‘‘Before cardamom, we used to face hardship”: Analyzing agricultural commercialization effects in Nepal through a local concept of the Good Life

Numerous low-income countries foster the commercialization of smallholder agriculture to achieve development outcomes and improve the lives of the rural population. The effects of commercialization policies, however, are measured using a limited set of indicators. This paper exempliﬁes a new approach to the study of agricultural change: analyzing commercialization effects through a local concept of the Good Life. In our case study of East Nepal, we ﬁrst elicited a local concept of the Good Life through qual- itative interviews and participatory photography. In the analysis, we disaggregated the data between men and women, elderly and young, farmers and laborers as well as members of different castes. Second, we applied the resulting Good Life concept to the evaluation of agricultural commercialization. Our results show that the local concept of the Good Life is multidimensional and includes both subjec- tively and objectively measurable dimensions. Respondents across all socio-economic groups consistently emphasized the notion of hardship ( dukha ) in both their Good Life concepts and their perspectives on agricultural change. Commercialization was evaluated positively predominantly because it reduced physical and ﬁnancial hardship, in addition to tangible improvements in other domains. However, respondents also pointed to the limitations of commercialization in contributing to the Good Life: the ultimate reduction of hardship was associated with the prospect of non-agricultural employ- ment. The notion of hardship elicited through the perspectives of the Good Life offers a nuanced perspective on commercialization. Including local views in analyses of agricultural change enables researchers and policy makers alike to direct their efforts to those aspects of agricultural change that are most meaningful to the local population. (cid:1)


Introduction
Agricultural systems worldwide have undergone drastic changes in recent decades. Commercial agriculture has enabled states to feed their growing populations, and average per capita food amounts have risen (Federico, 2009). Today, sustained productivity increases are still necessary for feeding the growing world population, for contributing to the elimination of poverty, and for meeting the Sustainable Development Goals (Barrett et al., 2018;Collier & Dercon, 2014). In the global South, the first Green Revolution policies were implemented in the 1960s (Birner & Resnick, 2010). To date, numerous low-income countries are pursuing commercialization strategies to increase production, raise farm incomes, create rural employment, and eventually transform their economies (see Dawson et al., 2016;Emran & Shilpi, 2018;Ivanic & Martin, 2018). In these countries, the agricultural sector accounts for nearly 60% of total employment (World Bank, 2019). Therefore, changes in agricultural policies affect millions of people in different parts of the world.
In this paper, following Carletto et al (2017) and Pingali & Rosegrant (1995), we define agricultural commercialization as a rise in the level of market-orientation of small-scale producers. The potential benefits of agricultural commercialization include poverty reduction, enhanced food security, employment creation (FAO, 2002), and, to some extent, an increase in women's economic empowerment (Kabeer, 2005). However, while it is established that commercialization raises average agricultural income (see Section 2.1.), the extent to which rural people believe this makes their lives better is not well understood. While the economic dimension forms an integral part of numerous well-being concepts, information on average income alone is insufficient for analyzing the well-being effects of agricultural commercialization. First, the economic benefits associated with agricultural commercialization may be distributed unequally across different social groups, potentially leaving the most disadvantaged behind (Bieri, 2014). Second, it is not a given that an increase in income will result in improvements in other domains of life. Consequently, in recent decades a growing body of literature has engaged in discussions on the concept of well-being and the construction of meaningful indices applicable to the development context (Gough & McGregor, 2007). Yet, these debates apparently have not had a major influence on agricultural development research; the bulk of studies investigating the effects of commercialization continue to rely on a limited set of quantitative indicators (see Section 2.1).
This article strengthens the link between research on agricultural commercialization and debates on well-being in development through a two-stage analysis of agricultural transformation in a rapidly changing region in East Nepal. First, we established a local concept of the Good Life, elicited through an innovative methodological combination of participatory photography and in-depth interviewing. In doing so, we deliberately did not draw from theory-based concepts such as the Good Life Elements (Skidelsky & Skidelsky, 2013;Delhey & Steckermeier, 2016). Instead we pursued a participatory approach, leaving the definition of the Good Life entirely to the respondents (see Calestani, 2009;Fischer, 2014;Lim, 2008). In our analysis of local Good Life perspectives, we carefully distinguished between members of different social groups: men and women, elderly and young, farmers and laborers, as well as members of different castes. In stage two, we applied the local concept of the Good Life to our analysis of agricultural commercialization in that region. Not only did the respondents produce a nuanced multidimensional concept of the Good Life that was consistent across social groups, they also evaluated commercialization through a lens that appears to be new to development research: the hardship perspective. By advancing these insights, we contribute to the field of agricultural development research and to debates on well-being in development.

