percent increase in labour cost. Perkin (1970) in preliminary study also showed the effectiveness of non-monetary incentives in family planning. Herzberg (2003) on other hand argued that fringe benefits and non-monetary perks do not motivate, spiraling wages on contrary motivate people to seek the next wage increase. Hansen (1980) compared the relative efficiency of monetary and non-monetary gifts (ball point pen) in mail response survey. The monetary incentive was more successful in generating a higher response rate in a shorter period of time. Author attributed this low response to not perceiving correct value of the gift.

In choice treatment, only 4 (13 percent) students out of 30 chose non-monetary gift. In kind gift is very unlikely to match its recipient's preferences in joint valuation with cash. There was a significant 6.4 percent productivity gain compared to control group in choice treatment.

Choice treatment performed statistically equal to money gift and significantly less than non-monetary gift (Table2). Kube, et al. (2008) have tested choice treatment in separate lab session; more than 92 percent of subjects have chosen money.

3.1.1. Discussion

An explanation for the productivity differences are based on the theory presented by Heyman and Ariely (2004). Non-monetary gift is perceived more as gift and produce higher reciprocal behaviour. In social market, effort level is mainly derived through altruistic behaviour and social norms and level of reciprocity thus remains robust for different level of compensations. In money market reciprocity is supposed to be affected by magnitude of compensation i.e. increasing the monetary incentive will increase the magnitude of effort accordingly. However, Falk (2007) in his fundraising experiment found positive linear relationship between magnitude of non-monetary gift (post cards) and reciprocal behaviour (donation).

Kube, et al. (2008) attributed higher output in non-monetary gift to kind intentions signalled from using non-monetary gift as compared to monetary gift. Kind intentions from employers were also remained high for non-monetary gift in our post experimental questionnaire; differences are significant at 10 percent level.

Jeffery (2002), in his seminal work, argued non-monetary incentives have ability to address variety of psychological needs, so would have a deeper and long-term effect on motivation. Non-monetary incentives, due to high visibility, have greater trophy value so possess greater utility level. To test the argument, a trophy value index was calculated from set of questions (they will enjoy gift for long period of time, pride associated to gift and they are likely to tell their friend and family about their gift, for details see Appendix A4) by taking simple average of three ranks. Results showed significantly higher trophy value for non-monetary gift. Similarly gift perception for diary was significantly higher than money wage. Money on other hand quickly gets confused with high salary or payment, agents adjust their perception of wage and effect will die out. Assuming this true, subject should respond strongly to perceive wage cut than discontinuation of gift incentive. This argument has been tested and supported for asymmetry of reciprocal behaviour for both kinds of gift in Section 4.

[FIGURE 1 OMITTED]

The regression results for the data are given in Table 4. The dependent variable is the number of correctly checked questions, while treatment effects are measured using treatment dummies for these groups in comparison to control group. Column 1 gives the OLS estimates using aggregated individual level output, results are in line with Table 2. Column 2 to 5 gives robustness of treatment effect using control variables. The model is extended by incorporating temporal dimension, and interaction of time with treatments. The list of control variables is divided into two categories, ability based and socioeconomic variables. Job ability or job behaviour characteristics are measured using Big Five Personality test--a famous test used by firm for hiring. The test scores five personality traits associated with work behaviours, i.e., openness, conscientiousness, extraversion, agreeableness and neuroticism. Socio-emotional variables included the data on age, major subject in university, previous wage (if any), monetary preference and gender.

Despite a deliberate attempt to minimise learning effect (average output is increasing with time), learning remained significant during experiment. Similar patterns were also observed in Kube, et al. (2006, 2008), while positive impact of wage increases fades over time in Gneezy and List (2006). Data analysis showed that non-monetary gift resulted into significant adaptation behaviour over time even after controlling for personality traits related to work behaviour and other socio-economic variables. This increase significantly improved over the period of experiment (column 1 to 4), supporting the general perception that social relationship take time to built and improve over time.

After controlling for variables, the treatment effect of both monetary and choice treatment is insignificant over time. Gneezy and List (2006) have found a transient effect of gift on long run outcomes. Kube, et al. (2006) also showed ineffectiveness of monetary gift in the long run. Transactions in social exchange via non-monetary gift produced long term and consistent effect on the response [Heyman and Ariely (2004)].

