This is the accepted version of the article: Gamboa Jiménez, Gonzalo; Mingorría, Sara; Scheidel, Arnim. «The meaning of poverty matters : trade-offs in poverty reduction programmes». Ecological eco- nomics, Vol. 169 (March 2020), art. 106450. DOI 10.1016/j.ecolecon.2019.106450 This version is available at https://ddd.uab.cat/record/251852 under the terms of the license 1 The meaning of poverty matters: trade-offs in poverty reduction 2 programmes 3 Gonzalo Gamboa1,*, Sara Migorría2, Arnim Scheidel2 4 1 Associació Arran de Terra. C/ Torre dels pardals, 60, 1-4, 08032 Barcelona 5 2 Institute of Environmental Science and Technology, Universitat Autónoma de Barcelona (ICTA-UAB), 6 Edificio Z, Campus UAB. 08193, Bellaterra (Cerdanyola del Vallès), Spain. 7 * Corresponding author (gonzalo@riseup.net) 8 9 10 Abstract 11 Poverty has many different dimensions, yet few poverty reduction policies take an integrated 12 approach to multidimensional poverty. Many continue to focus predominantly on income and 13 employment issues whereas others address different aspects of poverty separately. In this 14 paper, we illustrate that while such policies may be effective at reaching their objectives 15 individually, they may clash with other dimensions of poverty targeted by other policies. To 16 achieve our aims, we employ a case from rural Guatemala, where a series of development 17 policies have pursued different targets, based on different narratives of poverty reduction. We 18 apply a multidimensional assessment framework and analyze the household typologies of three 19 rural communities to address how these typologies perform in relation to the contrasting goals 20 of different rural-development policies. While for some household types classic indicators such 21 as monetary income and employment did increase, a series of further issues targeted by other 22 policies, such as self-sufficiency, disposable time for community activities, or access to land, 23 worsened. Hence, the problem of focusing predominantly on one dimension is not only that it 24 provides an incomplete picture: the main problem is rather that it can obscure the creation of 25 new types of poverty. 26 27 28 Keywords: Q’eqchi’; Guatemala; multidimensional poverty; rural development 29 30 1 31 1. Introduction 32 Poverty has been increasingly conceptualized as a multidimensional phenomenon that involves 33 deprivation in many different dimensions of life. Well-known examples of multidimensional 34 approaches are Sen’s capability approach (Sen, 1999) and Max-Neef’s Human Scale 35 Development approach that underlines the existence of different poverties (Max-Neef et al., 36 1989). However, when it comes to development practice, poverty-reduction programs and 37 projects have been often dominated by income and employment approaches (Sumner 2007, 38 Konkel 2014). 39 Several empirical studies have demonstrated the importance of the use of different definitions 40 of poverty in poverty-reduction efforts (e.g., Caizhen, 2010; Haveman and Wolff, 2005; Laderchi 41 et al., 2003; Rojas, 2008; Scheidel, 2016). According to these authors, the approach used to 42 define and measure poverty determines the individuals and groups which will be categorized as 43 poor, and the policies aimed at poverty eradication. For instance, the OECD (2006) proposes 44 fostering agricultural development to reduce poverty through four different lines of approach: 45 by increasing farm incomes, by creating employment on farms, by promoting the rural non- 46 farming economy and by reducing prices of staple foods. However, the question remains as to 47 how far such policies can bring about positive change across other, non-economic dimensions 48 of poverty. 49 This paper demonstrates that the pre-analytical adoption of different narratives on poverty 50 leads to the design and implementation of different policies for poverty alleviation. This pre- 51 analytical choice also leads to non-equivalent assessments of the performance of rural 52 households in terms of poverty reduction. In other words, this paper provides further empirical 53 evidence that the choice of adopting certain poverty-reduction narratives is incredibly important 54 to poverty studies, policies and the poor. To illustrate this, we carry out a case study in 55 Guatemala, in which we (i) identify narratives about poverty in two rural-development policies; 56 (ii) identify the pertinent attributes needed to describe and represent poverty within these 57 different narratives.; and (iii) carry out an integrated assessment of households involved in 58 different rural-development policies from different perspectives (i.e., adopting a different 59 coupling of narratives and attributes). 