Contrasting conclusions on the effects of commercialization
Researchers, governments, and international organizations alike consider agricultural commercialization a promising strategy for achieving development outcomes on a macro-economic level (Binswanger and von Braun, 1991;von Braun, 1995;Christiaensen et al., 2011;Diao et al., 2010;FAO, 2002;Ivanic & Martin, 2018;Maxwell & Fernando, 1989;Pingali, 2010;Pingali & Rosegrant, 1995;Strasberg et al., 1999;United Nations, 2009;Zhou et al., 2013). This particularly applies to the commercialization of smallholder agriculture. While some scholars argue that investments in both large-scale and small-scale farms would be the most promising strategy for rural development (Collier & Dercon, 2014;Glover & Jones, 2019;van den Broeck & Maertens, 2017), many are convinced that smallholder-based agricultural commercialization is the most effective poverty reduction strategy. For instance, Dorosh and Thurlow (2018) compare the effects of agricultural growth achieved by small and large farms and find that smallholder-led agricultural growth has greater poverty reduction effects. Hazell et al. (2010) demonstrate that small farms have a higher productivity rate per hectare and hire more labor per unit area, thus improving local employment opportunities and generating greater spillover effects on the rural non-farm economy. Similarly, Bieri (2014) highlights the employment potential of smallholder-based commercial export production, particularly for women. Wiggins et al. (2010) conclusively state that ''small farm development is not just desirable for poverty reduction, but also feasible, even in changing circumstances" (p. 1341).
While the macro-economic effects of commercialization are generally evaluated positively, impacts on a micro-economic level vary widely, depending on the respective investment schemes and agricultural policies in place (Bachewe et al., 2018;Beck et al., 2016;Birner & Resnick, 2010;von Braun, 1995;Glover & Jones, 2019;Pingali, 2010;Pingali & Rosegrant, 1995;Rahut et al., 2010). Studies measuring the effects of specific agricultural interventions show mixed results. Positive outcomes were found in the Philippines, for instance, where tobacco contract farming increased farm profitability while reducing inequality between households (Briones, 2015). In Bangladesh, increased rice productivity and agricultural wages significantly contributed to poverty reduction (Emran & Shilpi, 2018). Agricultural commercialization processes also decreased poverty rates in Senegal (van den Broeck & Maertens, 2017). In Kenya, commercial smallholder legume and banana production were associated with greater household welfare (Ochieng et al., 2015), and small-scale vegetable commercialization increased food security and dietary diversity (Muriithi & Matz, 2015). These positive impacts notwithstanding, other studies demonstrate that commercialization processes had either no effects or even adverse impacts on the rural population. In Liberia, a value chain intervention successfully increased yields and incomes, but this change did not translate into higher household welfare or improved child nutrition (Rutherford et al., 2016). Comparing data from Malawi, Tanzania, and Uganda, Carletto et al. (2017) found no substantial relationship between commercialization and nutritional status. An analysis of rural data from eight sub-Saharan countries showed that commercialization led to inclusive agricultural growth in some villages, but not in others (Andersson Djurfeldt, 2013). Finally, a recent study in Rwanda demonstrated that the policies in place benefitted mostly the comparatively wealthy households, leaving poorer families behind (Clay & King, 2019). In sum, while there is substantial evidence of positive commercialization effects on a macro-economic level, the specific local outcomes ''are not uniform and cannot be generalized" (Strasberg et al., 1999, p. 2).
What is striking in most of the research cited above is that the effects of agricultural commercialization are mainly measured in terms of income poverty and food security: broader indicators of well-being or the views of the rural populations affected by these commercialization processes are usually not considered. This is noteworthy given that for at least four decades there have been vigorous academic debates on measuring development outcomes in more holistic ways. Research that aims to measure the effects of agricultural change appears to seldom draw on these debates.

Multidimensional measures of well-being in development
Income only measures have been criticized for their inability to accurately measure the impacts of development policies at an empirical level. Alkire et al. (2014) explore the relationship between income measures and other objective indicators and do not find a consistent link between income poverty and other dimensions of deprivation such as malnutrition. The inadequacy of using income only measures to draw conclusions about other objective indicators is further explored by Carletto et al. (2017) who found that commercialization had no effect on nutritional outcomes in three African countries. Consequently, a measure used in the development context should not only capture income poverty but also encompass deprivations in other areas of life, such as a lack of adequate nutrition or education. The most prominent example of combining an income measure with other objective indicators is the Human Development Index (Anand & Sen, 1994), but there are also other approaches comprising a broader range of objective indicators (e.g., Alkire et al., 2015;Berenger & Verdier-Chouchane, 2007).
The second area in which income only measures fall short is in their ability to generate empirical findings on the relationship between income levels and subjective indicators. Since Easterlin found that increases in income did not make Americans happier (Easterlin, 1974), the validity of the so-called Easterlin paradox has been a matter of discussion, both in relation to the global North and the global South. A recent comparative study found no significant association between incomes and happiness in developing countries (Mikucka et al., 2017). Pure income measures, it seems, fail to capture the lived realities of people not only regarding broader objective indicators like nutrition, but also when considering the subjective dimension. Consequently, a growing body of development research focuses on subjective well-being, or happiness (Camfield & Esposito, 2014;Fernandez et al., 2015;Fontaine & Yamada, 2014;Graham, 2005;Helliwell et al., 2018;Kingdon & Knight, 2006;Kroll, 2015;Rojas, 2008;Rojas & Guardiola, 2017). The use of subjective well-being measures in the development context has been criticized on both methodological and conceptual grounds, inter alia for being inaccurate or biased, for neglecting physical deprivation, for disregarding the social dimension and other important elements of human life, and for having depoliticizing effects, potentially undermining the case for development assistance (Graham, 2005;Sen, 1999;Schokkaert, 2007;Stewart, 2014;White, 2010; for a broader feminist critique see Ahmed, 2010). Such criticism notwithstanding, research on subjective well-being has become an important field in the development research arena.
As debates have continued around the ability of various measures to provide insights into the lived existences of the poor, academic perspectives on development have been changing; in recent decades, there has been a fundamental shift from income and consumption-based approaches to multidimensional concepts (Hojman & Miranda, 2018). Development is increasingly seen as the ''organised pursuit of human wellbeing" (Gough & McGregor, 2007, p. 4) which is best measured through a combination of both objective and subjective indicators. This holistic perspective is reflected in a variety of academic approaches (Costanza et al., 2007;Diener & Tay, 2015;Gasper, 2005;McGregor et al., 2009;White, 2010; to name a few) as well as in implementationoriented indices, produced for instance by the WHO (1997) and the International Wellbeing Group (IWG, 2013). According to the latter, there are over 1,200 idiosyncratic instruments to measure quality of life. While combinations of subjective and objective dimensions have become increasingly common, there is by no means a consensus on which dimensions a comprehensive wellbeing measure should include. To summarize in the words of Dodge et al. (2012) ''wellbeing is a growing area of research, yet the question of how it should be defined remains unanswered" (p. 224).

The need to include local perspectives
The choice of a development measure -be it income, food security, subjective well-being or a multidimensional measure -always involves a normative decision on what is worth measuring. Based on this consideration, a growing body of literature suggests that researchers should not take this decision from a theoretical stance alone. Instead, the meaning of concepts like ''development" or ''well-being" should be established in conjunction with the people whose very lives are under investigation (Beauchamp et al., 2018;Chaves et al., 2018;Kant et al., 2014;Lim, 2008;van Norren, 2017;Zorondo-Rodríguez et al., 2014). Poor people in different parts of the world have their own, culturally diverse understanding of the Good Life and yet, they are mostly deprived of opportunities to contribute their views to global and local development discourses (Gough, 2004). This is problematic, as exemplified by a recent mixed-methods study of agricultural transformations in Rwanda by Dawson et al. (2016). Even though commercialization successfully increased yields and reduced poverty rates, the analysis of local perspectives revealed that commercialization increased inequality, disrupted social practices, and undermined the farmers' autonomy. According to Dawson et al., a study based on a limited set of quantitative indicators would have led to the erroneous evaluation of agricultural commercialization as unequivocally benefitting the poor. To avoid such shortcomings, ''we must seek to understand local conceptions of the good life through which a particular community pursues developmental goals" (Lim, 2008, p. 208). The paper at hand takes up this call.

