Choosing Meaning Over Money? Ethnographic and Field Experimental Evidence from a Handicraft Cluster in Southern India Aruna Ranganathan ∗ September 30, 2013 ∗ MIT Sloan School of Management. Direct all correspondence to arunar@mit.edu. I would like to thank Evan Apfelbaum, Stuart Bunderson, Lotte Bailyn, Curtis Chan, Jane Dutton, Roberto Fernandez, Oliver Hahl, Kate Kellogg, Paul Osterman, Mike Pratt, Ray Reagans, Ofer Sharone, Yanbo Wang, Amy Wrzesniewski, participants of the ASA Economic Sociology session, MIT’s IWER seminar, Economic Sociology seminar and Economic Sociology Working Group, and especially Abhishek Nagaraj, Susan Silbey, Tavneet Suri and Ezra Zuckerman for comments. I am also deeply indebted to Gowramma and her team at IFMR, Manjula Kale and the field research team, and myriad informants and interviewees in Channapatna. All errors remain mine. Abstract An ethnographic and field experimental study of artisans and traders in the Indian handicraft cluster of Channapatna illustrates the differential impact of calling and job work orientations on economic behavior. I find that traders, who merely retail craft products as a job, focus on monetary outcomes from the selling process, while artisans, in pursuing a higher calling in their craft work, develop psychological ownership over their handmade products and care additionally about buyers’ appreciation for their craft. Data from participant observation, interviews and a survey is used to characterize the lives and work practices of artisans and traders in Channapatna and to hypothesize the relationship between differing work orientations and selling outcomes like prices, bargaining intensity and transaction length. Empirical evidence from a field audit study supports these hypotheses by showing that, when faced with discerning auditors, artisans engage in longer transactions, bargain less and give significant discounts while traders do not, focusing solely on auditors’ willingness-to-pay. These results make empirical and theoretical contributions to the literature on meaning of work by causally establishing that work orientation impacts market outcomes and by identifying the mechanism of psychological ownership regulating these effects. “It’s like bringing up a child when you are an artisan. You get that satisfaction when you see your product transform from the raw wood to the final shape, like them taking the first footsteps, then getting shapes, then making their initial forays into the market, and finally, you have to sell them. It’s like holding their hands through this whole process and then giving them away. It’s more than the money.” (Artisan in Channapatna) Individuals derive vastly different meanings from the work that they do (Rosso et al., 2010; Pratt and Ashforth, 2003). As a way of characterizing the predominant meanings attached to the activity of work, scholars have described two contrasting work orientations – job versus calling (Bellah et al., 1985; Baumeister, 1991; Wrzesniewski et al., 1997).1 Those with a job orientation are seen to pursue their work for its material benefits rather than for other kinds of meaning or fulfillment. This orientation is one where work is primarily a means to a financial end, allowing for the enjoyment of other activities in time away from work. In contrast, those with a calling orientation are seen to work not for financial rewards but primarily for the fulfillment experienced in engaging with work (Bunderson and Thompson, 2009; Dobrow, 2006). Here, work is an end in itself, or an expression of one’s deepest self (Dik and Duffy, 2009; Hall and Chandler, 2005). Such a calling orientation is thought to develop in concert with one’s work, ultimately making work deeply meaningful, purposeful and significant (Pratt et al., 2013; Wrzesniewski, 2010). Scholarship on meaning of work not only describes different work orientations but also finds that individuals who have a calling orientation are associated with a number of distinct work-related attitudes and behaviors. Early work implicates a calling orientation with a strong work ethic and the pursuit of success, responsible for the modern capitalistic enterprise (Weber, 1930; Hardy, 1990). In recent empirical studies, workers with a calling orientation are seen to be more motivated in their job (Wrzesniewski et al., 2006), have lower levels of absenteeism (Wrzesniewski et al., 1997) and greater identification with their occupation (Dobrow, 2006). These characteristics mean 1 A third orientation - career - focusing on advancement in an organizational setting has also been discussed, though this approach to meaning-making is less universal, in that it is not applicable to the broad range of jobs that operate beyond the realm of full-time, regular employment at a living wage within formal organizations, including most jobs in the informal sector and in developing economies (Rosso et al., 2010; Brief and Nord, 1990). 1 that workers with a calling orientation seem to be associated with higher levels of performance (Hall and Chandler, 2005) within the organization and higher levels of work, life and health satisfaction (Wrzesniewski et al., 1997; Dobrow, 2006). While a calling orientation is mostly associated with positive work-related characteristics, scholars have also recognized that a sense of calling might be a “double-edged sword”, where individuals trade pay, personal time and comfort for the moral duty associated with work (Bunderson and Thompson, 2009). While the general claim in this literature seems to be that a calling orientation has important consequences for economic and work-related outcomes, such a causal claim suffers from three significant challenges. First, this research has had limited success in accurately and consistently measuring the outcomes that it aims to investigate. A number of these studies that relate work orientation to individual and firm-level outcomes rely on selfreported and survey-based measurements that suffer from well-known issues, namely their subjective and often retrospective nature, response bias and limited comparability across the population of respondents. Second, these studies are unable to rule out alternative explanations to the theorized link between work orientation and the outcomes of interest. For example, workers with a calling orientation could also have systematically higher levels of human capital, leading to superior performance. Such individual-level and organizational-level factors can often be mistaken for a calling orientation. Third, with the notable exception of Wrzesniewski et al. (1997), most studies analyze the behavior of only those workers who are suspected to have a calling orientation in isolation, therefore omitting counterfactuals or comparison groups that might have other orientations to work. The presence of such a counterfactual is crucial to identifying the differential influence of work orientation on resulting behavior. Therefore, while these studies are able to successfully show the association between a sense of calling and a number of important work-related outcomes, the causal direction of these claims is open to critique. Wrzesniewski (2010) summarizes the state-of-the-art by saying that “to date, much of the empirical research on callings has been correlational in design” and highlights the need for research that helps us unravel the causal relationship between work orientation and economic outcomes. Pratt et al. (2013) agree that “there is more theoretical than empirical support for [the literature’s] claims” and call for more research to investigate whether work ori2 entations vary in meaningfulness experienced by individuals. Further, while extant research on work orientation has been foundational in helping us describe the origins of a sense of calling and the experiences of those with a calling orientation, even theoretically, we have few explanations of how, if at all, work orientation should affect the economic activity of workers. Setting aside the empirical difficulties of measurement and causal identification of the impact of a calling orientation, we have little understanding of the mechanisms through which work orientation operates in practice. Under what conditions will workers with a calling orientation be seen to go beyond the call of duty or make sacrifices for their work? If one is to fully explain the causal link between a calling orientation and economic outcomes, then the lack of such theory is sorely felt. In this paper, I present results from a study that helps us make progress on both the empirical and theoretical link between work orientation and economic outcomes. The study is based on data from eight months of ethnographic fieldwork and an audit study where trained buyers are sent to purchase a standardized product from a community of wooden handicraft workers in an isolated town in southern India. Workers in this community can either be traders, who retail handicraft products with a job orientation or artisans, who both produce and retail the exact same products, with a calling orientation towards work. The setting helps us address problems in the existing literature by comparing two groups of similar workers- artisans and traders- who are engaged in the same activity of selling but vary significantly in their orientation to work, in a field-experimental setting that rules out alternative explanations, and by focusing on objective and comparable outcomes such as the prices offered by these different sellers. The setting has the additional advantage of studying a group of workers facing poverty, with no savings or alternative sources of income, offering perhaps the strongest test of whether a calling orientation has an impact on work outcomes. The qualitative data first demonstrates that artisans and traders subscribe to different work orientations and contrasts the quite different work practices associated with these different orientations. Building on recent work that examines the different elements of a calling orientation (Pratt et al., 2013), I document how artisans’ calling, in this case, stems from a “craftsmanship orientation” towards their work – an orientation where meaning is derived from a job done well, for its own sake. I then illustrate how this craftsmanship orientation leads artisans to develop an emotional connection 3 with the products that they painstakingly craft, drawing on the literature of psychological ownership (Pierce and Jussila, 2011), which examines the sense of possession that individuals experience over inanimate objects. The paper presents data to show that it is through this mechanism of psychological ownership that we can begin to make predictions about the impact of work orientations, namely that workers with a calling orientation, unlike those with a job orientation, will develop a preference to sell to discerning buyers who will take care of their products beyond the point of sale. In order to test this hypothesis that a sense of calling developed through a craftsmanship orientation manifests as a preference to sell to certain buyers, I conduct a field experiment that measures the impact of work orientation on the prices charged, bargaining intensity and transaction length of artisans and traders with three experimentally-manipulated auditor groups, varying in the taste that they demonstrate for craft products. The results show that artisans offer discerning and appreciative auditors significant discounts (upto 50%) for their products even when these buyers have a higher willingness-to-pay, such as in the case of foreigners. Traders, who have a more job orientation, on the other hand charge wealthier buyers higher prices, ignoring the auditors’ level of appreciation for the craft. Artisans are also seen to bargain less and engage longer with buyers who display a taste for their products, while traders bargain and engage equally with all buyers. The results, drawing on novel methods, point quite clearly to the causal link between work orientation and market outcomes. The study also helps us validate a novel theoretical mechanism that links one type of calling, i.e. a craftsmanship orientation, to market outcomes through the channel of psychological ownership over work-output. Finally, by uncovering the significant impact of a calling orientation and meaningful work in a developing country context where workers are extremely poor, the study informs the design of labor market institutions and jobs for low-income workers in these contexts. SETTING AND METHODS In the summer of 2011, I began fieldwork in various parts of India, conducting exploratory research on artisanal, craft markets. Over the course of 4 this exploration, I visited a number of clusters, each based around a single artistic technique (tie-and-dye, mirrorwork, block printing etc.) and discovered a theme that unified the labor markets in each of these settings. Each cluster was characterized by the presence of an artisanal community that crafted products by hand and sold them locally in their worksheds and a trading community that was responsible solely for the sale of these products to a wider market through retail establishments. While both sets of workers faced financial distress, artisans displayed a rather remarkable level of attachment towards their work, while traders had a more dispassionate outlook, where the products that they were selling were merely a means towards an end. Equipped with this inductively-developed understanding of artisanal and trading communities across India, I decided to focus in-depth on one cluster that was a part of my initial sample, the Channapatna wood and lacquerware craft cluster in southern India. Channapatna has a three-hundred year long tradition of wood and lacquerware handicrafts and about 10% of Channapatna’s 60,000 people are artisans and traders, involved in the production and retail of handmade articles like wooden jewelry and toys (Aziz, 1979). Apart from the richness of its craft heritage, as a research site, Channapatna offered several advantages. First, the artisans and traders in Channapatna were similar along a number of key dimensions, differing only in their work orientation. Members of both occupations are male, married, literate, usually in their mid-forties, belong to what are designated in India as backward castes,2 have families of about five or six people and have been practicing the occupation for about twenty years each, having inherited it from their family. A survey I conducted, whose findings are summarized in Table 1, points to a few small differences in these two groups. Traders are slightly more educated as compared to artisans, are more likely to be Hindu rather than Muslim and make about $160/month as compared to artisans who make $70/month.3 Both groups are however quite poor, buried in debt with an average loan size of $2,000 and live in rather shabby houses without tiled flooring or modern kitchens. INSERT TABLE 1 ABOUT HERE 2 includes Scheduled Castes (SC), Scheduled Tribes (ST) and Other Backward Castes (OBC) as defined by the Indian Constitution 3 While these gaps might be a cause for concern, Figures A.7, A.8 and A.9 establish that the final results are robust to these differences. 