Assuming that after controlling for treatments and time, the effort level provided by the subject is totally due to reciprocity, regressing output on the personal traits shown by the subjects we find personality trait "Agreeableness" insignificant at traditional significance levels, with hetero-corrected standard errors. High agreeableness score indicate cooperative and compassionate personality that tends to reciprocate good behaviour. Many earlier studies have found agreeableness significantly related to Gift exchange mechanism [Ben-Ner, et al. (2006); Englmaier and Leider (2010)].

3.2. High Preference for Money

Traditional utility theory is based on preferences, which are not observable. Under the influence of positivist philosophy, an attempt was made to reduce all theoretical concepts to observable ones. Samuelson (1938) introduced the idea of revealed preferences (RP) as an observable counterpart to preference. An agent is offered a choice between A and B; if she chooses A, then she has revealed a preference for A over B. This transforms an unobservable preference to an observable choice. There is a lot of discussion about whether or not choices do reveal preference in this fashion. See Hausman (2000) or Wong (2006) for a discussion and a critique.

In this paper, we differentiate between three variants of preference: True underlying preferences, stated preferences and observed preferences. True underlying preferences, in this paper context, whether the person feels happier when given money or whether he feels happier when given gift, are unobservable. However we can attempt to measure this by looking at responses to the two treatments, money or gift. If person increases his effort more when given gift, then we can say that he was motivated more by the gift, and therefore infer that he liked it more. The stated preferences are when we asked them to report/state their preferences like we did in our post experimental questionnaire, whether you would prefer money to gift. These preferences are made in hypothetical situations. Lastly the observed preferences are the actual decisions in the real world situations, like most of subjects in our experiment actually chose money gift. In both the real world and hypothetical settings, the researcher does not have information on all the factors that influence an individual's choice that determine the true underlying preferences.

Since the observed and stated preferences choice settings are quite different, there is no reason to believe that the variance of the unobserved factors in the RP setting will be identical to that of the variance of unobserved factors in the SP setting [Ben-Akiva and Morikawa (1990)]. To test the argument, our questionnaire asked subjects to rank (1 to 5 scores) their preference for money and a set of non-cash incentives for their motivational ability at work. To test the stability of stated and observed preferences for this experiment, we first defined a dummy for monetary preference on the basis of preference scores given in questionnaire. The difference between two scores is calculated as d, money preference dummy is defined as

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

The proportion of subjects with strictly money preference to total subjects was then tested for its consistency in with observed preference recorded in choice treatment (i.e. 13 percent chose money gift over non-monetary gift).

Testing for equality of proportion of subjects who strictly preferred money in questionnaire versus observed choice, we failed to reject any significant differences between the two proportions at 10 percent level of significance (Fisher exact test: p-value = 0.068). However, higher stated and observed preference for money is not translated into higher output in the monetary girl group. The preference-labour supply conflict in experiment can be attributed to mismatch of preference, fungibility of money is likely to create higher preference for money [Hsee, et al. (1999)]. Hsee (1999) argues that this prediction-decision inconsistency is caused by the presence of a "pseudo-value attribute"--an attribute which provides information about "rational" behaviour--which can cause people to choose their less preferred option. In joint valuation money market mechanism surmount the social norms and non-monetary gift loses its significance as social exchange relationship [Heyman and Ariely (2004); Ariely (2008)].

While we have showed the insignificant differences among stated and observed preferences, the main finding is that the true response to gift was significantly higher for non-monetary gift.

4. ASYMMETRY OF RECIPROCITY

During recessionary periods, wage cuts and discontinuation of incentive is common. While, workers are assumed to reciprocate higher wage levels from firms by increasing their effort (positive reciprocity) they also decrease their effort (negative reciprocity), in retaliation to lower wage. Asymmetry of reciprocity mean productivity gain due to x percent wage increase may not equal to a productivity loss due to similar x percent wage cut. Studies have shown that workers respond strongly to wage cut than wage increases [Campbell and Kamlani (1997)]. Some experimental studies have also supported the hypothesis of asymmetry reciprocity where incentives framed negatively (as fines and wage reductions) result in stronger response than positively framed incentives [Hannan, et al. (2005); Fehr and Falk (2002); Kube, et al. (2006)].