Agricultural change in the study area
For our case study, we selected a region characterized by rapid agricultural transformation: the mountainous Rong Rural Municipality in Ilam District, East Nepal, elevated between 275 and 1,836 m above sea level (see Fig. 1). Owing to its medium elevation, ward Rong 6 is suitable for the production of black cardamom (Amomum subulatum Roxb., henceforth referred to as cardamom). From 2003 onwards, numerous farmers chose to produce cardamom as its economic value exceeded the revenue of other cash crops (K. C. et al., 2016), particularly the high quality variety Jirmale/Salakpurey which grows best between 700 and 1000 m above sea level (Adhikari & Khanal, 2016;Timsina & Paudel, 2016). In areas unsuitable for cardamom production, farmers produce tea (high elevations), broom grass (steep slopes) and other cash crops such as betel nut (low elevations). According to the Office of the Rong Rural Municipality, the three main cash crops in terms of volume are broom grass (227 metric tons), tea (78 t), and cardamom (72 t); main subsistence crops include maize (180 t) and rice (80 t) (ORRM, 2018).
For most of the 20th century, farmers of Rong 6 pursued subsistence agriculture (K. C., 2019). In the 1940s, the first households started cultivating ginger and tangerines for domestic use. Gradually, farmers replaced their subsistence crops with ginger, and by the early 1980s, ginger had gained major economic importance. In the 1990s, tangerine farming became economically viable, but ginger remained the predominant income source until the year 2000. In 1993, broom grass was introduced as an additional cash crop. In 1995, farmers for the first time took small quantities of cardamom to the market -the crop had already been introduced in 1984 by a local farmer who had brought saplings from India, but initially it was used for domestic purposes only. Around the year 2000, diseases affecting ginger became a major issue in in the area. Very quickly, farmers replaced their ginger fields with cardamom plantations: by 2003, 95% of ginger farmers had switched to cardamom production (ibid.).
The cardamom price in Nepal has been characterized by significant fluctuation: after a relatively stable period from 1970 to 2009 the price rose sevenfold to a peak in 2015 and fell sharply afterwards (FAO 2018). At its peak, the average price for high quality cardamom was USD 28 per kg at the market in Ilam (ITC, 2017). In our fieldwork, farmers in Rong 6 reported local prices ranging from NPR 3,000 per kg in 2015 (USD 25) to NPR 600 per kg in February 2019 (USD 5). However, this still exceeds the tea price which was fixed by the local government in 2019 at NPR 40 per kg (USD 0.3). In this context, cardamom remains an attractive income source: for 69% of 514 households surveyed in 2018, cardamom was the most important crop (Subedi & Upreti, 2019).

Data collection and analysis
We carried out the fieldwork between September and December 2018, in a period of low agricultural activity after the cardamom harvest and returned to the study area in February 2019 to collect additional information. After an exploratory field trip, we selected the economically most and least developed village out of three villages within Rong 6, based on our observations and discussions with local key informants.
To maximize diversity among the respondents, we combined quota sampling with sampling for variation (Morse and Niehaus 2009). Our sampling frame provided information on gender and caste. We intersected the binary caste variable ''Dalit/non-Dalit" with the binary gender variable ''man/woman", thus creating four strata per village. We then randomly selected five respondents per village and stratum, resulting in a sample of 40 respondents. Initially, we had not intended to focus on caste because we wanted to avoid reifying caste discrimination which is banned by the constitution (Government of Nepal, 2015), but we realized in the field that caste still mattered. By oversampling Dalit women and men, we ensured appropriate representation of the Dalit minority.
In the randomized quota sample of 40 respondents, laborers (i.e., persons whose main income is agricultural labor, even though they might have a small cardamom field) and young adults were not well represented, so we purposively selected 18 additional respondents whom we approached through snowball sampling (Bryman, 2012) or during their work on the cardamom fields. The sample incorporates men and women, Dalits and non-Dalits, farmers and laborers, elderly people and young people, and thus grants insight into all social milieus in the study area.
With 53 of 58 respondents we had extensive conversations on what it means to lead a Good Life. Importantly, in these in-depth interviews (Ritchie & Lewis, 2003) we did not prompt a given list of Good Life dimensions. While prompting is a useful technique in many respects (Yeo et al., 2014), it might have led to bias in our case: we intended to understand the very concept of the Good Life from the perspective of the respondents, and through prompting potential Good Life dimensions we would have implicitly introduced academic concepts. Instead, we relied on open questions, inspired by Greco et al. (2015), such as ''What does it mean to you, personally, to lead a good life?". When necessary, we stimulated additional narration through questions about good and bad phases of life in the past. In rare instances, this led to ethical challenges, for instance when a widow burst in tears upon speaking about her husband's death. In this case, we did not pressure the respondent to answer all questions (Rubin and Rubin 2012); rather, we moved to less sensitive topics and concluded the interview on a lighter note. In addition to Good Life perspectives, interview topics included agricultural change, agricultural labor, and the life changes that followed the shift to cardamom. In all interviews, closed questions were only asked when necessary to obtain specific information such as the respondents' ages. After completing about half of the interviews, no new Good Life dimensions or other central themes emerged and data saturation (Saunders et al., 2018) was achieved.
In addition to the in-depth interviews, we carried out a photography project with eight middle-aged male and female participants, of whom six were non-Dalits and two were Dalits. Inspired by Yefimova et al. (2015), we requested the respondents to take about 20 pictures of their everyday life with a digital cam- era, including pictures of work, free time, things they liked and disliked, and something that was important to them. Based on the printed pictures, we had extended conversations with the participants in an informal atmosphere, allowing for elaborate narrations. Compared with in-depth interviews, photography-based interviewing had three advantages. First, the participants were actively involved as data collectors. Second, at the time of the interview we had met the participant at least three times which increased familiarization. Third, the photography interviews took considerably longer (twice as long on average; some lasted over two hours), and the depth of the narrations was very high. As such, our main goal in using participatory photography was not to produce pictures for analysis but to use the photographs as an effective stimulus for narration. Challenges associated with participatory photography include the question as to what can be disclosed through pictures and what is intentionally or unintentionally left out (Wang & Burris, 1997), a potential bias towards the positive aspects of life (Byrne et al., 2016), the risk of social control and surveillance (Prins, 2010), and different ethical issues related to privacy and consent (Yefimova et al., 2015). In the context of our study, it is important to be aware of the potential positive bias as people may wish to present their life in a favorable light, and photography may not adequately represent non-visible aspects of life, such as self-determination or peace. To overcome these challenges, it is useful to combine participatory photography with other qualitative research methods. We conducted the interviews in Nepali with the help of a translator and recorded the interviews with the consent of the participants. During the exploratory phase we worked with a female translator but later changed to a male translator who was gifted in establishing good rapport with both male and female respondents. Upon analysis we could not detect whether the gender of the translator was significant; the interviews conducted with the male translator resulted in a comparable range of topics but produced richer descriptions from both genders. We prepared full verbatim translated transcriptions of all interviews which were then imported into MaxQDA. Drawing from Grounded Theory (Strauss & Corbin, 1997), we assigned codes to the entire transcript without any pre-defined categories. In the process of coding, a flexible category system emerged and was constantly adapted and expanded as we added new codes. We then analyzed the data, summarizing shared views and highlighting contrasting perspectives wherever applicable. To account for intersectionality (Crenshaw, 1989(Crenshaw, , 1991 we created separate code matrices for four groups (gender/caste). However, across these groups, the exact same Good Life dimensions emerged, and we could not detect any qualitative differences when comparing the detailed descriptions of the dimensions between groups. Likewise, challenges associated with agricultural change were described consistently across groups. For the sake of completeness, we still indicate caste and gender of the respondents when presenting the results.