5 Channapatna also proved to be an excellent research setting in order to study artisans’ and traders’ selling behavior, the focus of this paper. It is a remote market located on a major highway between two large cities (Bangalore and Mysore) and receives local and international tourists, who purchase products from both traders and artisans (see a map in Figure 1). These transactions are typical of those in the informal economy: prices vary and are set by bargaining, relationships between buyers and sellers are onetime, devoid of reputational incentives (over 90% of artisans and traders reported never selling twice to the same direct customer), products are not branded or signed, sellers operate independently and there is no audience for any given sale. These conditions make it a favorable venue to situate an audit study to investigate the impact of work orientation on the selling process, where objective data on selling outcomes can be collected, selling outcomes can be compared with a couterfactual group and experimental variation and features of the setting help to rule out alternative explanations. INSERT FIGURE 1 ABOUT HERE Data Collection It was with these advantages in mind that I began fieldwork in Channapatna in January 2012. The data I collected over eight months is based on participant observation and twenty two semi-structured interviews, in addition to a formal survey and a field audit study. The qualitative data was collected inductively and experimental methods were used to test hypotheses emerging from an enriched understanding of the field. When I arrived in Channapatna, I stayed at a small lodge that offered proximity to both artisanal localities and trading establishments in town. By the end of my first week of fieldwork, the news of my presence in Channapatna had spread throughout the community, making access straightforward. Although the locals did not understand what research or academia meant, my interest in studying how handicrafts were made and sold and how work life was organized in Channapatna led them to trust me and, as a single woman in a predominantly male occupation, regard me as “harmless”. While, as a researcher from the US living amongst the artisans, I enjoyed some celebrity 6 status, with a train of children accompanying me wherever I went, being of India origin, I wasn’t completely foreign and blended in quite well. During the day, I would observe artisans and traders at work, paying special attention to their work practices and routines. For example, I would observe an artisan’s process of creating a product from start to finish and his attention to the minutest of details or a trader’s dealings with a large number of buyers of different kinds. In addition, in the evenings and over meals, artisans, traders and I would talk about the day’s work and events (Spradley, 1979). Apart from my ability to communicate in Hindi, I also developed a working understanding of Kannada (the state language) that allowed me to speak with a diverse set of local people over time. Artisans, in particular, were keen for their stories to be told and they seemed to open up to me, even more than to their colleagues or families because, unlike their family members, I asked about their work and took interest in their lives. I decided to carry a visible notebook from the beginning and let people see me jotting notes at all times, bringing out their expressive best, making them feel important each time I wrote down what they said. Finally, at the end of each day, I would make sure that I had documented all salient observations. I structured my time so that I was in the field for three days a week and spent the rest of my time in Bangalore typing up field notes, writing memos and making sense of the emerging data. This worked well because the time away from the field helped me identify puzzling observations that would inform the following week of my fieldwork. In this way, fieldwork included intensive participant observation of the artisans’ and traders’ dayto-day work, including observation of over sixty artisanal worksheds and visits to more than thirty trading establishments in the area. In addition to participant observation, I conducted twenty-two formal interviews with artisans and traders whom I had met in the course of fieldwork. My interview sample captured diversity in religion and size of establishment (Trost, 1986). I used these interviews to probe deeper into how artisans and traders understand and make sense of their work (Spradley and Baker, 1980). The semi-structured interviews were conducted in Hindi and lasted an average of one hour. Over the course of the interview I would cover most topics ranging from how the sellers had entered their occupation, how they experienced their work, what they liked and disliked about it, their daily routines and practices, the problems that they experienced and views about 7 their family, town, and occupation. Each interview was digitally recorded and after every interview, I recorded my impressions of the interviewee, his house, workplace and family members. Beyond these formal interviews, I also conducted over fifty informal ones in the midst of my participant observation (Barley and Kunda, 2001). Finally, in order to paint a complete quantitative picture of the artisans and traders in Channapatna, I conducted a survey of 52 artisans and 23 traders who would also participate in my field audit study.4 Conducted at the end of my fieldwork, these surveys collected (a) descriptive data including age, family information, household assets, religion, education and leisure activities, (b) workplace data such as tenure in the profession and machinery owned, (c) occupational data regarding work practices and norms such as time spent working, knowledge of other crafts, generational shifts and secondary occupations, (d) financial data about income, prices and expenditure on raw materials and (e) sales data by different customer groups. The nineteen page survey, translated into Kannada, was developed through an iterative process of referring to field notes, relevant literature (Aziz, 1979), and in consultation with handicraft and survey experts. The survey was administered by three trained surveyors who read the questions aloud, elicited responses and filled out the surveys on behalf of the respondents. An individual survey took about 30 minutes to complete and respondents were compensated for their time away from work. Data Analysis The open-ended data, comprising over 500 pages of fieldnotes and interview transcripts, was analyzed inductively using Atlas.ti. My inductive analysis (Glaser and Strauss, 1967) consisted of multiple readings of field notes and interview transcripts and extensive memo writing to decipher patterns in how artisans’ strong connection with their work manifested in their economic behavior. In my coding of the data, I associated passages of text with one or more mnemonic codes, labeling different ways in which artisans and traders talk about and understand their work. Some of these codes referred to uniqueness of products, artistic quality, product variety, creativity, atti4 There were 25 traders in the experimental sample but only 23 trader survey responses because one trader owned two shops and one trader refused the survey. 8 tude towards money, passion, selling behavior, sincerity towards work, work routines, competition, references to organization of work, control over work, reputation, work commitment and pride. In this way, the theorizing and analysis of the data proceeded iteratively. This process of data collection and inductive data analysis revealed the starkly different orientations that artisans and traders possessed towards their work, which is the focus of this study. The following sections draw on the qualitative data to describe the different orientations of workers in Channapatna and develop hypotheses for the field experiment to follow. JOB VERSUS CALLING ORIENTATION Everything is about money only, no? Why else would we work? People steal also, for money only. Without money, we have nothing. Even the cocks wouldn’t wake up in the morning without money. (Trader) I love this work...I put my heart and soul into it..it’s more than the money. (Artisan) Despite being similar along a number of dimensions, my fieldwork identified how traders and artisans differed in their orientations towards the work that they do, as hinted by these quotes. Using my participant observation and interviews, I describe the work practices and beliefs that show how traders and artisans approached their work in accordance with job and calling orientations respectively. Bellah et al. (1985) argued that individuals with a job orientation approach work as an instrumental activity to provide the resources necessary to pursue activities outside of work and to attain financial security. Traders in Channapatna imbibed such a job orientation, repeatedly talking about “money” and the “monetary gains from the sale of their products,” and explicitly articulated these monetary motivations. One trader confessed: We work for money. Even my wife understands this. My wife asks [me] for money and if I don’t give her enough, then she asks “what is the point [of] your work?”. So ultimately I work for 9 money. As Bellah et al. (1985) further noted, even though people with a job orientation do not approach work as an end in itself, they are likely to value hard work and individual initiative. Traders in Channapatna similarly valued the drive and skill to make a “good” sale. For traders, the ability to sell and make profits is linked not to the quality or innovation of the product, but the skills of the traders themselves, a classic indicator of a “job” orientation to work. Traders whom I interviewed spoke of this skill: Everyone can sell a good product. But us traders, we take pride in the ability to sell even a broken product. For that you need intelligence. You have money but no brain..I have brains but no money..who will do better? Me. Traders were skilled appraisers of handmade products that they sold and yet they demonstrated no attachment to the wood and lacquerware craft or appreciation for its intrinsic artistic quality. I repeatedly heard quotes like “of course we know if the wood is adequately seasoned” and “we’ve been seeing this for generations,” and yet, very rarely, did I hear a trader regard a particular product as being beautiful or well crafted. For traders, products were simply the means to a financial end: You have to just repackage the old designs as new designs. Each one is a different price. We just need the intelligence to make sales. Another signal of the lack of attachment that traders had towards the products that they sold was revealed in the relative ease with which they would stock cheaper, machine-made Chinese products and Chinese imitations right next to their Channapatna counterparts, a practice that infuriated artisans. I saw a number of trading establishments where traders had imported a large quantity of plastic Chinese products, responding to financial signals and market conditions in their decision-making rather than to an attachment with the craft. When explaining why more than half of the traders in Channapatna now stock products from China that “copy [local] designs”, “are of lower quality” but “look the same” a trader explained: Well the material costs are increasing, what do you do? The cost of production doesn’t get adjusted [on its own], so one solution is to buy cheaper raw materials..or else lower the quality of the 10 products. In contrast to the traders, artisans displayed a very different orientation towards the work of selling handicraft products. Artisans personified a calling orientation, validating (Bellah et al., 1985, pp 66)’s original and most-cited definition of a calling as something that “has value and meaning in itself, not just in the output or profit that results from it.” Here, work is pleasurable, fulfilling and intrinsically valued. Artisans, distinguished by the fact that they make and sell their own handicraft products, frequently described “loving their work” and “enjoying their work” and took pains to describe that they were not in this profession for money and that “if [they] wanted to make more money, [they] could have left a long time ago.” An artisan I spoke to made this orientation clear: This is my work. I get a lot of satisfaction...It’s not like other work, where you work for money or some other material thing. In fact, artisans in Channapatna