To test the argument that non-monetary gift is perceived as "Gift". while monetary gift is quickly mixed with wage increase, we introduced small manoeuvre within our experimental setup to take an additional observation of asymmetric behaviour through monetary and non-monetary gifts separately. After completion of experimental session, we asked last ten subjects (7) in two gift treatments i.e., monetary and non-monetary were given choice to participate in project for the second day as well. Discontinuation of gift resulted into significant decrease in the output of the same subjects significantly.

[FIGURE 2 OMITTED]

The summary statistics for the second day's output is given in Table 1. In monetary treatment, the positive reciprocal behaviour resulted into an increase of 5.5 percent; however, discontinuation resulted in 28.7 percent decrease in productivity. For non-monetary gift the productivity gain due to positive reciprocity (15 percent) is higher than negative reciprocal behaviour (8.3 percent). The monetary gift resulted in stronger negative behaviour compared to non-monetary gift when discontinued. Significant differences were noted when both kinds of gifts were discontinued, the comparison is given in Table 6. This supports the common perception that once introduced, its hard to roll back the cash based incentive programmes.

Fehr and Gachter (2000) and Masclet, et al. (2003) worked on monetary and non-monetary punishments respectively; both studies resulted in same conclusion that punishment increases average contributions sharply. The existence of the "non-monetary" punishment, however, increases the average level of contributions and earnings less than the monetary punishment. Herzberg (2003) also suggested that spiraling wages motivate people to seek the next wage increase, if rising wages won't motivate, reducing them might work.

4.1. Potential Explanation

There are many potential explanations given for asymmetry of reciprocity. First, the Loss aversion, according to Campbell and Kamlani (1997) people tend to value loss more than equivalent gain so effort provided by the labour is more sensitive to wage cut as compare to wage increase. In another experimental study on hot response game, Offerman (2002) showed that people react strongly to intentional hurtful choice than to intentional helpful choice. They contributed this asymmetry of response to self-serving bias. Intentionally helpful act are in line with positive self image of themselves, while intentional hurtful make sharp contrast with positive self image so results and produce strong behaviour.

We can explain the differences in asymmetric behaviour by two gifts, based on idea on mental accounting. People place all their assets in separate mental accounts, for example investment income, home appreciation and precious household item holding. Cash incentive since earned along with employment income, it is likely that subject combine this with rest of employment income. If this occurs, the neutral reference point for evaluating the cash bonus will be the employee's base salary, and will make the award more subject to the value-reducing effects of diminishing marginal utility [Kahneman and Tversky (1979), copied from Jeffery (2002)]. Cash bonuses lack separability as they go into the basic salary mental account; participants often continue to view this money as an increase in total compensation, because it is cognitively aggregated with salary. The value of the cash award for performance does not stand out anymore. Companies can counter this through a ceremony and the like (to commemorate the performance); however,

Non-cash incentives, due to separability, placed in to more specific separate mental accounts (e.g., Travel, Entertainment), etc., and not aggregated with salary account so values separately from basic salary [Jeffery (2002)]. Higher gift perception ranks for non-monetary incentive in our post experiment questionnaire support this argument. A discontinued cash program is perceived as a compensation benefit reduction rather than the end of an incentive program. This becomes even more difficult in a low paid environment. (9) Finally, the nonmonetary gift is usually taken as "gift", where cash incentives take form of rights instead of recognition [Flanagan (2006)].

5. CONCLUSION

Empirical evidence shows that wages in labour markets do not always clear the market: in many cases, firms pay a higher than market-clearing wage, resulting in higher labour supply and involuntary unemployment. A substantial amount of experimental literature favours positive relation among wages and effort, confirming efficiency enhancing reciprocal behaviour [Fehr, et al. (1993, 1997); Hannah, et al. (2002); Brown, et al. (2004)]. Social exchange theory came up with investigation of incentives as monetary and non-monetary incentives [Heyman and Ariely (2004)].