60 To achieve our aims, we conduct a multidimensional analysis of different rural Q’eqchi’ 61 communities located in the Polochic Valley, which is characterized by communities who exhibit 62 various degrees of market participation. 63 We illustrate empirically that trade-offs between different poverty dimensions become evident. 64 Analogous to the distinction of weak and strong sustainability (Daly, 1990) we discuss aspects 65 of weak and strong poverty reduction (Scheidel, 2013). While policies and programs focusing too 66 narrowly on monetary income and employment may bring related improvements, they likewise 67 may force fundamental structural changes in the cultural and productive system of rural Q’eqchi’ 68 communities that cannot be substituted or compensated by enhanced incomes. Hence, a central 69 problem of focusing predominantly on one dimension of poverty reduction, such as monetary 70 income generation, is not only that it provides an incomplete picture of the situation of the 71 poor, but rather that it may obscure the creation of other types of poverties in the lives of rural 72 dwellers. 2 73 At this point it is interesting to shed light on the fact that, despite the use of multi-dimensional 74 approaches to measure poverty in in Guatemala 1, the predominant approaches to poverty 75 reduction observed in rural-development policies and programs in Guatemala are aimed at 76 generating formal employment and monetary income. This article illustrates how such policies 77 can have adverse impacts on other dimensions of the lives in rural communities. 78 The article proceeds as follows: Section 2 introduces our theoretical and methodological 79 framework, which provides the basis on which the multidimensional assessment is conducted. 80 Section 3 provides an overview of Guatemala’s rural-development policies and the case study 81 area. Section 4 describes the multidimensional assessment of the three communities 82 characterized by different degrees of market participation, and discusses how different trade- 83 offs across various poverty dimensions are produced and how these relate to structural changes 84 in the peasant economy and cosmovision. Section 5 discusses the implications that different 85 poverty-reduction narratives have for development policy and practice, and section 6 concludes. 86 87 2. Concepts and Methods 88 2.1. Multidimensional poverty: implications for rural studies and policies 89 Poverty may have different meanings for different social groups. The choice of different poverty 90 definitions and approaches to measuring poverty determines who is considered poor, as well as 91 the development of poverty-reduction efforts (Caizhen 2010; Haveman & Wolff 2005; Laderchi 92 et al., 2003; Rojas 2008; Scheidel 2013). 93 From an epistemological perspective, the interpretation of complex issues, such as poverty 94 reduction in rural households and communities, is undertaken through a set of narratives, 95 expectations and goals delimiting the problem at hand (i.e., the issue definition). In other words, 96 the pre-analytical adoption of different narratives about poverty leads to non-equivalent 97 representations of the system under analysis. By pre-analytical choice we refer to the decisions 98 about the relevant attributes used to describe and represent a system, which are made before 99 data collection and analysis. These decisions determine the scale and methods of observation, 100 and consequently the results of the analysis (Kovacic and Giampietro, 2015b). 101 A rural household can be described and represented using different attributes and indicators 102 (e.g., in terms of income per capita, in terms of literacy, in terms of access to healthcare or in 103 terms of access to productive land), and the same household can be considered poor from one 104 perspective, but not poor from a different perspective. As a result, choosing different narratives 105 of poverty leads to different assessments of the performance of rural households and 106 communities regarding poverty reduction. It also leads us to different conclusions in terms of 107 the level of poverty in households and communities, as well as the type of poverty alleviation 108 policies that are designed and implemented (Laderchi et al., 2003). 1 For instance, the National Living Conditions Survey (ENCOVI) carried out by the National Stadistical Institute (INE) uses indicators that go beyond income; such as, time use, access to education or health services 3 109 Furthermore, if we consider poverty reduction as an issue of long-term sustainability of 110 livelihoods, the distinction between weak and strong sustainability becomes relevant. According 111 to Daly (1990), the paradigm of weak sustainability is based on assumptions on the 112 substitutability of different types of capital; these imply that maintaining the total stock of 113 capital, no matter how it is composed, is enough for sustainability. Strong sustainability is based 114 on the assumption that different types of capital cannot be substituted, but rather are 115 complementary. As such they must be maintained independently. 