took their calling orientation to an extreme, where they deified the very process of work and saw their craft as being sacred (Harrison et al., 2009). As one artisan said, “work is God for us.” To be clear, artisans developed devotion towards their craft in concert with doing their work, loving their work and becoming emotionally and spiritually attached to the work processes, rather than as a result of an invitation from God to a vocation through which God’s will is accomplished. Artisans worshiped their work, not as a service to God but for its own sake. In addition to interview data, a few workplace practices drove home the fact that artisans viewed their work to be a form of worship. First, artisans in Channapatna always worked without footwear despite the workplace being filled with sharp wood chips and sawdust, citing Indian custom where it is inappropriate to wear footwear in a place of worship like a temple or a mosque. The consequence of this was that there was not one workshed that I visited where I did not see an artisan with cuts and cracks on his feet from the sawdust and splinters of his workshed floor. When asked why they don’t wear footwear despite the safety implications, the idea seemed nonsensical to one artisan - “Do you wear chappals[footwear] inside the house of god?” he replied. In accordance with the workshed being a place of worship, artisans elevated their tools and machinery to the level of God, treating them like idols inside 11 a temple. While lax about both personal safety and standards of hygiene, artisans would go to extreme lengths to ensure that their machinery and tools are kept clean. I even observed several artisans worship their implements by offering flowers, lighting incense sticks and rigorously cleaning their machines. This practice was widespread even though machines were rendered dirty and covered in sawdust the minute they were turned on and the work commenced. A factory owner in the trade whom I spoke to, explained to me why this practice of multiple cleanings per day was so prevalent: People carry that kind of devoted attitude towards work. They clean their machinery daily. People treat their machinery as god. I see people cleaning their machines in the morning and evening, twice a day just like we pray twice a day. In order to understand how the sense of calling that we see in the case of Channapatna artisans might influence market outcomes, it becomes incumbent to understand it in the context of the typology first introduced in Pratt et al. (2013). According to this model, a sense of calling can arise from three distinct sources, through “doing well” leading to a craftsmanship orientation, through “doing good” leading to a serving orientation and through “doing with” leading to a kinship orientation. While a serving orientation is thought to characterize those who evaluate work as worth doing when it betters the lives of others or serves a cause, and a kinship orientation describes those who love work when it creates close, familial bonds with others, the focus of the craftsmanship orientation is on the work itself. Those with a craftsmanship orientation evaluate work as worth doing when it is done well for its own sake and conducted with skill and expertise, developing pride in the output and viewing it, not as a means to an end, but the end itself (Sennett, 2008; Crawford, 2009; Mills, 1956). It is this notion of calling through a “craftsmanship orientation” that best describes the artisans in Channapatna. This is reflected in the following quote by an artisan: Every time I make something creative,.. I feel blessed [to have] the skill and heritage I have gotten... [and] want to keep him [the product] close. While Pratt et al. (2013) guides us in broadly defining a “craftsmanship” orientation, the case of artisans in Channapatna can help us understand how a “craftsmanship” orientation might play out in practice, particularly with respect to how artisans come to view their hand-made products and how this 12 ultimately affects their selling behavior. In the next sections, I outline three broad themes that organize artisan’s outlook towards their work output and build on the literature of psychological ownership to develop hypotheses that will guide the design of the field audit study. Psychological Ownership over Products Through my fieldwork, I discovered that a key factor distinguishing artisans’ craftsmanship orientation is their relationship with the products that they make and sell. This relationship develops in the process of hand-crafting products. The production process consists of multiple stages, and at each stage the artisan must apply considerable skill and expertise. Through this process, artisans transform raw wood to finished products that have been carved, dyed and polished. Every stage of the production process calls for a painstaking application of skilled technical principles while at the same time allowing for a full expression of the artisan’s creative talents, leading to a strong bond with the final creative product. It is through this production process, illustrated in detail in Figure 2, that artisans seem to develop an emotional connection with every piece that they make. In fact, artisans come to treat their products like “babies”, part of their very embodied selves. As one artisan said: When I make the piece, I get attached to it. After seeing the design, I convert the shape on wood and after seeing the final product, I become happy. I [develop] affection for it. In order to understand how artisans’ strong relationship with their products affects their selling behavior, in this section, I draw on my fieldwork to organize artisans’ outlook towards their products under three broad themes: personal sacrifices, intimate knowledge of products and a desire for control. INSERT FIGURE 2 ABOUT HERE Personal Sacrifices. First, a craftsmanship orientation means that artisans subordinate themselves to the skilled practice of making handicraft products with their hands, putting all of their energy and time into making 13 sure that each piece turns out to be beautifully made. To this end, artisans make a number of personal sacrifices for the well-being of their work and products. For example, while it is recommended practice that woodworkers wear safety glasses5 and despite being handed eyeglasses for free by the government, artisans in Channapatna refuse to wear them. Glasses impede the artisan’s ability to pay close attention to minute details in their designs, and artisans happily trade artistry over their personal safety. One artisan I spoke to attested, I [...] get dust in my eyes and would fall sick. I put my finger in to take it out but sometimes cannot. Because they mix chemicals in [the wood], it’s painful ... [but] When I work on the lathe, if I put on the shades [eye glasses], I am unable to see the wood as carefully [as I want to]. So no one wears them. Intimate Knowledge of Products. Second, a craftsmanship orientation implies a deep knowledge and connection with the products that artisans make. Each product begins its life as a piece of wood, which must be turned on a motorized lathe in order to give it a desired shape, much like a potter does clay on a wheel. Once a piece of wood has been given the desired shape, artisans apply layers of lacquer that have been previously dyed using vegetable colors, then shade it expertly to produce a variety of effects, and finally use a locally available leaf to polish the product to obtain its distinctive glossy appearance. Thus, each piece that the artisan makes reflects his artistic vision and the unique choices that he has made during the production process and it is through these choices that artisans come to identify their own products. One artisan I spoke to described how he has intimate knowledge of every product that he makes: No two bangles look exactly the same due to the nature of the polishing and coloring process. I can identify my red bangle, among a sea of seemingly similar red bangles. 5 An online manual (http://woodworking.about.com/od/safetyfirst/tp/safetyRules.htm) states explicitly that “there is no time in the wood shop that you should be without your safety glasses. Your eyesight is too important to take chances.” 14 Control. A final feature that describes artisans’ relationship with their products is the desire to retain control over the product both during the production process and after it has been manufactured. Given the artistic nature of the process, one of the things that artisans enjoy the most is their ability to control the shape, color and design of each product. One artisan I spoke to said: We like to do our own pieces, create them with new designs, new colors and keep samples for our memory..we don’t want to become operators in a factory. Similarly, once a piece has been produced, artisans display a keen desire to control where it ends up. One practice that I observed was that artisans would keep products that they treasured most, enshrined in their closet. This is a widespread practice, as evidenced by the fact that on my first visit to any given artisan’s home, he invariably fetched a handful of unique self-made artifacts from a closed cupboard in his bedroom and introduced me to each one of them, like a parent would introduce their children. In this way, the sense of calling that artisans experience through their work is distinguished from other forms of calling by the unique relationship that artisans come to have with their products, characterized by the personal sacrifices that they make, deep knowledge of their products and a desire for control over work. Instead of talking about these ideas separately, the literature on psychological ownership offers a useful construct that helps us unify these themes. The construct of psychological ownership is based upon a sense of possession (“It is mine!”) as well as an affective relationship with the creation since calling something “mine” implies an emotional linkage (Pierce et al., 2001; Pratt and Dutton, 2000). Similar to the case of a calling developed through a craftsmanship orientation described above, psychological ownership is theorized to develop through three key mechanisms: (a) investment of self, (b) intimate knowledge and (c) control over the target that is being owned (Rouse, 2013; Pierce and Jussila, 2011; Pierce et al., 2003). The notion of investment of self is linked with the idea of psychological ownership because, as Locke (1690) argued, “we are likely to feel that we own that which we create, shape, or produce ... through our labor, we not only invest our time and physical effort but also our psychic energy into the product of that labor.” Scholars have also pointed to the willingness to assume personal risk or make personal sacrifice on behalf of a creation as an 15 important aspect of investing oneself in work (Pierce et al., 2003). Intimate knowledge of the product defines the notion that people are psychologically tied to things as a result of their active participation or association with those things (Weil, 1952). The most obvious and powerful means by which individuals develop such intimate knowledge of objects is by creating them (Pierce et al., 2003). Finally, control illustrates the idea that much like the parts of our body that we can control, material objects that can be controlled come to be regarded as part of the self, establishing a sense of psychological ownership (McClelland, 1951). Artisans, as described above, are also seen to invest their self in their products through personal sacrifice, develop intimate knowledge of the products that they craft and have a keen desire to control the production and sale process of their output. A sense of psychological ownership that develops through these three key mechanisms can therefore by used to describe artisans’ relationship with their products emanating from a craftsmanship orientation. More broadly, a sense of calling developed through a craftsmanship orientation can be summarized to be different from a kinship orientation or a serving orientation by the fact that workers with such an orientation evaluate work as worth doing when they have the opportunity to develop psychological ownership over their work. Traders, on the other hand, have a very different relationship with the products that they sell. In my fieldwork, I observed that even though traders had the capacity to distinguish well-made products, as evidenced by quotes like “we can see and tell apart the good wood [from the bad wood]” and “we can identify if the color has been applied well,” never did this expertise translate into an appreciation for exquisitely-crafted products. In fact, the reason that traders cared about the wood being well-seasoned or the color being smoothly-applied was to ensure that these products would still be in good condition by the time they reached the marketplace. Well-made products, for traders, were products that customers would find beautiful irrespective of their personal taste. Ultimately, every product was valued only according to the profit that it would fetch as demonstrated by this quote: Today I sell one thing, tomorrow another, its all the same. 16 Preference over Buyers in the Selling Process The previous section demonstrates that artisans with a calling orientation have psychological ownership over the products that they make and sell, while traders with a job orientation do not. This key difference manifests in the preferences that artisans and traders have over buyers to transact with in the selling process. The literature on psychological ownership highlights that creative workers, in particular, pay inordinate attention to the process of handing over their creative ideas or physical creations to coworkers or customers, once their work on the project culminates (Rouse, 2013). Such a “handoff” entails a