Ariely (2008). According to social exchange theory, non-monetary gift facilitate social relationships, signal kind intentions, and hence results in higher reciprocal behaviour. There are only few studies to test the ability to extract reciprocal behaviour in real field settings. One exception is Kube, et al. (2008), they have supported social exchange phenomenon in field settings. Our study not only proved the existence of social exchange theory in field but provided an alternative explanation for it as well. Results of experiment strongly support the prediction of social exchange theory and productivity gain due to non-monetary gift was significantly higher than monetary incentives.

Kube, et al. (2008) attributed high performance to perception of gift; non-monetary gift is considered as the "gift", and so gives a signal of more kind intentions. Our experiment provided evidence for separability property of non-monetary incentives; tangible non-monetary incentives segregated from the salary carry utility beyond the pure consumption value of incentive. We have also calculated trophy value indicator and results supported the argument by Jeffery (2002). A non-monetary gift on other hand may sit in the living room for years reminding the gift. Employee will evaluate the utility of tangible gift through pleasurable experience he will get from it. Jeffrey and Shaffer (2007) suggested that non-monetary incentives can produce better and cost efficient results as compared to monetary results. Employee may say they want cash but it isn't the most effective incentive always. Tangible rewards are both extrinsic and intrinsic motivators; they have a strong emotional appeal to participants' personal wants and interests. They also provide lasting satisfaction and long-term performance improvement. Cash incentives can produce short term incentives, but have little connection with sustained long run performance improvement.

Despite clean evidence of social exchange mechanism by non-monetary incentive, a higher stated and observed preference for money raised interesting puzzle. Hsee, et al. (1999) explained a similar prediction-decision inconsistency by presence of "pseudo-value attribute"--a feature which provides information about "rational" behaviour attaching pseudo value to their less preferred option in joint evaluation. In isolation, higher job satisfaction kind intentions and trophy value rank were given to non-monetary incentive, supporting the argument that monetary gift due to separability evaluated separately from salary.

Most interesting finding is significant differences in the asymmetry between both monetary and non-monetary gift. Result show negative reciprocity is stronger for money wage increase compared to non-monetary gift that can be explained by mental accounting of provided incentives.

In low incentive environment where people are unable to fulfil their basic needs, non-monetary incentive may not be as efficient results as predicted by social exchange theory. Similarly, during sluggish economic conditions, the use of non-monetary incentive in combination of monetary incentives may produce cost efficient results. Interactions in labour market are long term and have ability to incorporate social exchange very well. Cash incentives are not the only option, firms may use many alternative tool that can be used to facilitate socio-emotional relationship like more attention, care and appreciating their efforts. In more complicated situations in real world, it worth noting that exchanges in real lift may be very complicated and may not product result strictly similar to such controlled experiment. Workers might respond in many unique directions like sticking to firms in bad times, decide not to quit even if more attractive outside options are present.

Appendix A1

Sample Answer Key

[ILLUSTRATION OMITTED]

Appendix A2

Session Record Sheet

[ILLUSTRATION OMITTED]

Appendix A3

Picture of Diary

[ILLUSTRATION OMITTED]

Section B

Please rate the following statements on a scale from 1 to 7, where 1 means "fully disagree and 7 means "fully agree".

1. ... will enjoy diary/money for long period of time

1 2 3 4 5 6 7

2. ... am proud to receive diary/money

1 2 3 4 5 6 7

3. ... will tell friends and family about diary/money

1 2 3 4 5 6 7

4. ... perceived as gift

1 2 3 4 5 6 7

5. ... perceived as payment

1 2 3 4 5 6 7

6. ... risk lover

1 2 3 4 5 6 7

7. ... am satisfied with this job

1 2 3 4 5 6 7

8. ... treated kindly from employer

1 2 3 4 5 6 7

REFERENCES

Akerlof, George (1982) Labour Contracts as Partial Gift Exchange. Quarterly Journal of Economics 97:4, 543-63.

Akerlof, George (1984) Gift Exchange and Efficiency Wage Theory: Four Views. American Economic Review 74, 79-83.

Ariely, Dan (2008) Predictably Irrational: The Hidden Forces That Shape Our Decisions. Harper Collins.