Eleven dimensions and an emphasis on hardship
The Good Life, according to the participants in our study, consists of eleven dimensions (see Fig. 2). Not suffering hardship was the most salient element: three quarters of the respondents considered the absence of hardship an integral part of the Good Life. Good relationships with family members and friends as well as happiness were the second and third most frequently mentioned aspects. About 40% of the respondents listed health, income, education, self-determination and a good life for their children.
Less frequently mentioned aspects included the ability to work, peacefulness in the home and in the heart, as well as food and clothes which were always mentioned together. The local concept of the Good Life hence is multidimensional and includes dimensions that can be measured with objective indicators (e.g., education) and dimensions that can be assessed subjectively (e.g., happiness).
To account for potential group differences, we analyzed the frequency data separately for men and women, farmers and laborers, Dalits and non-Dalits, as well as the elder 50% and the younger 50% of respondents. In doing so, we noted striking similarities: in all groups, the exact same eleven dimensions emerged, and the hardship dimension always ranked first (see Table 1). The biggest differences were found along the lines of caste and occupation: more non-Dalits than Dalits valued happiness (difference of 40 percentage points), and the elder 50% of respondents listed health and the ability to work more often than the younger 50% (differences of 36 and 34 percentage points). Differences by gender and occupation were less pronounced: the frequencies by which the respective Good Life domains were mentioned all differed by less than 25 percentage points. These variations in frequency notwithstanding, the overarching concept of the Good Life is comparable across social groups: regardless of social background, the respondents listed the same eleven dimensions as demonstrated in Fig. 2, and the absence of hardship was the most salient factor.
Not only did the respondents consider hardship important in terms of the Good Life. Even more so, they repeatedly invoked the notion of hardship when explaining their perspective on agricultural change. While the other ten dimensions of the local Good Life concept are common in agricultural studies (income, nutrition) and/or well-being concepts (health, happiness, social relationships, self-determination etc.), we have not come across any research that uses hardship as an analytical category in connection with well-being in development. Because of the respondents' strong emphasis on this dimension and the contrasting gap in research, we focus the following analysis of agricultural commercialization on the notion of hardship. Hereafter, we first explore the meaning of hardship in depth and then evaluate the effects of commercialization using hardship as a central criterion.
4.1.2. Three kinds of hardship: labor-related dukha, financial dukha, and emotional dukha The Nepali term for hardship is dukha. This expression is associated with a variety of English terms, including not only hardship but also trouble, problem, distress, shortage, need, sorrow, and grief (Schmidt, 2005). The corresponding expression dukha garnu (i.e., ''doing dukha") implies suffering, persevering, doing hard work or having a hard time (ibid.). The wide range of possible meanings is reflected in the respondents' use of the term which we classified in labor-related, financial, and emotional aspects of dukha.
First, the respondents used the term dukha to refer to physical hardship associated with agricultural labor, as explained by a 29year-old Dalit woman: ''We have to do dukha. For example, going for the agricultural labor, sometimes carrying loads and sometimes digging, sometimes walking to a distant place. For example, [. . .when] we need to carry the doko (basket carried with a strap around the head) and loads the entire day we will have dukha." Work termed dukha may result in ''blisters and pain in the hands and back pain [. . .] I might not be willing to do it, but I have to do it". Hence, when used in its labor-related sense, dukha describes physically challenging tasks.
Second, the respondents used the term dukha when illustrating a tangible scarcity of funds, whereby the source of such dukha predominantly was the cost of sending children to school. For instance, a 45-year-old Tamang woman explained: ''When the children were small, we had dukha to educate them.
[. . .] We sent our children to an expensive school in the city, to a private school. We had to invest a lot of money, so at that time we used to struggle for the money." Likewise, a 53-year-old Dalit man stated that some years ago his life had not been as good as it could have been ''because my children were small at that time. We used to depend on paid labor, and we needed to educate the children and run the household, so there was dukha at that time". Consequently, in its second sense, dukha stands for hardship due to financial insecurity.
Third, dukha is used to denote sorrow. Evaluating his life course, an elderly Dalit man disclosed: ''I haven't had dukha for affording food or clothes. But regarding another type of dukha, my elder son died when he was twelve years old studying in class seven, and my first wife also died. So, this type of dukha I have faced." The term dukha was, however, only rarely associated with emotional burdens. In most cases, the respondents used the expression when referring to hardship due to physical labor and financial challenges.