shift in formal ownership, where the creator no longer has access to her creation and the receiver, to whom the creative work is handed over, assumes primary responsibility for the creation. Because workers are psychologically invested in their creations, having built a prolonged relationship with them, studies find that they care about the identity of the receiver, to whom the creative work is handed over (Rouse, 2013). Above all, workers need to trust that the receiver of their creation is competent before handing off their product (Colquitt et al., 2007). Similarly, workers need to know that the receiver will perform actions important to them, such as taking care of their “baby” after the point of handoff, irrespective of their ability to monitor the receiver (Mayer et al., 1995). In studying “handoffs” between different game developers in the video game industry, Rouse (2013) documents workers’ desire to give their products to “safe hands” for the safety and future development of their creations. In line with the predictions of this literature, I too observed that artisans’ psychological ownership over their products engenders a preference to sell to certain buyers. Artisans seem to care about selling to discerning buyers, who display a taste for craft work and will treat their products well even beyond the point of sale. In my observation, instead of behaving like traditional sellers who care primarily about the monetary consideration that they will receive for the product, artisans behaved more like adoption agencies who care acutely about the home that their babies will find. In other words, artisans derive value not just from finding a buyer for their product and charging the highest price possible, but also in ensuring that the right buyer is found, one who will take care of the product, appreciate its value, and 17 show it off in aesthetically pleasing ways. In particular my interviews suggest that artisans prefer to sell to buyers who display appreciation for their craft products and in the case of jewelry, those who will display it aesthetically. One artisan I spoke to confirmed: Our necklaces which are brightly colored look good with a white shirt and usually foreigners know to wear this combination; then the necklace shines. And yet another artisan described his preference for discerning buyers: I want my products to be displayed well in the customer’s home..I don’t want it[the product] to lie on a dusty shelf somewhere or in a closed cupboard...some buyers will put my product on a center table, that’s what i like. Traders, on the other hand, lacking a sense of psychological ownership in their job orientation of work, see products merely as a means to monetary gain, not paying attention to buyers’ appreciation for the product. As one trader said: I have to make money...what do you do? you just adjust prices and keep going..this is business. Building on discussion that explains how a craftsmanship approach to a calling leads artisans to develop psychological ownership over their products and a preference for certain kinds of buyers, this section hypothesizes how the presence or absence of such a strong sense of ownership over products regulates artisans’ and traders’ selling outcomes in this setting. In particular, I focus on the impact of psychological ownership developed through a calling orientation on three key outcomes of the selling process: prices, bargaining intensity and the length of transactions. These hypotheses derive from my experiences in the field. In order to better understand the selling behavior of artisans and traders in Channapatna, I engaged in several transactions purchasing handicraft jewelry, typically along with local women from the area. Despite looking Indian, I was distinguished from the local women by the handmade jewelry and handwoven clothing that I wore, bought from craft clusters around India. Further, even though I spoke the same language, it was evident that I was visiting from abroad from the way I spoke. Thus, I was differentiated from the local women along two 18 dimensions, level of discernment, through my appearance signaling a taste for craft products, and level of willingness-to-pay, by the fact that I was visiting this isolated town from far away. This difference between me and the local women manifested in three outcomes that I observed in the selling process, which guide the hypotheses developed below. Prices. The first key difference that I noticed was that artisans would often charge me lower prices than they would charge the local women, while the traders priced in line with more traditional notions of profit maximization – charging me a higher price than they did the locals. The supposition that artisans provided me with discounts due to the taste I portrayed for craft products was validated by my observation that on occasion, artisans offered discounts even to conspicuously wealthier buyers, like foreigners, for the opportunity to display the products in a foreign land in an aesthetic manner. My survey also indicates that when it came to pricing decisions, artisans paid attention not to buyer’s financial backgrounds but rather their intention of use for the product. When I asked them, “ Why do your prices vary?”, the most common responses were “depends on the customer’s interest” or“customer’s purpose from the purchase.” On the other hand, when it came to traders, I was consistently charged higher prices as compared to the local women. This was consistent with traders’ response to the same question of “ Why do your prices vary?”, where the most common responses were “different ability to pay among buyers” or “maximum profits.” These observations regarding artisans’ and traders’ selling preferences allows for the development of the following hypotheses: Hypothesis 1a: Artisans provide discounts to discerning buyers. Hypothesis 1b: Traders price according to buyers’ willingnessto-pay. Bargaining Intensity. While price is perhaps the most visible indicator of the differences between artisans’ and traders’ selling behavior, the nature of the bargaining process also offers important clues about how work orientation might affect market outcomes. When I bought products from artisans and tried to haggle for them, as is common in informal markets like Channapatna, artisans were often very willing to agree to the price I requested. Overall, 19 the transaction did not appear to be confrontational and artisans were more interested in making the sale to a discerning buyer, rather than achieving a particular monetary outcome. The idea is that an artisan who is keen to sell to a certain customer, not only charges lower prices, but is less interested in haggling over a monetary outcome. Consistent with this notion, the local women found artisans to be being quite hesitant in reducing the original price that they had been quoted. In this case, artisans had a more “take it or leave it” approach. Traders, as in the case of prices, did not seem to pay attention to a customers’ appreciation for the craft. Instead, traders bargained as much as possible, irrespective of whether the local women or I was making the purchase. Traders seemed to be focused on closing the sale while at the same time, extracting the highest possible price from each buyer. These observations motivate the following hypotheses: Hypothesis 2a: Artisans bargain less with discerning buyers. Hypothesis 2b: Traders bargain “hard” with all kinds of buyers. Transaction Length. Finally, sales transactions provide an opportunity to compare the selling behavior across traders and artisans by measuring not just the prices that sellers charge, or the intensity with which they bargain but also the time that they choose to spend with different kinds of buyers. In my initial fieldwork, I found that artisans differed quite starkly in their interactions with different kinds of buyers, while traders engaged uniformly across buyer groups. In particular, there was a significant difference in the length of transactions between me and the local women in the case of artisans. While shopping for products, artisans would often notice the handmade jewelry I wore and would ask about their origin: where the products were from and what materials and processes went into their production. I would also get asked where I was from and whether I had ever seen handicrafts of this nature before. Some artisans would further educate me about their painstaking production process. In contrast, when it came to the local women, the transactions proceeded swiftly, with the conversation being restricted to the purchase at hand. This pattern, where artisans had longer transactions with me as compared to the local women, did not seem to replicate itself in the transactions I had with local traders. They seemed indifferent to whether the local women or I wore craft jewelry or not, and it was never a topic of conversation, leading to transaction lengths that were more or less similar. 20 In the spirit of this observation, I hypothesize that: Hypothesis 3a: Artisans have longer transactions with discerning buyers. Hypothesis 3b: Traders have equally long transactions with different buyer types. Mechanism. Finally, if the mechanism through which work orientation affects market and selling outcomes is psychological ownership, then we expect artisans with a greater degree of psychological ownership to provide larger discounts. In particular, Pierce et al. (2003) have argued that an important feature that moderates levels of psychological ownership is how involved the creator is with the production process. This is because, the opportunity to do a whole and identifiable piece of work, from a logical beginning to a logical ending point, provides workers with the opportunity to become familiar with each of the tasks that are associated with completing a piece of work (Pierce et al., 2009; Hackman and Oldham, 1980). In addition, it affords the opportunity to see the sequencing of tasks, to come to understand those connections, and to develop an understanding of how the whole comes together. The implication is that, the deeper the knowledge of the target obtained through a living relationship with the object, the more psychological ownership will develop since: “the more information possessed about the target of ownership, the more things are felt thoroughly and deeply and in the process [the more] the self becomes attached to the object” (Pierce and Jussila, 2011, pp 80). My setting is unique, because every artisan is not involved in all nine steps of the production process, and in some cases, a few initial or final steps are delegated to other smaller artisans or family members. I exploit this feature to test the claim that a greater sense of psychological ownership is developed when artisans are involved in a larger number of processes, which leads to greater discounts to discerning buyers. Put simply, in testing the mechanism of psychological ownership through which a calling work orientation leads to greater discounts, I hypothesize that: Hypothesis 1d: Artisans’ discounts to discerning buyers increases with greater involvement in the number of processes they participate in. 21 A visual depiction of all the hypotheses can be seen in Figure 3. INSERT FIGURE 3 ABOUT HERE IMPACT OF WORK ORIENTATION ON SELLING OUTCOMES: A FIELD EXPERIMENT An audit study design, which has previously been used in the study of discrimination in sociology (Pager (2007), Pager and Quillian (2005), Correll et al. (2007)), is adapted to study work orientation and its impact on market outcomes in Channapatna. Specifically, the experiment is designed such that artisans’ psychological ownership over their products is activated by particular categories of trained buyers who project a more discerning orientation towards craft products and not by others, allowing me to test hypotheses about the impact of a calling and job orientation on prices, bargaining intensity and transaction length. Specifically, the experiment6 proceeded as follows. Six auditors were trained to be buyers of a standardized craft product, a pair of half-inch bangles, in Channapatna. Each auditor visited a sample of sellers in a randomly assigned order to make these purchases. The auditors negotiated for a price, according to a prescribed bargaining script, and recorded price and other information on a form after each sale. The experiment, which was conducted over a two week period in the middle of May 2012, coincided with a large cricket tournament in the area. This meant that there were more tourists than usual visiting Channapatna and the auditors for the experiment did not stand out. This section describes the design and implementation of the experiment in detail. Product. The “Channapatna Bangle”, of half-inch width, was the standardized craft product chosen for the purpose of the experiment (See Figure 4 for a picture). This bangle is ubiquitous and widely produced and sold in Channapatna owing to its current popularity in Indian fashion. In addition, 6 A big thank you to Tavneet Suri for all her help with the experiment. Schoar and Iyer (2013) was also instructive in the design of the audit study. 