Bandiera, Oriana, Iwan Barankay, and Imran Rasul (2009) Team Incentives: Evidence from a Field Experiment. University College London. (Mimeographed.)

Ben-Akiva, M. and T. Morikawa (1990) Estimation of Travel Demand Models from Multiple Data Sources. Proceedings from the 11th International Symposium on Transportation and Traffic Theory, Yokohama, Japan, Elsevier, New York, pp. 461-476.

Ben-Net, Avner and Louis Putterman (2006) Trust, Communication and Contracts: Experimental Evidence. Brown University, Department of Economics. (Working Papers 2006-23).

Brown, Martin, Armin Falk, and Ernst Fehr (2004) Relational Contracts and the Nature of Market Interactions. Econometrica 72:3, 747-780.

Campbell, C. and Kanal Kamlani (1997) The Reason for Wage Rigidity: Evidence from Survey of Firms. Quarterly journal of Economics 112, 759-89.

Englmaier, Florian and Steve Leider (2010) Gift Exchange in the Lab--It is not (only) How Much You Give--CESifo Working Paper Series. CESifo Working Paper No. CESifo Group Munich.

Falk, A. (2007) Gift Exchange in the Field. Econometrica 75:5, 1501-1511.

Fehr, E. and A. Falk (2008) Reciprocity in Experimental Markets. In C. R. Plott and V. L. Smith (eds.) Handbook of Experimental Economic Results.

Fehr, E. and E. Tougareva (1995) Do Competitive Markets with High Stakes Remove Reciprocal Fairness?--Evidence from Russia. University of Zurich. (Discussion Paper).

Fehr, E. and Simon Gachter (2000a) Cooperation and Punishment in Public Goods Experiment. American Economic Review 90, 980-994.

Fehr, E., G. Kirchsteiger and A. Reidl (1993) Does Fairness Prevent Market Clearing? An Experimental Investigation. Quarterly Journal of Economics 108, 437-60.

Fehr, E., S. Gachter, and G. Kirchsteiger (1997) Reciprocity as a Contract Enforcement Device: Experimental Evidence. Econometrica 65:4, 833-860.

Fehr, Ernst and Armin Falk (1999) Wage Rigidity in a Competitive Incomplete Contract Market. Journal of Political Economy 107:1, 106-34.

Fehr, Ernst and Armin Falk (2002) Psychological Foundations of Incentives. European Economic Review 46, 687-724.

Fehr, Ernst and Georg Kirchsteiger, and Arno Riedl (1998) Gift exchange and reciprocity in competitive experimental markets. European Economic Review 42:1, 1-34.

Gneezy, U. and J. List (2006) Putting Behavioural Economics to Work: Field Evidence on Gift Exchange. Econometrica 74:5, 1365-1384.

Green, T. B. (1992) Performance and Motivation Strategies for Today's Workforce: A Guide to Expectancy Theory Applications. Westport, CT: Quorum.

Hannah, R. L., J. H. Kagel and D. V. Moser (2002) Partial Gift Exchange in an Experimental Labour Market: Impact of Subject Population Differences, Productivity Differences, and Effort Requests on Behaviour. Journal of Labour Economics.

Hannah, R. L., V. B. Hoffman, and D. V. Moser (2005) Bonus versus penalty: Does the Contract Frame Affect Employee Effort? In R. Zwick and A. Rapoport (eds.) Experimental Business Research 2. Boston: Kluwer.

Hansen, Robert A. (1980) A Self-Perception Interpretation of the Effect of Monetary and Nonmonetary Incentives on Mail Survey Response Behaviour. Journal of Marketing Research 17, 77-83.

Hausman, Daniel M. (2000) Revealed Preference, Belief, and Game Theory. Economics and Philosophy 16, 99-115.

Hennig-Schmidt, Rockenbach Bettina Heike, and Abdolkarim Sadrieh (2008) In Search of Workers' Real Effort Reciprocity--A Field and a Laboratory Experiment. Journal of the European Economic Association 49:55, 1-16.

Herzberg, F. (2003) One More Time: How Do You Motivate Employees'? Harvard Business Review 87-96.