116 Along these lines it is also possible to distinguish between weak and strong poverty reduction 117 (Scheidel, 2013). Weak poverty reduction assumes that improvements in one dimension of well- 118 being can compensate for deprivation in other dimensions. For instance, food security may be 119 attained by increasing income (a flow) through off-farm jobs that allow people to buy food, 120 which in turn can compensate for the loss of land (a fund, or asset) used to produce food. 121 However, there are situations in which improvements in one dimension cannot negate the 122 deterioration in other poverty dimensions. For instance, an increase in (short-term) flows (e.g., 123 money, or food) to enhance deprived consumption cannot always compensate or serve as a 124 substitute for the loss of underlying funds (e.g., fertile land, healthy labor conditions) that allow 125 for the production of such flows in the long-term. Hence, strong poverty reduction needs to be 126 attentive to such situations where trade-offs are not desirable, and must focus on increasing 127 access to and control over the productive funds that enable producing flows and living a 128 dignified life in the long-term (for further theoretical discussion, see Scheidel, 2013). 129 The objective of this article is to demonstrate the existence of relevant trade-offs and related 130 aspects of weak and strong poverty reduction, by showing the consequences of adopting 131 different poverty narratives in the classification of households as poor, and in the design and 132 implementation of poverty-alleviation policies. To attain this objective, we performed the 133 following steps: 134 1. Reviewed Guatemalan rural-development policies to identify different narratives 135 behind those policies and programs. 136 2. Chose the policies with extreme narratives underlying their understanding of poverty. 137 3. Defined attributes relevant for each narrative for describing and representing poverty. 138 4. Defined different indicators to measure and represent attributes. 139 5. Evaluated these indicators for the households of the case study. 140 In the following sections, we present the main aspects of each of these steps. 141 2.2. Case study approach 142 The success of poverty-alleviation programs and policies is highly dependent on the specific 143 context in which they are implemented. Therefore, this study is based on the empirical analysis 144 of a case study; such an analysis allows us to understand this issue (i.e., poverty reduction) and 145 perform an in-depth analysis of a real context, rather than simply allowing us to look for 146 statistical generalizations (Ford et al., 2010; Yin 2003). The case study approach is also 147 appropriate for analyzing complex problems and systems, in which the main research questions 148 start from a HOW (Robinson 2008). Our questions are: How does the pre-analytical choice of a 149 given narrative on poverty determine policy design and implementation? And, how does the 4 150 pre-analytical choice of a given narrative on poverty determine both the households considered 151 to be poor and the policies aimed at poverty eradication? 152 2.3. Defining narratives, attributes and indicators 153 Narratives are understood here as stories that identify the relations of causality used to 154 structure the perception of the observed system (Magrini 1995, Allen and Giampietro 2006; 155 Kovacic and Giampietro 2015a). In this way, narratives define the relevant attributes to be 156 considered when dealing with poverty-eradication policies. 157 In this article, we analyze in particular the importance of the pre-analytical choice of narratives 158 behind the definition of poverty, and the problems inherent in structuring the national strategies 159 for poverty alleviation. To do so, we first analyze the strategies, policies and programs dealing 160 with rural development and rural poverty eradication in Guatemala. Then, we chose two policy 161 documents based on purposive sampling: a non-probabilistic sampling of individuals with some 162 characteristics relevant to addressing the research questions. In this case, we chose two extreme 163 policies that define poverty in very different ways, two clear-cut instances of the studied 164 phenomenon (Given, 2008). These are the Competitiveness Agenda (Government Agreement 165 No. 306-2004) and the National Policy of Integrated Rural Development (Government 166 Agreement No. 196-2009). In this way, we aim to contrast the official definition of the issue and 167 the problem of structuring poverty (Competitiveness Agenda) against the narratives and formal 168 representation used by social and peasant movements (Integrated Rural Development). 