22 this bangle is standardized in its cost of production and yet, offers an avenue for creative expression and craftsmanship through differential color, pattern, and design. Since these bangles are usually worn in pairs, each auditor also purchased a pair of Channapatna Bangles of half inch width. INSERT FIGURE 4 ABOUT HERE Sellers. For this experiment, a sample of 77 sellers, 52 artisans and 25 traders, was created from Channapatna’s population of over 5,000 sellers.7 See Figure 5 for a GPS plot of the chosen artisans and traders. In choosing the sellers for the experiment, only those artisans and traders who had ample experience making and selling half-inch bangles, and who had sufficient stock of this product were considered. Further restrictions were imposed on the sample to select artisans and traders who were at least 500 meters away from other sellers in the sample. The final sample consisted of sellers across 8 localities of Channapatna. The sellers in the sample were divided into 20 groups, each consisting of 3 to 4 artisans or traders, based on geographical proximity. This aided in devising a schedule for the experiment. INSERT FIGURE 5 ABOUT HERE Auditors and Treatment. Six auditors were hired for this experiment, four Indians and two foreigners. The auditors were all women in their early twenties with 12 to 14 years of education. They came from families of 4 to 6 members and had 2 to 3 siblings each. They also looked alike, with similar height and weight. None of them had been to Channapatna before, or had prior familiarity with the craft work there. The six auditors were similar except for their “look,” or material presentation, which constituted the key aspect of the treatment for the experiment. My fieldwork indicated that artisans preferred clients who were appreciative of their products, especially if they had traveled from a foreign country or if they were discerning Indian appraisers of the craft. Accordingly, the auditors were 7 My participant observation and interviews encompassed a wider sample of artisans and traders, which allowed me to establish that the sellers in my sample seem similar to other sellers in Channapatna. 23 assigned to three experimental roles, “International” foreign auditors who conveyed a high level of appreciation (apart from having a high willingnessto-pay), “Indian-Craft,” Indian auditors wearing several handmade products thereby conveying taste for craft work and ”Indian-Baseline,” a control group that consisted of Indian buyers who dressed like the locals in the area. In order to enact these roles, I hired four Indian auditors (who were randomly assigned to the Indian experimental categories) and two international auditors. The Indian auditors I hired were all from small towns about 450 kilometers away from Channapatna in the same state of Karnataka and spoke the same language - Kannada- as the locals in Channapatna. Although they conducted their transactions in Kannada, they were outsiders to Channapatna, which was evident from the accent with which they spoke the language. They dressed traditionally in the Indian attire of salwar-kameez, a pair of loose trousers, a long tunic and a scarf, just like the average female Indian tourist in Channapatna. However, the two auditors in “Indian-Craft” were directed to display an aesthetic taste for handmade, craft products in their fashion choices instead of the more common synthetic alternatives. Hence, they wore several craft items including handmade terracotta earrings, a handcrafted metal necklace, salwar-kameezes made from handwoven cotton and they each carried handbags woven out of natural fiber. These products were specifically from craft clusters in other parts of India, not from Channapatna, and were meant to indicate appreciation for craft work rather than familiarity with Channapatna’s products. On the other hand, the two auditors in “Indian-Baseline” dressed like they normally would, with plastic jewelry, salwar-kameezes made of polyester material and they carried synthetic handbags. I also hired two international auditors for this experiment who were assigned to the “International” role. These auditors were from Thailand and Mauritius, spoke neither Kannada nor Hindi, and therefore conducted their transactions in English.They looked discernibly foreign, with their lighter skin color, distinct features and their Western clothing consisting of dresses, skirts, pants and shorts, and as such, were similar to the average international tourist shopping in Channapatna. The six auditors were each paid Rs.10,000($200) for this assignment lasting two weeks. This is equivalent to the wage for an entry-level job at a call center in India. 24 Training. Prior to implementing the experiment, the auditors spent three days in training, consisting of classroom presentations and discussion, practical observation of the area, followed by a pilot exercise. The first part of the training involved introducing the auditors to the setup of the experiment without revealing the research questions of interest. This included educating the auditors about the wood and lacquerware products of Channapatna and specifically, the product that they would purchase in the experimenta pair of bangles of half inch width. This would ensure that the auditors could easily identify the Channapatna bangle at the time of sale. Auditors were instructed to transact with the trader or artisan himself, not a relative or wife, and similarly, not to purchase from sellers not listed in the sample. Subsequently, the auditors were put through a range of role-playing exercises to practice acting like a buyer and to achieve consistency in their acting. This portion of the training involved memorizing a script describing what the auditor was doing in Channapatna, namely making a stop en route to the tourist city of Mysore, learning to deny any prior exposure to the wood and lacquer craft and learning the bargaining routine to haggle for the products. The final element of the classroom training entailed practicing how to fill out a “Transaction Form”, a custom form designed to capture prices and other details of the transaction, after each purchase. The classroom education was followed by field training in Channapatna. Here, a mini-van was rented to drive through Channapatna and build familiarity with the area. Auditors were given detailed maps marking the locations of every artisan and trader in the sample and were shown the different localities of the town so that they could find their way alone during the actual experiment. Finally, a pilot experiment was conducted in a nearby town called Yarabnagar that makes carved wooden products. Here, auditors visited a small sample of sellers to rehearse their script, make purchases in accordance with the bargaining protocol and practice filling out the transaction form after each sale. Randomization. The schedule for the experiment, governing the transactions that an auditor would conduct on a given day, was created using randomization code. Each day of the experiment was divided into two time slots, a morning slot and an evening slot. As established, the actors for the experiment comprised 6 auditors and 20 seller groups, each consisting of 3 25 to 4 sellers each. The goal of the randomization code was to assign one auditor and one group of sellers to each time slot such that: 1) each auditor visits a given seller group only once, 2) two auditors do not visit the same seller group in the same time-slot, 3) sellers do not receive auditor visits on consecutive days and 4) a seller does not receive more than 3 auditor visits in a week.8 Imposing these constraints on the randomization code mitigated concerns of “demand effects” or stockouts9 and ensured that sellers in the sample were not bombarded by auditors. Buying Process. On any given day of the experiment, each auditor was given her schedule for the day and money for the transactions. The auditors would set out at 9AM and commute within Channapatna independently by rickshaw or bus. On reaching an assigned seller, the auditor would leisurely survey the seller’s products before choosing a pair of half-inch bangles for purchase. The auditor would then get the seller’s attention and ask the price of the bangles. Upon hearing the seller’s initial price, in the first round of bargaining, the auditor would offer half of this quoted price. If the seller did not accept this offer, he would suggest a second price to which the auditor would raise her initial price by Rs.2 in this second round of bargaining. If the seller did not accept this offer as well, the interaction would repeat with the seller offering yet another price and the auditor raising her price by Rs.2 again in the third round of bargaining. The bargaining would cease at this point and the auditor would pay the final price demanded by the seller. This bargaining routine, modeled on typical informal market behavior in India (Schoar and Iyer, 2013), was standardized across all transactions conducted in the experiment. After completing a transaction, the auditor would leave the seller’s premises and find an isolated spot to fill out the transaction form before proceeding to the next seller. If a transaction could not be performed due to unavailability of the seller, the auditor took note of this and revisited the seller on a different day. At the end of the day, auditors would submit their purchased bangles and completed transaction forms to me. Each pair of bangles was labeled with 8 Survey data indicate that sellers retail 70 bangles per week on average and therefore, the 3 auditor purchases per week of the experiment is unlikely to be noticed by the sellers or affect their behavior. 9 A situation where the sellers price differently because of depleting stock. 26 both seller and auditor information. The forms were also checked for missing values and data entry was completed the very same day. RESULTS In total, the experiment comprised 455 audit visits to 52 artisans and 25 traders conducted by 6 auditors. Table 2 shows the distribution of sales transactions by each seller and buyer category. The table reports that each of the three auditor categories conducted between 100 and 103 transactions with artisans and 50 transactions with traders over the course of the experiment, thus making for a fairly balanced overall distribution in exposure to the experimental treatment, namely the 3 “looks”.10 Data from the experiment also confirm that the ordering of auditor visits to any seller was random, as planned. For example, a seller had a roughly 33% chance of receiving an International, a 31% chance of receiving an Indian-Baseline and a 35% chance of receiving an Indian-Craft auditor for their first transaction. In unreported analyses, I also verified that the mean prices offered by sellers on any given day of the experiment displayed no discernible trend over time. To be cautious however, all reported regressions include date fixed effects allowing me to control for any possible time trends in the results. INSERT TABLE 2 ABOUT HERE To confirm that randomization was conducted correctly in the experiment, Table 3 displays an OLS regression of the auditor categories on 3 baseline transaction-level characteristics that could have an impact on price. We would not expect any significant heterogeneity across auditor categories in these baseline measurements if the randomization was done correctly, since treatment assignment was random and orthogonal to these baseline characteristics. The first characteristic that the auditors took note of was whether there was electricity at the time of sale. This is relevant because Channapatna faces unpredictable power outages of 3 to 4 hours a day. Second, the 10 While all 150 transactions with traders were conducted as planned, there were 7 planned transactions with artisans that could not be completed due to unavailability of the artisans in question. However, these incomplete transactions are also distributed evenly across the 3 auditor categories. 27 auditors estimated the remaining stock of half inch bangles a seller had, after they had made their purchase; this stock varied widely based on the seller’s day-to-day business. And third, the auditors indicated whether the seller’s spouse was present while the transaction was being conducted; the spouse was typically present when there were other errands to complete such as cleaning or arranging the items on display. The table reports that electricity was available in about 58% of the 455 transactions, that a seller typically had between 12 and 17 pieces of half-inch bangles left and the seller’s spouse was present during 22% of the transactions. However, as predicted, these baseline transaction-level characteristics do not vary significantly across the auditor categories, providing evidence that randomization was implemented according to plan. INSERT TABLE 3 ABOUT HERE Prices. This section first explores the variation in initial prices offered by artisans and traders to the auditors displaying different looks. Initial price is the price first quoted by the seller. Initial price is a particularly useful measure because, even though the bargaining protocol was standardized, initial prices reflect the sellers’ appraisal of each buyer role and should be independent of the way in which bargaining proceeded.11 Figure 6 illustrates the distribution of initial prices (in Indian Rupees, where approximately Rs.50=$1) that auditors received when transacting with artisans and traders. The graph to the left shows that the mean initial price offered by artisans is Rs.27.91, but that there is significant variation in this price: it ranges from Rs.0 to Rs.60 and its standard deviation is 14.79. The graph to the right illustrates the corresponding initial prices that auditors received from traders. The mean initial price offered by traders is considerably higher than that of artisans at Rs.36.35. Traders’ initial prices also vary widely, ranging from Rs.18 to Rs.75 with a standard deviation of 10.88. Note that, unlike artisans, traders do not offer any zero initial prices. 11 The focus of the analyses will mostly be on initial prices, but the results are robust to using final prices also. Final price is the price at which the product is eventually purchased after auditors engage in three rounds of standardized bargaining. On average, the buyers receive about a 15% price reduction after bargaining with artisans and a 10% price reduction from traders. 