Heyman, James and Dan Ariely (2004) Effort for Payment: A Tale of Two Markets. Psychological Science 15:11, 787-793.

Hsee, C. K. (1999) Value Seeking and Prediction-decision Inconsistency: Why Don't People Take What They Predict They'll Like Most? Psychonomic Bulletin and Review 6:4, 555-561.

Jeffrey, S. A. (2002) The Benefits of Tangible Non-monetary Incentives. University of Chicago [Online]. Available: http://home.uchicago.edu/~sajeffre/amrmotive.pdf. November 4, 2002.

Jeffrey, Scott A. and Shaffer, Victoria (2007) The Motivational Properties of Tangible Incentives. Compensation & Benefits Review 39, 44-50.

Kube, S., M. A. Marechal, and C. Puppe (2006) Putting Reciprocity to Work--Positive Versus Negative Responses in the Field. University of St. Gallen. (Economics Discussion Paper No. 2006-27).

Kube, S., M. A. Marechal, and C. Puppe (2008) The Currency of Reciprocity--Gift-Exchange in the Workplace. Institute for Empirical Research in Economics, University of Zurich. (Working Paper).

Masclet, D., C. Noussair, S. Tucker, and M. C. Villeval (2003) Monetary and Non-Monetary Punishment in the Voluntary Contributions Mechanism. American Economic Review 93, 366-380.

Offerman, T. (2002) Hurting Hurts More Than Helping Helps. European Economic Review 46, 1423-1437.

Perkin, Gordon W. (1970) Non-Monetary Commodity Incentives in Family Planning Programs: A Preliminary Trial. Studies in Family Planning 57, 12-15.

Pfister, Thomas (2007) Mainstreamed Away? The European Employment Strategy and its Gender Equality Dimension. In European Union Studies Association (EUSA) > Biennial Conference > 2007 (10th), May 17-19, 2007, pages 25, Montreal, Canada.

Samuelson, P. (1938) A Note on the Pure Theory of Consumers' Behaviour. Economica 5, 61-71.

Villano, M. (1999) The Soft Touch. CIO Magazine. Retrieved March 16, 2001 from: http://www.cio.com/archive/031599_ec.html

Wong, Stanley (2006) Foundations of Paul Samuelson's Revealed Preference Theory, Revised Edition: A Study by the Method of Rational Reconstruction. Routledge.

(1) There are only few exceptions for non-monetary incentives as motivational tool in labour market like Kube, et al. (2008) and Bandiera, et al. (2009). Ariely (2008), Heyman and Ariely (2004) in their series of experiments showed that non-monetary incentives work more than reciprocal behaviour, these help to build consistent long-term response which is independent of magnitude of payment.

(2) We would like to thank Hisham Tariq, Muhammad Amjad Malik and their teams for providing their valuable research assistance during execution of field experiment and university administration for providing structural support, without them this was never possible.

(3) The list of all candidates was arranged in the order they called to show their willingness to participate in project. First 120 subjects were selected from randomly shuffled list of candidates.

(4) The task was not cognitive, so according to Pink (2008) it will respond to monetary incentives.

(5) Papers were later rechecked by especially designed software.

(6) Using Mann-Whitney U test (also known as Wilcoxon rank-sum test), is non-parametric test used for two independent sample to test the equality of variable mean

(7) Due to budget constraint, only ten subjects were tested for asymmetry of reciprocity.

(8) By Mann-Whitney U-test.

(9) Arnold Light, President of the Light Group at http://www.incentivesmotivate.com/art_cash_vs_merchandise.shtml.

Saima Mahmood is affiliated with the Pakistan Institute of Development Economics, Islamabad and Sate Bank of Pakistan. Asad Zaman is associated with the International Islamic University, Islamabad.