Dukha and sukha as complementary aspects of a fulfilled human life
Even though dukha is generally used to describe undesirable states, a young Dalit woman explained during a participatory photography interview that dukha in her view was an integral part of a fulfilled human life. She first clarified how dukha relates to its counterpart sukha: ''Sukha means, for example, not doing very hard work. Like, if we have money, then we can go to the market and buy things, whatever you can afford, you purchase it and sit with the family. This is what we call sukha. For being able to afford that sukha we must do dukha, like doing the labor and other things." She elaborated that before they had children, she and her husband used to have an easy life without major responsibilities. However, in that time, something was missing: ''We didn't know what life is like. [. . .] If we do the work and earn some money and we are able to buy food and feed the family, then that is a different kind of happiness and satisfaction. We feel: 'I am able to do this much for my family; though I have done dukha but also I can feed my family.' [. . .] Before, we didn't have satisfaction. We only enjoyed, but now we are satisfied. Now we identify both dukha and sukha."  Hence, in her opinion, a life without dukha might be enjoyable, but only through doing dukha is one able to achieve a ''different kind of happiness and satisfaction".
The young woman was the only respondent who explicitly stated that dukha was necessary for living well. Yet, a few other male and female respondents of different castes mentioned dukha and sukha as two sides of the same coin. For instance, in another participatory photography interview, a 33-year-old Tamang woman explained that ''good life means having sukha, [. . .] like, not doing the work, having the delicious food, visiting different places, that would be sukha: happiness". Her own life, overall, ''is running in dukha and sukha -we are living in this way", she said and laughed.

Analyzing agricultural change through a hardship perspective
As much as certain hardships inevitably form part of every human life, most respondents agreed that a good life is a life with little dukha. An analysis of agricultural commercialization from the perspective of the respondents hence needs to investigate whether and how the prevalence of dukha has been affected by agricultural change. The following sections provide an insight into the respondents' views on labor-related and financial dukha in the wake of cardamom production. The analysis again carefully differentiates between the perspectives of different social groups and highlights contrasting perspectives wherever applicable.

Commercialization and physical hardship
According to the respondents, cardamom production requires less effort than most other cash crops: because cardamom is a perennial plant, the field does not have to be ploughed between seasons. Comparing cardamom with ginger production, a 38year-old male Tamang farmer stated: ''I prefer the work required for cardamom production. It is easier; in ginger production there is more dukha." A 23-year-old male Rai farmer confirmed: ''Cardamom production is better; the work is a bit lighter". A 50year-old female Tamang farmer explained that for ginger production, ''we have to plough and prepare the soil two to three times before planting and while planting." After planting cardamom, however, ''we just weed, maintain the plant and weed. We will get income after two years. It is not as much dukha as in ginger." Agricultural laborers likewise confirmed this view. For instance, a 52-year-old male Dalit laborer preferred cardamom over ginger production because ''the work is a bit lighter; it does not require hard work like digging and ploughing. Only during the time of the first planting do we have to plough and dig the field, otherwise there is no need".
When comparing cardamom with other perennial cash crops, the opinions diverged. Harvesting broom grass, for instance, was considered an easy task by a 41-year-old male Dalit laborer, ''because we can just do the work standing [. . .]. For the cardamom work, mostly, we have to sit and do the work, and so I feel some pain in the hands and legs". In contrast, an 41-year-old Newar woman argued that harvesting broom grass involved more dukha because the sharp edges of the broom grass leaves can cause injuries. If arms and hands are not covered, ''they will be cut by the leaves. Moreover, the broom grass grows in the steep forest area, so we have to go there to collect it. In contrast, the cardamom work is much easier." Comparing cardamom production with subsistence farming, the latter was unequivocally viewed as involving more dukha. A 40year-old female Rai farmer confirmed: ''Before, when we used to have maize, millet and paddy, we used to spend most of our time in the field. [. . .] We had to work year-round. But for cardamom, we work during the season only." A 23-year-old male Rai farmer agreed that ''compared to cereal crops, cardamom production is a bit easier, it takes less effort".
While most farmers and laborers preferred cardamom work over the labor required for other crops, it should be noted that cardamom production still is a physically challenging task. A 39-yearold female Tamang farmer explained: ''We have to do dukha in cardamom also. We have to do the work in the cold area, like harvesting in the monsoon season. While working [in the rain] we might suffer from the cold." Consequently, cardamom production is still a challenging occupation, but it involves less labor-related dukha than subsistence farming and the production of most other cash crops.