28 INSERT FIGURE 6 ABOUT HERE In order to investigate whether average prices differ significantly between auditor groups, for both artisans and traders, Figure 7 plots the mean initial price offered to the three auditor groups, with error bars representing a 95% confidence interval around the mean. Consistent with the predictions of Hypothesis 1a and 1b, the figure indicates that artisans charge the lowest mean prices to Indian-Craft (Rs.18.63) followed by International (Rs.28.15) and Indian-Baseline (Rs.36.87). Artisans seem to distinguish between these three groups and charge them prices in accordance with their displayed taste and appreciation for handmade products. It is interesting to note that the discounts offered by artisans are greater for Indian-Craft as compared to the International auditors. This suggests, perhaps, that artisans are most responsive to the customers who not only signal an appreciation for craft (like the international tourists), but who demonstrate a physical and monetary commitment to handicraft products by purchasing and wearing them. Traders, on the other hand, price quite differently. They charge the highest mean price (Rs.45.30) to the International category, which is in accordance with pricing based on customers’ willingness-to-pay. There seems to be no significant difference in the traders’ prices offered to Indian-Craft (Rs.32.72) and Indian-Baseline (Rs.31.04), suggesting that traders disregard the handicraft products worn by the Indian-Craft category in making pricing decisions. In this way, traders seem to price in accordance with customers’ willingness-topay for the products being sold rather than appreciation for the craft. INSERT FIGURE 7 ABOUT HERE Having explored the variation in prices offered to different buyer groups, Table 4 tests whether log initial and final prices (excluding zero prices) offered to different auditor groups are statistically different from each other. The OLS models I employ also allow me to account for error structures robust to group-level covariance,12 add date and time-level fixed effects and log specifications allow me to interpret differences in percent-terms. Specifically, I 12 Standard errors are clustered at the level of sellers (77 clusters); this allows for potential correlation in the error terms within sellers and uses repeated observations on transactions with the same seller to estimate standard errors robust to this problem (Angrist and Pischke, 2008). Standard errors are clustered in a similar way for all the models reported in this study. 29 regress price outcomes on three dummy variables, one for each buyer category, Indian-Baseline, Indian-Craft and International, using separate models for artisans and traders, focusing on the variation within sellers in prices offered to different buyer categories. Column (1) of the table regresses initial price on Indian-Craft and International, using Indian-Baseline as the omitted category. The constant term gives us the estimated mean initial price of the omitted category, Indian-Baseline, conditional on covariates, while the coefficients on the Indian-Craft and International dummy variables provide estimates of the difference in mean prices between Indian-Baseline and the respective category. The table reports that artisans offer a 50% discount to Indian-Craft and a 27% discount to International. A post-estimation F-test is performed to test the equality of the Indian-Craft and International coefficients, which shows that the difference in these coefficients for artisans is also statistically significant. This regression is then repeated in column (2) for traders. The table reports that traders charge almost the same price to Indian-Craft as they charge to Indian-Baseline, but charge 45% higher prices to International as compared to Indian-Baseline. The post-estimation F-test indicates that the difference in prices between Indian-Craft and International is also statistically significant. Columns (3) and (4) repeat these regression for final prices and find the same pattern in price discrimination for both artisans and traders, confirming that the pricing behavior is resistant to bargaining. Overall, the regressions provide evidence to support Hypotheses 1a and 1b that artisans offer discounts to those groups of buyers who display a taste for handicraft products in a way that traders, who focus on the buyers’ willingness-to-pay, do not.13 INSERT TABLE 4 ABOUT HERE It is further remarkable in my setting, however, that in some cases artisans will offer bangles at a zero price to some auditors, even before they have bargained for it. Artisans seem so eager to sell to a particular buyer that they are willing to part with their product without any financial returns, serving perhaps as the strongest confirmation of my theory. Table 5 explores patterns in the zero prices offered by artisans to test whether there are signif13 The results are robust to using raw prices, dropping fixed effects and including zero prices, indicating that the results are not likely to be driven by differences in transactionlevel characteristics, but by the role played by the auditor herself. 30 icant differences across the 3 auditor categories. This is important because more than 10% of initial prices quoted and 12% of final sale prices were zero for artisans.14 To test whether these zero prices differ across the auditor categories, I created dummy variables for both initial and final price indicating whether a given price is equal to 0. Table 5 reports the OLS regression on zero prices offered by artisans using a specification similar to Table 4. Column (1) shows that Indian-Craft are 18% more likely to receive a zero initial price than Indian-Baseline, and International auditors are 7% more likely to receive a zero initial price than Indian-Baseline. In addition, a postestimation F-test rules out equality of the coefficients of Indian-Craft and International. Column (2) further shows that the results are even stronger for a zero final price. This demonstrates that the broader price discrimination pattern observed amongst artisans, namely offering the lowest prices to Indian-Craft and the next lowest prices to International, also holds for zero prices, with Indian-Craft having the highest likelihood of being offered a zero price followed by the International category. INSERT TABLE 5 ABOUT HERE Bargaining Rounds and Transaction Length. As predicted in Hypothesis 2a, if it is indeed true that artisans respond differently to the presence of more appreciative customers as a result of their strong identification with work, they can also be expected to haggle less with such clients. I collected detailed data on not only the prices charged, but also the number of bargaining rounds that took place before the transaction was completed. Table 6 compares the total number of bargaining rounds required to complete a purchase (where 0 is the minimum and 3 is the maximum) across auditor groups using an OLS model. Column (1) finds that artisans engage in fewer bargaining rounds with both International auditors (2.05 rounds) and IndianCraft auditors (2.05 rounds) as compared to Indian-Baseline auditors (2.67 rounds). Column (2) shows that traders typically bargain more than artisans, but do not seem to bargain differently with different auditor groups in any way i.e. they bargain “hard” with all buyers. This provides support for hypotheses 2a and 2b. 14 Note that there are more zero final prices than zero initial prices, implying that not all zero prices are prior to bargaining. 31 INSERT TABLE 6 ABOUT HERE In order to investigate hypotheses 3a and 3b that measure the impact of work orientation on transaction length, I use data collected by the auditors who were trained to use a stopwatch to measure the time taken for each transaction to complete. Columns (3) and (4) of Table 6 report results from an OLS regression that regresses transaction length on each auditor category. Column (3) shows that even though artisans bargain significantly less with Indian-Craft and International, as compared to Indian-Baseline, they engage in longer transactions with these two groups of discerning buyers. Since auditors were rigorously trained to use a standardized script in conducting their transactions, differences in transaction length stemmed solely from variation in seller-initiated conversations for different auditor groups. Anecdotal descriptions confirmed that artisans asked Indian-Craft auditors questions about the handmade products they wore as part of the experimental treatment and asked International auditors questions about their home country. This resulted in artisans having longer transactions with Indian-Craft and International auditors, in an attempt to get to know the discerning buyers who were taking home their “babies.” The estimates reported in column (4) show that traders display no difference in transaction lengths between the different auditor categories, confirming that traders only care about buyers’ willingness-to-pay. These results confirm hypotheses 3a and 3b. Mechanism. Hypothesis 4 predicts that as artisans engage in a greater number of work processes, their craftsmanship orientation will lead them to have more psychological ownership and give greater discounts to discerning buyers. In order to investigate such an effect, I use survey data that measures artisans’ involvement in each of the the nine different stages that comprise the overall production process.15 I find that on average artisans engage in about 6.4 work processes with a standard deviation of 2.25. Figure 8 graphs the impact of involvement in work processes using estimates from an OLS model. The underlying model regresses the initial price offered on dummy variables for auditor categories, as before, and also includes Processes as an independent variable. Processes is a continuous variable ranging 15 These processes include cutting the wood, seasoning the wood, turning the wood on a lathe, assembling the products, polishing, lacquering, painting, finishing and selling. 32 from 1-9 measuring the number of work processes that a given artisan engages in. Crucially, in addition to these main effects, the regression includes interactions between auditor category and Processes. This interaction term estimates how prices offered change for different buyer groups as an artisan engages in a greater number of work processes. These estimates can then be used to calculate the predicted price than an auditor will receive from an artisan depending on the auditor’s category and the artisan’s Processes. A convenient way of displaying these predictions obtained from regression coefficients is using a “marginal plot,” as shown in Figure 8. INSERT FIGURE 8 ABOUT HERE Figure 8 makes clear that the difference in prices offered to the three buyer groups increases with greater work process involvement. For example, while artisans who are engaged in only one work process are seen to provide a discount of about 40.5% to Indian-Craft auditors as compared to IndianBaseline auditors, this discount increases to 53% for artisans engaged in all nine work processes. This finding is in accordance with the theory that as artisans engage in more work processes, they have more psychological ownership over their products, and are therefore willing to give greater discounts to transact with discerning customers. This confirms Hypothesis 4 that psychological ownership seems to be an important mechanism for the artisans’ pricing pattern described in this study. Appendix A describes results from a number of robustness exercises performed to investigate the role of alternative explanations in explaining artisans’ distinctive pricing behavior. These include age, work tenure, presence of neighboring sellers, income, distance from highway, business with traders, visits to Bangalore, education and religion of artisans. DISCUSSION This paper presents an in-depth examination of artisans and traders in a handicraft cluster in India to illustrate the significant impact of their different work orientations. While scholars have long described the differences between a job versus calling orientation towards work (Bellah et al., 1985; Wrzesniewski et al., 1997) and have shown that a calling orientation is associated 33 with distinct work related attitudes and behaviors (Bunderson and Thompson, 2009; Dobrow, 2006; Wrzesniewski et al., 1997), this paper presents causal evidence that work orientation significantly impacts economic outcomes. In particular, the paper empirically establishes this causal claim by collecting data on objective variables, using a field-experimental design to rule out alternative explanations and by presenting a counterfactual analysis through the study of two otherwise similar occupations that differ chiefly in their orientation towards work. Additionally, the paper uncovers the mechanism of psychological ownership that helps us link a sense of calling developed through a craftsmanship orientation with economic outcomes. Data from participant observation, interviews and a survey is used to characterize the lives and work practices of artisans, who see their handicraft work as a calling, and traders, who retail craft products as a job. Artisans are shown to have a calling motivated by a craftsmanship orientation, which results in a unique relationship of psychological ownership with the handicraft products they make, characterized by personal sacrifices, intimate knowledge of the products and a desire for control over the work(Pratt et al., 2013; Pierce et al., 2003). Traders, in contrast, are shown to value products only for the profits that they bring. These differing relationships with the products engender different preferences over buyers to sell to: artisans prefer selling to discerning customers whereas traders care solely about buyers’ willingnessto-pay. Building on this theory, hypotheses are developed, which describe the