Authors' Note: We gratefully acknowledge funding and support of HEC for Social Research Unit Project jointly based at PIDE, International Islamic University, and Harvard.
Appendix A4

Post Experiment Questionnaire

Name Assigned Code
Registration No.
Department:
University
Contact No.
Age Years
Gender: Male Female

Previous Wage (if any) Rupees/hour, Not employed previously

Section A:
1. is talkative

1. Is talkative
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

2. tend to find faults with others
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

3. does a thorough job
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

4. is depressed, blue
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

5. is original, comes up with new ideas
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

6. is reserved
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

7. is helpful
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

8. can be somewhat careless
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

9. is relaxed, handles stress well
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

10. is curious about many different things
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

11. Is full of energy
Strongly       Disagree
I     Neutral    Agree     Strongly Agree
Disagree

12. starts quarrels with others
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

13. is reliable worker
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

14. can be tense
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

15. is ingenuous, a deep thinker
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

16. generates a lot of enthusiasm
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

17. has a forgiving nature
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

18. tends to be disorganised
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

19 . worries a lot
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

20. has an active imagination
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

21. tends to be quite
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

22. is generally trusting
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

23. tends to be lazy
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

24. is emotionally stable, not easily upset
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

25. is inventive
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

26. Has an assertive personality
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

27. can be cold and aloof
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

28. preserves until the task is finished
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

29. can be moody
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

30. values artistic, aesthetic experiences
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

31. is something shy, inhibited
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

32. is considerate and kind to almost everyone
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

33. does things efficiently
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

34. remains calm in tense situations
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

35. prefers work that is routine
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

36. is outgoing and sociable
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

37. is sometimes rude to others
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

38. makes plans and through with them
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

39. gets nervous easily
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

40. likes to reflect, play with ideas
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

41. Has few artistic interests
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

42. likes to cooperation with others
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

43. is easily distracted
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree T

44. is sophisticated in arts, music, or literature
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

45. extraverted, enthusiastic
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree 7

46. critical quarrelsome
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

47. dependable, self-disciplined
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

48. Anxious, easily upset
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

49. open to new experiences, complex
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

50. reserved, quiet
Strongly       Disagree 7 Neutral        Agree     Strongly Agree
Disagree

51. sympathetic, warm
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

52 disorganised, careless
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

53. calm and emotionally stable
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree

54. conventional and uncreative
Strongly       Disagree       Neutral    Agree     Strongly Agree
Disagree


Table 1
Details and Summary Statistics of All Groups

                        Monetary Gift     Non-monetary Gift
            Control                                          Choice
             Group   Gift *  No Gift **  Gift *  No Gift **  Group

Gift Given    No      Yes        No       Yes        No       Yes
Day           1st     1st       2nd       1st       2nd       1st
Mean        633.75   666.98    451.75    730.83    581.5     675.25
St. Dev.     133.3   118.9     100.7     139.9      101      128.3
Median        625     670       475       735       580       680
Min           330     380       280       400       390       330
Max           940     1000      630       1120      760       1100

* Unexpected gift was given with wages. **: upon arrival on second
day, subjects were told that they will be  given their originally
announced wage only (i.e., no gift will be given to them).

Table 2
Comparison between Treatment Groups

                     Choice               Monetary

Control        Z = 2.3 (0.020) **     z = 1.8 (0.069) *
Non-Monetary   Z = 3.3 (0.001) ***   z = 3.6 (0.000) ***
Monetary         Z = 0.5 (0.620)

                   Non-monetary

Control        z = 5.05 (0.000) ***
Non-Monetary
Monetary

*** significant at I percent , ** significant at 5 percent, and
* significant at 10 percent level of significance.

Table 3
Mean Scores of Control Variables

                             Control   Monetary   Non-monetary
                              Group      Gift         Gift
                               (1)       (2)          (3)

Enjoy Gift for Long Period     --        3.8          4.6
Proud to Receive Gift          --        3.9          4.5
Tell Friend and family         --        3.9          4.4
Trophy Value Index             --        3.9          4.5
Gift Perception                --        3.7          4.4
Payment Perception             --        4.0          4.0
Fairness of contract           4.5       4.1          4.2
Risk Behaviour                 4.7       4.0          3.9
Job Satisfaction               4.3       4.5          5.1
Kindness from Employer         4.9       3.9          4.6

                              Choice
                             Treatment    p-value
                                (4)      [+ or -]

Enjoy Gift for Long Period      3.8      0.032 **
Proud to Receive Gift           4.0        0.105
Tell Friend and family          4.1        0.234
Trophy Value Index              4.0      0.009 ***
Gift Perception                 4.3       0.073 *
Payment Perception              4.5        0.829
Fairness of contract            3.7        0.588
Risk Behaviour                  4.3        0.680
Job Satisfaction                4.4      0.003 **
Kindness from Employer          4.1       0.072 *

[+ or -] p-value of difference of ranks among monetary and
non-monetary gift groups only Columns I to 4 give the
average ranks given to given variables.