Commercialization and financial hardship
As compared to all other crops in the study region, cardamom production yields substantially higher incomes -this was confirmed by respondents from all socio-economic backgrounds. For instance, a 34-year-old Tamang woman summarized: ''From this cardamom we have sukha. [. . .] Cardamom came, then the income increased, and then good progress occurred." With the term ''progress" she refers to a variety of changes in her family: they bought a motorbike, constructed a new cow shed, and procured several new household items such as a gas stove, various kitchen utensils, and a wide range of tableware. Similarly, a 47-year-old male farmer explained: ''Before, we had some dukha, we used to have tension from where to collect the money." Today, he is a successful cash crop farmer, having acquired additional plots and pursuing a sophisticated agricultural strategy based on a diverse portfolio of high-value crops.
While not all respondents reported such wide-ranging transformations of their lives, almost everyone stated that their incomes rose thanks to cardamom production. With the increased means, the respondents realized different changes in their everyday life, as exemplified in Fig. 3. For poor families, a significant improvement was year-round coverage of basic needs, as described by a 37-year old male Tamang farmer: ''Before, when we planted rice, there was less production and less income, and we had a shortage of some things, like food and clothes. But now, because of cardamom, it is good." Similarly, a 63-year-old male Dalit farmer stated: ''After selling cardamom we got to eat [laughs], and now we have money." Concordantly, a 40-year-old female Rai farmer disclosed: ''Before, when we used to plant maize, millet, and paddy we didn't have much income. We used to have difficulties to get enough food for the whole year, so we used to take loans. But now, with cardamom, we have income. We do not have to take loans anymore, and we can save some money." In the same vein, numerous respondents explained that they could overcome financial instability thanks to cardamom production. A 23-year-old female Tamang farmer summarized: ''We can run our household easily from the income, and during the time of emergency or sickness we can use that money." For agricultural laborers, incomes also rose because of both higher wages and extended employment opportunities. A female employer explained that cardamom production increased the labor demand on her farm in different seasons, and a 36-year-old female Dalit laborer confirmed: ''We can get more work when compared to the past." A 46-year-old male Rai worker added that salaries had been rising considerably: ''When we used to do work for maize and millet production, we used to get 150 rupees per day. But now, in cardamom, we get 300 rupees per day." A 29-year-old female Dalit laborer said that thanks to cardamom, her life was better ''because before we had a lower wage, but now we have a higher wage from cardamom. [. . .] I am spending that money for educating my son and for running the household." While a wide range of respondents reported that their financial dukha decreased thanks to cardamom production, several successful farmers pointed to a potential downside for those who recently crossed the poverty line. A 33-year-old male Rai farmer explained: ''The ones who previously used to be poor, they suddenly got more money and they uplifted their standard. [. . .] They started the habit of spending money beyond the necessary things [. . .] But nowadays, gradually, the cardamom price is declining and so the economic benefits are decreasing. [. . .] If they cannot maintain their standard, then in coming days they might have mental tensions." Several farmers observed that some households took loans for motorbikes or small luxuries, but had trouble repaying the credit once the cardamom prices fell, which then led to increased financial hardship. A 28-year-old male Rai farmer summarized: ''The ones who don't use the money properly, they are not doing well. But the ones who have saved the money in the bank, they have done well. [. . .] It depends upon the talent of the individual, on how they use the money." 4.3. Beyond hardship: Economic benefits and future prospects 4.3.1. Who reaps the economic benefits of cardamom production?
Despite the new financial challenges mentioned above, almost all respondents evaluated the impact of cardamom production on their lives positively. However, it seems that particular members of society benefitted disproportionally. A wealthy Rai farmer explained: ''In general, we cannot say that all of the people in the village benefitted equally. It depends upon different factors: education is one, income is another one, and work effort is another. Besides that, the land size is an important factor, [. . .] and irrigation is another crucial point. For the irrigation, the farmers need water and the sources are in very distant places, so they have to make huge investments in irrigation. Because of these reasons we cannot generalize that all people got economic benefits in an equal way." Establishing an irrigation system is a major investment: a young male Rai farmer indicated that he had invested about 350,000 NPR (3,000 USD) for the pipeline connection. Such a major investment is beyond the financial possibilities of most respondents. A 41-year-old Dalit farmer explained why this is problematic: ''The ones who have good irrigation, they will have good production in their field. During the time of the flowering [. . .] it is a bit dry and we need to irrigate the cardamom. So, the ones like us who don't have irrigation, in those people's cardamom fields the fruits will not be of good quality because of dryness. If the rainfall comes during that time, we will have good production but if there is no rainfall, we don't have that much hope for good production." Hence, a lack of agricultural investment capacity seems to entail lower agricultural productivity for poorer segments of society, potentially aggravating existing inequality. Spending priorities grant additional insight. About one third of the respondents mentioned that they had improved their house -this figure was stable when disaggregating the data by gender, caste, occupation, and age. However, marked caste and occupational differences became apparent regarding mobility and education. All respondents who reported to have bought a motorbike were non-Dalit farmers; neither Dalits nor agricultural laborers (the groups partly overlap) were among those who made such a major investment. Regarding education, 61% of the non-Dalit respondents mentioned that they had invested in the education of their children, while for the Dalit group, the share was 12% only. For farmers, this figure amounted to 50% and for laborers to 27%. These differences appear even more striking when compared with the Good Life concepts of these groups: Dalits and laborers placed greater emphasis on education than non-Dalits and farmers (see Table 1). Potentially, the income increases for Dalits and laborers were not substantial enough to cover both household needs and school fees. Possibly, these groups are economically disadvantaged, lacking the means to pursue the goals they have reason to value (cf. Sen, 1999). Given the above differences, we hypothesized that the gap between rich and poor might have increased. However, the leader of a women's cooperative argued that the opposite was true. She reasoned that the gap between rich and poor has declined: ''When there was no cardamom, the ones who had less land, they only had income from their own land for one or two months. For the rest of the year, they had to do paid labor [. . .]. But now, with cardamom, though their land is small they will get more income." When we asked a 46-year-old male Rai agricultural laborer whether or not people have benefitted equally from cardamom production, he said: ''All of the people have benefitted. For example, there are the ones who didn't have income before, but by selling cardamom they could earn some money. Likewise, some of the people have bought land in the neighboring district, and land for the house, like that. But in our case, we have not been able to add land. However, we now have some money to afford the education of our children." A 52-year-old male Dalit laborer explained that ''everybody benefitted. [. . .] It depends upon the land: those who have more land, they might have more income, the ones having less land, they have less income." To summarize, there are evident variations in the extent to which different people benefitted from commercialization, and these are at least partly determined by land size and the capacity to invest in irrigation. However, it seems like there were no real losers from agricultural change, as all population groups were able to reduce both financial and physical hardship to some degree.

Is a life without hardship life without farming?
Not only did the respondents use the notion of dukha to reason about their own labor and life; even more so, they invoked the notion of hardship when explaining their aspirations for their children's future. Across all ages, castes, genders, and economic situations, respondents consistently argued that their own life had been affected by hardship, and that the life of their children should be easier. For instance, a 33-year-old wealthy Rai farmer and businessman explained: ''Whatever dukha and sukha I have faced up to now, I do not want my son to get that much dukha. Therefore, he should go to a better school and get a better job." Concordantly, a 46-year-old Rai laborer said that if his children studied well, ''then they will not get as much dukha as I faced. If they do well in education, then they won't need to struggle as much as we did." As a result, village life is generally seen as involving hardship at different levels, as opposed to a life in the city. Speaking about the future of his grandchildren, a 63-year-old male Dalit farmer explained: ''I don't want them to stay in the village and do dukha. I would like them to go and live in town and study, not carry grass like we did. [. . .] If they stay in the village, they have to cut grass and carry loads. If they go to town, they can open a shop and that can be their job. No stress in that." Some parents emphasized that their children should be able to make their own choices. For instance, a 23-year-old Rai farmer said that his daughter should ''not get as much dukha as I faced. I have not had the opportunity to study well, so I wish to give higher education to her. But later, when she has grown up, she will decide what she wants to do." From the perspective of parents, a good life for their children is a life with little hardship, involving opportunities other than cardamom production. As much as the respondents are convinced that agricultural change contributed to reducing dukha, their ultimate vision of a Good Life involves a non-agricultural occupation because it is perceived to involve less hardship as compared to cardamom production.