impact of work orientation on selling outcomes including prices charged, bargaining intensity and transaction length. A field experiment, where trained buyers, varying in the taste that they display for craft work, are sent to purchase identical products is then used to compare the differences in traders’ and artisans’ selling behavior. The results validate the hypotheses: artisans provide significant discounts (up to 50%), bargain with lower intensity and engage in longer transactions with discerning customers while traders charge wealthier customers higher prices and do not differ along other selling outcomes. Broadly, the qualitative data and the results of the field audit study paint a picture supporting the initial claim that work orientations have a significant impact on economic decisions of workers. Although the focus of this paper is on selling outcomes, one could reasonably expect that having that a craftsmanship-motivated calling influences other, more complex market outcomes too. This study offers a theoretical and empirical framework to understand these other outcomes as well. For example, 34 musicians often have to make decisions regarding where to perform. While it would seem advantageous to pick large concert halls that would generate more revenue from ticket sales, one can imagine that musicians who have psychological ownership over their carefully-prepared concerts also consider the likely quality of the audience- in particular, the ability to appreciate nuances in the compositions- in picking performance venues (Smith, 2009; FonsecaWollheim, 2012). Similarly, workers often also have to make investment and financing choices, where again a close association with one’s output can influence the decision made. For instance, filmmakers, in choosing a production company for their film, are likely to consider which companies will preserve the artistic vision of their painstakingly-crafted films, and not just regard the monetary rents from this contract (Lovell, 2012). Further, a wide range of workers additionally have to choose what intellectual property (IP) or licensing scheme they would like to use to protect their work (Murray, 2010). In these decisions too, it is reasonable to expect that work orientation plays a role in the choice of IP regime: Silbey (2014) reports that recouping investment through IP arose infrequently in interviews with workers, in contrast to abundant discussion about the value and meaning of their work. Methodologically, this paper makes several contributions. First, a randomized field experiment, which is a common tool in the study of discrimination in sociology and in development economics, is used to precisely measure the impact of different buyer groups on prices and selling outcomes, while controlling for other factors that are not under study. Second, the paper shows how conditions common in developing country contexts (like informal price setting or the prevalence of craft communities) can be exploited to illuminate general theoretical questions of interest in meaning of work. Third, this study demonstrates how one might combine field experimental evidence, which provides a causal but narrow outcome measure, with rich ethnographic analysis to ground the phenomenon and investigate mechanisms. Finally, this research informs the design of labor-market institutions in developing economies by highlighting the need for more sophisticated models of how low-income workers make sense of their work, beyond theories rooted solely in financial interests. Some scholarship in the meaning of work literature has argued that the tradeoff between meaning and money is not relevant to workers facing poverty, for whom the economic value of work becomes more salient (Leana et al., 2012; Brief and Nord, 1990). These scholars have argued that poverty is a classic “strong situation” which can constrain 35 the meanings to be found in work practices such that individuals suffering economic distress tend to deemphasize the latent value of work (i.e. selffulfillment) in favor of the manifest value of work (e.g., monetary rewards) (Leana et al., 2012). However, artisans in Channapatna, despite being extremely poor, have significant psychological ownership over their handicraft products which motivates them to make monetary sacrifices for the future of their products. This suggests that an understanding of work that solely emphasizes its instrumental benefits is insufficient to design labor market institutions and human resource policies among the working poor. References Angrist, Joshua D. and Jorn-Steffen Pischke. 2008. Mostly harmless econometrics: An empiricist’s companion. Princeton University Press. Aziz, A. 1979. “Socio-economic Conditions of wooden toys and lacquerware artisans of Channapatna Taluk.” Report from the Institute of Social and Economic Change, Bangalore . Barley, S. R. and G. Kunda. 2001. “Bringing Work back in.” Organization Science 12:76—95. Baumeister, Roy F. 1991. Meanings of life. Guilford Press. Bellah, Robert N., Richard Madsen, William M. Sullivan, Ann Swidler, and Steven M. Tipton. 1985. “Habits of the heart. Berkeley.” University of California Press.” Academic Ability and Performance.” Social Psychology Quarterly 49:103—109. Brief, Arthur P. and Walter R. Nord. 1990. Meanings of occupational work: A collection of essays. Free Press. Bunderson, J. Stuart and Jeffery A. Thompson. 2009. “The call of the wild: Zookeepers, callings, and the double-edged sword of deeply meaningful work.” Administrative Science Quarterly 54:32—57. Colquitt, Jason A., Brent A. Scott, and Jeffery A. LePine. 2007. “Trust, trustworthiness, and trust propensity: a meta-analytic test of their unique relationships with risk taking and job performance.” Journal of applied psychology 92:909. 36 Correll, Shelley J., Stephen Benard, and In Paik. 2007. “Getting a Job: Is There a Motherhood Penalty?” American Journal of Sociology 112:1297— 1339. Crawford, Matthew B. 2009. Shop class as soulcraft: An inquiry into the value of work . Penguin. com. Dik, Bryan J. and Ryan D. Duffy. 2009. “Calling and Vocation at Work Definitions and Prospects for Research and Practice.” The Counseling Psychologist 37:424—450. Dobrow, Shoshana R. 2006. Having a calling: A longitudinal study of young musicians. Ph.D. thesis, Harvard University. Fonseca-Wollheim, Corinna. 2012. “Schoenberg’s Version of Mahler’s ’Song of the Earth’.” The New York Times . Glaser, B. G. and A. L. Strauss. 1967. The discovery of grounded theory: Strategies for qualitative research. Aldine de Gruyter. Hackman, J. Richard and Greg R. Oldham. 1980. Work redesign. Reading: Addison-Wesley. Hall, Douglas T. and Dawn E. Chandler. 2005. “Psychological success: When the career is a calling.” Journal of Organizational Behavior 26:155—176. Hardy, Lee. 1990. The fabric of this world: Inquiries into calling, career choice, and the design of human work . Wm. B. Eerdmans Publishing. Harrison, Spencer H., Blake E. Ashforth, and Kevin G. Corley. 2009. “Organizational sacralization and sacrilege.” Research in organizational behavior 29:225—254. Leana, Carrie R., Vikas Mittal, and Emily Stiehl. 2012. “PERSPECTIVEOrganizational Behavior and the Working Poor.” Organization Science 23:888—906. Locke, John. 1690. “An Essay Concerning Human Understanding.” Book IV p. 328. Lovell, Joel. 2012. “Kenneth Lonergan’s Thwarted Masterpiece.” The New York Times . 37 Mayer, Roger C., James H. Davis, and F. David Schoorman. 1995. “An integrative model of organizational trust.” Academy of management review 20:709—734. McClelland, D. C. 1951. Personality. New York: William Sloane Associates. Dryden Press. Mills, Charles Wright. 1956. The power elite. New York: Oxford University Press. Murray, Fiona. 2010. “The Oncomouse That Roared: Hybrid Exchange Strategies as a Source of Distinction at the Boundary of Overlapping Institutions.” American Journal of Sociology 116:341—388. Pager, Devah. 2007. “The use of field experiments for studies of employment discrimination: Contributions, critiques, and directions for the future.” The Annals of the American Academy of Political and Social Science 609:104—133. Pager, Devah and Lincoln Quillian. 2005. “Walking the talk? What employers say versus what they do.” American Sociological Review 70:355—380. Pierce, Jon Lepley and Iiro Jussila. 2011. Psychological Ownership and the Organizational Context: Theory, Research Evidence, and Application. Edward Elgar Publishing. Pierce, Jon L., Iiro Jussila, and Anne Cummings. 2009. “Psychological ownership within the job design context: Revision of the job characteristics model.” Journal of Organizational Behavior 30:477—496. Pierce, Jon L., Tatiana Kostova, and Kurt T. Dirks. 2001. “Toward a theory of psychological ownership in organizations.” Academy of Management Review 26:298—310. Pierce, Jon L., Tatiana Kostova, and Kurt T. Dirks. 2003. “The state of psychological ownership: Integrating and extending a century of research.” Review of general psychology 7:84. Pratt, Michael G. and Blake E. Ashforth. 2003. “Fostering meaningfulness in working and at work.” Positive organizational scholarship: Foundations of a new discipline pp. 309—327. Pratt, M. G. and J. E. Dutton. 2000. Owning up or opting out: The role 38 of emotions and identities in issue ownership. Emotions in the workplace. NM Ashkanasy, CE Hartel and WJ Zerbe. London, Quorum. Pratt, M. G., C. Pradies, and D. Lepisto. 2013. “Doing well, doing good, and doing with: Organizational practices for effectively cultivating meaningful work.” Purpose and meaning in the workplace. Washington, DC: APA Books (forthcoming) . Rosso, Brent D., Kathryn H. Dekas, and Amy Wrzesniewski. 2010. “On the meaning of work: A theoretical integration and review.” Research in Organizational Behavior 30:91—127. Rouse, Elizabeth. 2013. Kill Your Darlings? Experiencing, Maintaining and Changing Psychological Ownership in Creative Work . Ph.D. thesis, Boston College, Boston. Schoar, Antoinette and Rajkamal Iyer. 2013. “The Role of Culture in Financial Negotiations: Evidence from an Audit Study.” International Differences in Entrepreneurship, NBER Conference Volume . Sennett, Richard. 2008. The craftsman. Yale University Press. Silbey, Jessica. 2014. Harvesting Intellectual Property: IP Interventions and the Roles of Intellectual Property in Creative and Innovative Work . Stanford University Press. Smith, Steve. 2009. “Mahler, Extra-Reduced for the River.” The New York Times . Spradley, J. P. 1979. The Etnographic Interview . Orlando, Florida: Harcourt Brace Jovanovich College Publishers. Spradley, J. P. and K. Baker. 1980. Participant Observation. New York: Holt, Rinehart and Winston. Trost, J. E. 1986. “Statistically nonrepresentative stratified sampling: A sampling technique for qualitative studies.” Qualitative sociology 9:54— 57. Weber, Max. 1930. The Protestant Ethic and the Spirit of Capitalism. T. Parsons, trans. London and New York: Routledge. Weil, Simone. 1952. “The Need for Roots, trans.” Arthur Wills. New York: GP Putnam’s Sons . 39 Wrzesniewski, Amy. 2010. “Callings.” In Oxford Handbook of Positive Organizational Scholarship. Oxford University Press. Wrzesniewski, Amy, Clark McCauley, Paul Rozin, and Barry Schwartz. 1997. “Jobs, careers, and callings: People’s relations to their work.” Journal of Research in Personality 31:2133. Wrzesniewski, A., J. Tosti, and J. Landman. 2006. “If I could turn back time: Occupational regret and its consequences for work and life.” Unpublished manuscript, Yale University . 40 1 Figures and Tables Figure 1: Channapatna on Bangalore-Mysore Highway 41 42 Figure 2: A Diagram illustrating the Production Process 43 Figure 3: Summary of Hypotheses of Impact of Work Orientation on Selling Outcomes Figure 4: Some Channapatna Bangles at an Artisan’s Home 44 45 Figure 5: GPS Plot of Artisans and Traders in Sample Figure 6: Distribution of Initial Price(in Rupees) for Artisans 1 Histogram of Traders' Initial Prices .6 .4 .8 Cumulative Density (Initial Price) .04 Density .02 0 20 40 Initial Price 0 0 0 .2 .2 .02 Density .04 .06 .4 .6 .8 Cumulative Density (Initial Price) .06 1 .08 Histogram of Artisans' Initial Prices 60 20 40 Initial Price Initial Price: price first quoted by seller Exchange Rate: Rs.50=$1 45.30 36.87 31.04 32.72 28.15 18.63 10 Mean Initial Price 20 30 40 50 Figure 7: Mean Initial Price(in Rupees) Artisans Traders Sellers Indian-Baseline Indian-Craft Initial price: price first quoted by seller Error bars represent a 95% confidence interval around the mean Exchange Rate: Rs.50=$1 46 International 60 80 53% Discount 40% Discount 20 Linear Prediction of Initial Prices 25 30 35 40 Figure 8: Predicted Impact of Work Process Involvement on Initial Prices 0 1 2 3 4 5 6 7 8 Work Process Involvement 9 10 Indian-Baseline Indian-Craft International Error bars represent 95% confidence intervals Note:The entire work process can be split into 9 steps, some of which can be subcontracted. Processes is a continuous variable ranging from 1-9 measuring the number of work processes an artisan engages in. The graph captures how artisans’ prices change differentially for each category of buyers as they are involved in more work processes. Estimates are reported from a regression with Initial Price as the dependent variable and Buyer Group dummies and Processes as independent variables, including both main effects and interactions between buyer groups and processes. Each data point represents the predicted initial price at a given level of (buyer group, process level). For example, the predicted initial price for buyer group International for artisans engaging in 4 work processes is 27.48. 