Table 4
Robustness Analysis

Variable                         (a)           (1)           (2)

Constant                        2535          588.2         756.8
                             (64.52) ***   (28.56) ***   (73.53) ***
Monetary Treatment             132.93          53           48.03
                               (92.0)        (39.0)        (38.4)
Non-monetary Treatment          388.3          31           27.9
                             (91.25) ***     (42.0)        (41.7)
Choice Treatment                 166           48           42.1
                              (91.25) *      (38.0)        (38.1)
Time                             --           18.2          18.2
                                 --         (9.44) *      (9.11) **
Monetary * Time                  --           -7.9          -7.9
                                 --          (13.0)        (12.8)
Non-monetary * Time              --           26.4          26.4
                                 --         (13.74) *    (13.40) **
Choice * Time                    --           -2.6          -2.6
                                 --          (14.2)        (14.0)
Agreeableness                    --            --            --
                                 --            --            --
Socio-Economic                   No            No            Yes
Ability                          No            No            No

Wald Test +

Monetary vs. Non-monetary                     0.54          0.50
                                             (0.588        (0.616)
Monetary vs. Choice                           0.14          0.16
                                             (0.890)       (0.870)
Non-monetary vs. Choice                       0.43          0.36
                                             (0.669)       (0.718)

Variable                         (3)           (4)

Constant                        670.8         518.3
                             (85.54) ***   (40.21) ***
Monetary Treatment              45.06         47.5
                               (38.8)        -39.24
Non-monetary Treatment          27.6          29.6
                               (41.1)         -41.1
Choice Treatment                40.9          45.3
                               (38.5)        -38.24
Time                            18.2          18.2
                              (9.04) *      (9.29) *
Monetary * Time                 -7.9          -7.9
                               (13.0)        -13.01
Non-monetary * Time             26.4          26.4
                             (13.18) **    (13.38) **
Choice * Time                   -2.G          -2.6
                               (14.1)        -14.13
Agreeableness                    0.6           0.7
                                (0.4)       (0.34) *
Socio-Economic                   Yes           No
Ability                          Yes           Yes

Wald Test +

Monetary vs. Non-monetary       0.45          0.44
                               (0.654)       (0.660)
Monetary vs. Choice             0.06          0.11
                               (0.953)       (0.912)
Non-monetary vs. Choice         0.40          0.34
                               (0.697)        (711)

We have estimated same equations Column (2) to (5) through panel GLS
with random effect and Got similar  results, so decided to report
only OLS results. Socioeconomic variables include subject major,
previous wage  and age, money preferences and gender, while ability
includes five personality traits openness,  conscientiousness,
extraversion, agreeableness and neuroticism. Values given are
coefficients, while standard  errors are given in parenthesis.
+: values are Wald statistics with p-value in parenthesis. ***
Significant at 1  percent, ** significant at 5 percent, and *
significant at 10 percent level of significance.

Table 5
Test of Preference Proportions

                                                       Proportion

Stated Monetary Preference                             0.70
Observed Monetary Preferences                          0.87

Z Score = 2.3 (Fisher Exact Test: p-value = 0.068) *

*** Significant at 1 percent, ** significant at S percent, and
* significant at 1U percent level of significance.

Table 6
Comparison of Average Treatment Effect for Sessions with No-gift (8)

                        Monetary             Non-monetary

Control            Z = -6.7 (0.000) ***     Z = -2.3 (0.024) **
Non-monetary       7 = -4 7 (0.000) ***

*** Significant at 1 percent , ** significant at 5 percent,
and * significant at 10 percent level of significance.
Gale Copyright: Copyright 2010 Gale, Cengage Learning. All rights reserved.