Multidimensionality of local well-being concepts
The respondents' perspective on the Good Life is essentially multidimensional. Many of the eleven dimensions they cited overlap with findings from similar research in different regions of the world. For instance, respondents from African, Asian and Latin American countries alike emphasized social relationships as central to well-being of individuals, households, and communities (Beauchamp et al., 2018;Calestani, 2009;Greco et al., 2015;Hanrahan, 2015;Hoffmann & Metz, 2017;de L'Estoile, 2014;Lu & Gilmour, 2004;McGregor et al., 2009;Narayan et al., 2000). Likewise, education and the well-being of one's children are common themes (Beauchamp et al., 2018;Calestani, 2009;Greco et al., 2015;Kant et al., 2014;McGregor et al., 2009;Narayan et al., 2000). On the individual level, health is of central importance (Beauchamp et al., 2018;Bigler et al., 2019;Kant et al., 2014;McGregor et al., 2009;Narayan et al., 2000) as is having sufficient income and assets for covering one's basic needs (Beauchamp et al., 2018;Bigler et al., 2019;Caria & Domínguez, 2016;Greco et al., 2015;Kant et al., 2014;McGregor et al., 2009;Narayan et al., 2000). Accordingly, making money is usually not considered an end in itself; the income dimension was mostly framed as ''having enough" (Fischer, 2014;Narayan et al., 2000). In our case, some respondents explicitly emphasized this aspect as well; the majority, however, simply mentioned income as one Good Life aspect out of many. Further thematic overlaps between the Nepali concepts and other local definitions of the Good Life include peace in terms of both inner well-being and political security (Bigler et al., 2019;Greco et al., 2015;Narayan et al., 2000), happiness (Greco et al., 2015), as well as self-determination and freedom of choice (Narayan et al., 2000).
While in other regions men and women often emphasized different aspects in their concepts of the Good Life (Narayan et al., 2000), we could not determine any striking gender difference in our data. Moreover, in contrast to findings from other regions, the participants in our study did not mention the relationship with nature and land (Beauchamp et al., 2018;Caria & Domínguez, 2016;Kant et al., 2014;McGregor et al., 2009;Scott et al., 2018), nor culture (Kant et al., 2014;Scott et al., 2018) or spirituality (Calestani, 2009). Likewise, dignity, aspiration, and commitment to a higher purpose were valued elsewhere (Fischer 2014), but these aspects did not emerge from our data.
Notably, all the studies on local perceptions on well-being or the Good Life known to us portray multidimensional concepts. Many of these combine objectively measurable dimensions such as living standard, health or education with subjectively measurable dimensions such as happiness. Hence, local concepts of the Good Life -as elicited through our study and through similar research as discussed above -resonate best with academic wellbeing approaches that comprise both subjective and objective dimensions (e.g., Costanza et al., 2007;Diener & Tay, 2015;Gough & McGregor, 2007;White, 2010).

New perspectives on agricultural change using the local Good Life concept
In emphasizing the absence of hardship as the most salient dimension, the Nepali concept of the Good Life adds a new aspect to the well-being approaches we reviewed. This hardship perspective is rooted in the respondents' everyday life experiences shaped by their agricultural livelihoods and the burdens involved. Hence, an analysis of well-being in that region must incorporate the notion of hardship in addition to the other dimensions that form part of both existing academic well-being approaches and the Nepali concept of the Good Life.
It is important to note that the demarcation between physical, financial and emotional hardship as described in Section 4.1 is a distinction made by the authors to better understand the meaning of the term as used by the respondents. Financial, labor-related and emotional dukha did emerge from the data as conceptually delimitable types of hardship. However, this distinction was not emphasized by the respondents themselves. The male and female farmers and agricultural laborers consistently referred to the Good Life as a life without dukha, regardless of whether this dukha involved hard physical labor or worries about getting enough money for food on the following day. While it may be useful to separate different types of hardship when considering constructing a quantitative index, a qualitative approach that takes the respondents' understanding of the Good Life seriously must embrace an overarching hardship perspective. This is even more important in an analysis of agricultural change, as agricultural practices are both a cause of considerable hardship and a starting point for its relief. Farmers and laborers consistently used the notion of hardship to frame their perspectives on agricultural commercialization, and respondents across all socio-economic milieus evaluated the shift to cardamom positively on the grounds that it helped reduce dukha. This result is particularly important in the context of out-migration in Nepal and South Asia in general which increases labor constraints on those who are left behind, especially women (Aryal & Kattel, 2019;Devkota et al., 2020;Lahiri-Dutt & Adhikari, 2016). While agricultural commercialization has exacerbated existing labor constraints elsewhere (Brown & Waldron, 2013), this did not seem to be the case in our study area.
Could the hardship perspective be operationalized in quantitative terms? An attempt to do so would not only need to differentiate between physical and financial hardship: it would also need to acknowledge overlaps with other dimensions. For instance, financial hardship and income are closely related, but they are not congruent. We might expect a subjective financial hardship variable to be inversely correlated with an income variable up to a certain threshold after which further increases in income would not be associated with further reductions in hardship levels (similar to the relationship between income and subjective well-being, see Easterlin, 1974). Hence, an income measure based on a cut-off point rather than an open-ended continuous variable (as operationalized in multidimensional poverty indices, see Alkire et al., 2014) would reflect the respondents' perspectives on financial hardship.
Physical hardship, however, seems to be more difficult to measure. While time-use measurements are increasing in popularity as a proxy for workloads, especially with regard to gender in agriculture (for an example see IFPRI, 2012), this does not capture dukha which rather relates to the intensity of agricultural work. Attempts to measure the latter are scarce -possibly not least due to measurement difficulties (for an approach using accelerometry devices see Srinivasan et al., 2020). To capture physical dukha, a subjective indicator for physical hardship could potentially function analogously to the measures of subjective well-being, with similar advantages and drawbacks. In addition to facing the challenge of determining a meaningful measure for physical hardship, a quantitative operationalization of the hardship perspective would need to address the problem of overlap with other dimensions. Defining mutually exclusive dimensions might not always be possible, as Greco and co-authors pointed out in their case study of Good Life perceptions in rural Malawi (Greco et al., 2015). Alternatively, hardship could also be understood as a latent concept influencing other Good Life dimensions, without being directly measurable itself.