47 Table 1: Comparision of Seller Characteristics Artisan Trader Difference Fraction Male Age Number of Family Members Fraction Married Fraction Muslim Fraction Backward Castes Years of Education Completed Fraction Literate Work Tenure (years) Fraction in Cooperatives Exhibitions Attended/year Visits to Bangalore/month Radio Listening Hours/day Income in Dollars 0.923 (0.269) 42.65 (9.903) 6.423 (4.421) 0.942 (0.235) 0.769 (0.425) 0.885 (0.323) 6.788 (3.472) 0.885 (0.323) 23.71 (10.13) 0.288 (0.457) 1.292 (1.732) 3.241 (2.325) 4.667 (2.371) 77.39 (50.18) 1 (0) 46 (9.582) 5.826 (2.146) 1 (0) 0.130 (0.344) 0.826 (0.388) 10 (3.516) 0.957 (0.209) 19.32 (17.54) 0.261 (0.449) 2.200 (1.095) 2.450 (1.538) 3.200 (3.676) 155.2 (85.53) 49 21 Observations -0.077 -3.346 0.597 -0.058 0.639∗∗∗ 0.059 -3.212∗∗∗ -0.072 4.396 0.028 -0.908 0.791 1.467 -77.850∗∗∗ mean coefficients; sd in parentheses * p<0.1, ** p<0.05, *** p<0.01 Source: Survey conducted in June 2012 with sellers in experimental sample; 100% response rate for artisans and 96% response rate for traders; one trader refused the survey and one trader owns two shops in the sample 48 Table 2: Number of Sales Transactions by Sellers and Buyers SELLERS BUYERS Artisans (n=52) Traders (n=25) Total Indian-Baseline (n=2) Indian-Craft (n=2) International (n=2) Total 103 102 100 305 50 50 50 150 153 152 150 455 Experimental Design: 6 buyers were hired to purchase bangles from a sample of artisans and traders; Buyers visited the 2 types of sellers in a randomly assigned order; The buyers represented 3 experimental categories: Foreigners, Indians wearing handmade craft products and Indians dressed as usual Table 3: OLS Balance Checks on Experimental Transactions (1) (2) (3) Availability of Electricity Stock Left Presence of Seller’s Spouse Indian-Craft -0.095 (0.056) 3.865 (2.502) 0.119∗ (0.048) International -0.034 (0.057) 3.742 (2.510) 0.063 (0.048) Constant 0.627∗∗∗ (0.040) 12.438∗∗∗ (1.766) 0.163∗∗∗ (0.034) 455 455 455 Observations Indian-Baseline is the omitted category Standard errors in parentheses * p<0.1, ** p<0.05, *** p<0.01 Note: Auditors collected data on these baseline characteristics after every transaction; Availability of Electricity: Dummy variable indicating if there was electricity at time of sale (Channapatna faces power outtages of 3-4 hours a day); Stock Left: Estimate of number of bangles left with seller after sale; Presence of Seller’s Spouse: Dummy variable indicating if spouse was present during sale 49 Table 4: OLS Regression on Log Prices (1) (2) (3) Artisans-Initial Traders-Initial Artisans-Final (4) Traders-Final [D2]Indian-Craft -0.508∗∗∗ (0.048) 0.113∗ (0.052) -0.558∗∗∗ (0.061) 0.049 (0.058) [D3]International -0.274∗∗∗ (0.051) 0.451∗∗∗ (0.074) -0.325∗∗∗ (0.057) 0.401∗∗∗ (0.067) Constant 3.733∗∗∗ (0.069) 3.157∗∗∗ (0.124) 3.569∗∗∗ (0.079) 3.057∗∗∗ (0.113) Date FE Yes Yes Yes Yes Time of Day FE Yes Yes Yes Yes 275 18.054 0.000 150 34.728 0.000 268 10.795 0.002 150 43.924 0.000 Observations F-test [D2]=[D3] p-value [D1]Indian-Baseline is the omitted category Standard errors clustered by seller are in parentheses Zero prices have been dropped in this table to calculate log prices Zero prices represent 10% of artisans’ initial prices and 12% of artisans’ final prices Results are robust to raw prices (in Rupees), exclusion of fixed effects and inclusion of zero prices * p<0.1, ** p<0.05, *** p<0.01 50 Table 5: OLS Regression on Likelihood of Zero Prices offered by Artisans (1) (2) Zero Initial Price Zero Final Price [D2]Indian-Craft 0.180∗∗∗ (0.044) 0.237∗∗∗ (0.051) [D3]International 0.068∗ (0.037) 0.091∗∗ (0.038) Constant -0.034 (0.055) -0.067 (0.055) Date FE Yes Yes Time of Day FE Yes Yes 305 4.722 0.034 305 8.788 0.005 Observations F-test [D2]=[D3] p-value [D1]Indian-Baseline is the omitted category Standard errors clustered by seller in parentheses Results are robust to exclusion of fixed effects * p<0.1, ** p<0.05, *** p<0.01 Zero Initial Price: Dummy variable indicating if initial price=0 Zero Final Price: Dummy variable indicating if final price=0 10% of initial prices quoted by artisans and 12% of final sale prices were zero 51 Table 6: OLS Regression on Bargaining Rounds and Transaction Length (1) (2) (3) (4) Artisan-Rounds Trader-Rounds Artisan-Length Trader-Length [D2]Indian-Craft -0.621∗∗∗ (0.129) -0.120 (0.078) 4.014∗∗∗ (1.149) 0.800 (1.244) [D3]International -0.280∗∗ (0.120) 0.020 (0.035) 2.402∗∗∗ (0.798) 0.100 (0.629) Constant 2.670∗∗∗ (0.088) 2.940∗∗∗ (0.033) 7.398∗∗∗ (0.568) 7.860∗∗∗ (0.306) 305 5.184 0.027 150 3.086 0.092 305 1.081 0.303 150 0.385 0.541 Observations F-test [D2]=[D3] p-value [D1]Indian-Baseline is the omitted category Standard errors clustered by seller in parentheses * p<0.1, ** p<0.05, *** p<0.01 Bargaining Rounds: Number of rounds of bargaining on price; ranges from 0 to 3 Transaction Length:Time taken for sale in minutes 52 Appendix A : Robustness Checks This section describes results from a number of robustness exercises performed to investigate the role of alternative explanations in explaining artisans’ distinctive pricing behavior. First, the field experimental setting is able to rule out the competing alternative explanation of reputational incentives, that artisans give discounts in the short-run for lucrative gains from an enhanced reputation in the long-run. Artisans in Channapatna produce and sell in anonymity, lack branding (do not sign their products) and given the isolation of the town, engage in onetime transactions with tourists devoid of reputational concerns (over 90% of artisans reported never selling twice to the same direct customer). In many settings, one might suggest that pricing behavior is driven by pervasive socialization of actors in an occupational community. How artisans price to different customers might be driven by values and practices that they are exposed to during their apprenticeship, which they gradually habituate through increased contact with their peers in the local community. If this process was important in my setting, we would expect to see that older and more experienced artisans would have a greater degree of socialization, offering greater discounts to Indian-Craft and International auditors. However, this prediction does not hold true in the data (Figure A.1 and A.2). While socialization does not seem to be an important determinant of pricing behavior in this case, a similar but distinct alternative explanation could be that artisans’ pricing behavior is driven by community norms around pricing, where sanctions are imposed if visitors from far away or people that display a taste for craft products are charged high prices. The idea here is that, while an individual artisan might not be willing to offer a discount, he is forced to do so because of peer pressure from other artisans in the vicinity who might be monitoring his behavior. While the paper has qualitatively established that artisans in Channapatna operate in relative isolation and are unable to observe each other’s dealings, a formal test for this idea can be performed using precise locational data obtained from a handheld geopositioning (GPS) device. This device enabled not only the collection of the exact latitudinal and longitudinal position for each artisan and trader, but also the calculation of the number of other sellers living with a certain radius of the business. Figure A.3 reports the mean prices offered by two groups of artisans, distin53 guished by the extent of other sellers present in their vicinity. The results indicate that artisans who live in a “sparse” surrounding (with less that 12 artisans in a 0.2 mile radius and thus, are less likely to observed) do not price significantly differently than artisans who live in a more densely packed neighborhood (with relatively greater probability of being observed during transactions), suggesting that community sanctions do not seem to be an important factor as far as pricing decisions of artisans are considered. The GPS data also facilitated the investigation of the alternative explanation that artisans’ pricing behavior is driven by their level of market exposure. If artisans do not have enough exposure to different types of customers, then we do not expect them to price “optimally” in a narrow financial sense, like the traders. This can be tested in two ways. First, artisans who live close to the highway (and therefore have relatively more exposure to tourists) should offer smaller or no discounts to discerning buyers as compared to artisans who live in interior regions. Second, artisans who transact with traders on a regular basis should be expected to offer lower discounts because they have a better sense of traders’ “optimal” pricing structures. However, the data seem to reject both of these predictions. Group means as indicated in Figures A.4 and A.5 suggest that artisans’ distance from the highway or whether they transact with traders does not seem to affect the size of discounts offered to Indian-Craft and International auditors. This finding provides some evidence that a lack of market exposure does not seem to be driving the pricing behavior in this setting. Another alternative explanation that could account for the results is based on notions of artisans’ naivete or their fascination with customers from high status groups. Here, the idea is that artisans become so enamored with international tourists or customers wearing craft products that they find themselves offering discounts that they would not give to other customers. In the survey that I conducted, one question asked artisans about the frequency of their visits to nearby cities, especially Bangalore because it is the capital of the state of Karnataka where Channapatna is situated, it is not too far away and because it is a prominent shopping destination both for international tourists and affluent, art-loving Indian customers. Insofar as Bangalore provides sellers exposure to high status customers and foreign tourists, one would expect that artisans who visit Bangalore have a lower degree of fascination towards International or Indian-Craft auditors. However, Figure A.6 demonstrates that this prediction is not true in the data, suggesting that fascination with 54 high status customers or artisans’ naivete does not seem to be an important driver of the main results. Finally, while it seems unlikely, it is possible that international customers or those wearing craft jewelry might have relatively greater experience shopping for craft items in India and are therefore more conversant with the prices that are typically charged for such products, while other Indian customers are unaware of such details. This explanation predicts that artisans, knowing that experienced customers have a keen knowledge of “true” prices, offer such customers lower prices as compared to other inexperienced customers. Fortunately, the survey I conducted collects data on prices that artisans charge, on average, to retail clients for the Channapatna Bangle. The survey finds that the average price that artisans obtain for the Channapatna bangle is Rs.32.50, as determined by their answer to the question – “How much do you typically sell half-inch bangles for?” This price is significantly higher than the average initial price that both International and Indian-Craft auditors were offered, indicating that not only did artisans offer these buyers prices that were not equal to the average prices for bangles in the area, but they also seemed to forgo substantial profits in order to be able to sell to more appreciative customers. Experienced customers’ prior knowledge of true prices therefore, does not seem to be driving differential pricing decisions here. The results provided in this section, thus offer further confirmation of the theory that artisans’ pricing behavior is driven by identification with work. 55 0 10 Mean Initial Price 20 30 40 50 Figure A.1: Robustness Check for Artisans: Age Indian-Baseline Indian-Craft BUYERS Young Adult International Middle-Aged Young Adult<= 40 years, Middle-Aged>40 years Exchange Rate: Rs.50=$1 0 10 Mean Initial Price 20 30 40 50 Figure A.2: Robustness Check for Artisans: Work Tenure Indian-Baseline Indian-Craft BUYERS Junior Junior<25 years in the trade, Senior>=25 years in the trade Exchange Rate: Rs.50=$1 56 International Senior 0 10 Mean Initial Price 20 30 40 50 Figure A.3: Robustness Check for Artisans: Presence of Neighboring Sellers Indian-Baseline Indian-Craft BUYERS Sparse International Dense Sparse<=12 sellers, Dense>12 sellers in a 0.2 mile radius Exchange Rate: Rs.50=$1 0 10 Mean Initial Price 20 30 40 50 Figure A.4: Robustness Check for Artisans: Distance from Highway Indian-Baseline Indian-Craft BUYERS Near Near<=0.2 miles, Far>0.2 miles from highway Exchange Rate: Rs.50=$1 57 International Far 0 10 Mean Initial Price 20 30 40 50 Figure A.5: Robustness Check for Artisans: Business with Traders Indian-Baseline Indian-Craft BUYERS No Trader Business International Sells to Trader Sells to Trader: In addition to selling to direct customers, artisan sells to trader Exchange Rate: Rs.50=$1 0 10 Mean Initial Price 20 30 40 50 Figure A.6: Robustness Check for Artisans: Visits to Bangalore Indian-Baseline Indian-Craft BUYERS Bangalore Visit International No Visit Bangalore Visit: Artisan went to Bangalore at least once in the month preceding the experiment Exchange Rate: Rs.50=$1 58 0 10 Mean Initial Price 20 30 40 50 Figure A.7: Robustness Check for Artisans: Income Indian-Baseline Indian-Craft BUYERS Poor International Rich Poor earns less than or equal to Rs.5000 a month, Rich have income greater than Rs.5000 Exchange Rate: Rs.50=$1 0 10 Mean Initial Price 20 30 40 50 Figure A.8: Robustness Check for Artisans: Religion Indian-Baseline Indian-Craft BUYERS Hindu 77% of Artisans and 13% of Traders are Muslim Exchange Rate: Rs.50=$1 59 International Muslim 0 10 Mean Initial Price 20 30 40 50 Figure A.9: Robustness Check for Artisans: Education Indian-Baseline Indian-Craft BUYERS Primary International Secondary Primary<=6 years of education, Secondary>6 years of education Exchange Rate: Rs.50=$1 60