Limitations of agricultural commercialization in contributing to the Good Life
The analysis using the Good Life concept showed that agricultural commercialization had increased well-being in various ways at that point in time. However, one should be careful not to infer potential future effects due to ecological, economic, and social limitations. First, ecological sustainability is not granted as cardamom diseases have started to spread in the region (K. C. . Second, economic sustainability is threatened due to price fluctuation (ITC, 2017;Upreti et al., 2016) and inequality issues. While some respondents argued that even the poorer sections of society benefitted from commercialization in some way, from our observations of the different living conditions in the study area we suspect that previous inequalities have likely been reinforced through cardamom. Such a trend would be in line with commercialization processes in other regions where the comparatively wealthy benefitted more than the poor (Beck et al., 2016;Bigler et al., 2019;Brown & Kennedy, 2005;Dawson et al., 2016). Finally, parents strive to provide a good education to their children so they can avoid the financial and physical hardship engendered by farming. In the view of the respondents, it seems, agricultural commercialization is but one step on a path that eventually leads away from agriculture. In Nepal, like in other countries with a high share of agricultural GDP, working in agriculture often counts as an undesirable occupation (Agarwal & Agrawal, 2017) only pursued by those who did not succeed in make their living otherwise (Jones et al., 2017;Rigg, 2006). Such attitudes of course depend on the context -for instance, a recent study in Cambodia showed that the local concept of the Good Life was inextricably linked to agricultural land (Beauchamp et al., 2018), and research in Ethiopia demonstrated that the respondents' favorite activities in livestock keeping were precisely those that entailed the most physical hardship (Hertkorn et al., 2015). In contrast, the participants in our study aspire a life with little hardship for themselves and their children, even if this means that their children do not continue working on the family farm. This insight underscores the importance of incorporating local perspectives in analyses of agricultural change: a conventional study may have led us to the conclusion that commercialization had positive effects on income and food security, hence encouraging policy makers to further invest in the commercialization of agriculture. With the Good Life approach to agricultural change, however, we understand that income and food are only two out of several important dimensions. In emphasizing the notion of hardship, the respondents do acknowledge the improvements achieved through commercialization. Nevertheless, their aspirations of the Good Life may eventually lie outside the agricultural sector, because many of the non-agricultural professions are perceived to entail less physical and financial hardship.

Methodological limitations
Our study elicited a local concept of the Good Life which makes a meaningful addition to well-being concepts and analyses of agricultural change. Yet, it is insightful to critically reflect on the limitations of the method. For instance, while food, clothes, and health were mentioned by a substantial number of respondents, housing did not emerge as an important category. However, adequate shelter is an indispensable basic need (Streeten, 1981), and about one third of respondents indicated that they invested their increased income in improving their houses. This suggests that housing probably is a tacit dimension of the local Good Life concept. This noteworthy gap can be explained by the fact that in the study area almost everyone lives in private dwellings (Subedi & Upreti, 2019), so the respondents might take adequate housing for granted and thus might not think of mentioning this dimension when elaborating their perspective of the Good Life. In addition, topics like religion or domestic violence did not emerge from the interviews, but we assume that these topics are important: we observed the vital role of religion in everyday life in the study area, and we are aware that 15% of East Nepali women experience intimate partner violence (Dhakal et al., 2014). Potentially, dimensions like ''family" and ''peace" bear implicit reference to freedom from domestic violence, but the issue was never openly discussed. Arguably, themes like religion and domestic violence might not be the dimensions of life that are most affected by agricultural change. However, researchers investigating local well-being concepts with qualitative and particularly visual methods must recognize the absence of particular topics due to taboos, shame, or the respondents' desire to portray themselves in a good light (Pauwels and Mannay, 2020). These considerations demonstrate that while local perspectives can point researchers to important dimensions they might otherwise overlook, it is most useful to ground a wellbeing framework in both local perspectives and a set of indicators derived from theory and the results of previous research.

Conclusion
Agricultural commercialization policies in low-income countries affect millions of farmers and casual laborers in different parts of the world. Hence, it is important to investigate the effects of agricultural change on the lives of the local population. While well-being has been established as an important field in development studies, this trend apparently has not had a major impact on agricultural research: only rarely do studies of commercialization effects include local perspectives and multidimensional measures. Linking research on agricultural change with debates on well-being in development opens up nuanced perspectives on the effects of agricultural commercialization on different dimensions of rural life in low-income countries.
In this paper, we propose a new approach for investigating development outcomes through a concept of the Good Life as defined by local stakeholders. This approach could enable development researchers and practitioners alike to better understand the priorities of the people they work with. This is important because these priorities may differ from the priorities set by other development stakeholders. For the respondents in this case study, hardship is the most salient dimension, both in their definitions of the Good Life and in their perspectives on agricultural change. In contrast to the ten other Good Life dimensions that emerged through our analysis, the notion of hardship is not reflected in any of the well-being approaches we reviewed.
The shift to commercial cardamom production contributed to the Good Life through reducing financial and physical hardship. However, the positive effect of agricultural commercialization on the Good Life may be compromised by increases in inequality, a lack of economic and ecological sustainability, and the fact that most respondents conceptualized a life free from hardship outside agricultural livelihoods. Hence, while agricultural commercialization undoubtedly has contributed to the respondents' Good Life in the recent past, it is not possible to infer that fostering commercialization will have further hardship-reducing effects in future.
Our research is an example of how an assessment of commercialization effects can change when using locally defined categories of analysis: an investigation of income and nutrition indicators alone would have led us to a more positive outlook. Including holistic concepts of well-being is important because such an approach sheds a different light on the effects of agricultural commercialization. We therefore advocate increased collaboration between well-being scholars and researchers interested in the effects of agricultural change, especially in contexts where the commercialization of agriculture is an explicit policy goal. Using comprehensive concepts of well-being that combine local perspectives with academic indicators would allow for more rigorous and holistic analyses of the effects of commercialization on the rural population.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.