169 Secondly, we analyze the chosen documents according to the following issues: i) objective of the 170 policy, ii) approach to rural development, iii) actors prioritized by the policy, iv) approach to 171 agriculture, v) food, vi) land and vii) employment. In this way, we identified the main attributes 172 used to perceive and describe rural poverty. Attributes are the essential elements used within 173 the specific narrative to describe a system. For example, the assertion that “rural areas present 174 low employment rates that are one of the main causes of poverty” contains a value judgment, 175 which is used to identify “rural employment” as a relevant attribute within this sentence. 176 Thirdly, we have defined formal categories to map these attributes and permit the measurement 177 of the state of the system according to this attribute: i.e., the definition of indicators used to 178 perform a quantitative characterization of the system under study. Indicators are thus a means 179 of representing an attribute of the system – an image of an attribute, which is formalized in 180 terms of a specific measurement process (Galopin, 1997). For example, the number of people 181 employed in agriculture can be used as the indicator for the attribute “rural employment.” 182 In order to define and quantify indicators, we used the accounting framework of the Multi-Scale 183 Integrated Assessment of Societal and Ecosystem Metabolism (MuSIASEM) approach 184 (Giampietro et al., 2009). The MuSIASEM approach uses the flow-fund model (Georgescu- 185 Roegen, 1971), which distinguishes between fund elements as structural components of a 186 system, and flow elements that are processed by the system and exchanged with its context. 187 Fund elements analyzed in this study are human beings and Ricardian land, measured in human 188 time and land-use surface respectively. Human time and land are not only the main production 189 factors but are also important biophysical constraints for the production and reproduction of 190 peasant households (Grunbuhel and Schandl, 2005). Analyzed flows are income, expenditures, 191 and maize production and consumption. 5 192 Based on this approach, a large number of indicators have been developed for the analysis of 193 households’ metabolic pattern, which are described in detail further in Mingorria (2016). For the 194 purpose of this article, a set of six indicators was combined to assess the performance of rural 195 households in terms of poverty reduction in relation to the two different policies. 196 2.4. Data collection 197 Data were collected between 2009 and 2011 in three non-consecutive fieldwork periods, using 198 a mixed methodological approach encompassing in-depth interviews and questionnaires 199 (Huntington 2000). In the first period (March to June 2009), we conducted 12 semi-structured 200 interviews with indigenous leaders, NGO members and representatives of peasant movements 201 involved in the valley’s land struggles. The aim of these interviews was to identify the main 202 characteristics of the communities of the valley: their production systems, forms of organization 203 and participation in policies and programs. 204 In the second period (July to November 2009), male and female leaders from the selected 205 communities were interviewed. Interviews at community level were structured using five 206 themes: (1) the main productive and reproductive activities of the households; (2) the calendar 207 of agricultural seasons; (3) the traditional and formal rules affecting natural resource 208 management; (4) the socio-environmental history of the communities; and (5) the main 209 constraints on fulfilling their developmental needs and aspirations. 210 Interviews were undertaken to understand the socio-economic and environmental context in 211 which households and communities behave. Also, the questionnaire for the land- and time-use 212 survey was designed according to the information gathered in these interviews. Furthermore, 213 we use the interviews to identify the main narratives regarding poverty alleviation and rural 214 development from the point of view of communities. 215 Subsequently, during the data-analysis phase, we used the interviews to obtain detailed 216 knowledge about the socio-economic context and the people’s livelihoods, and this helped us 217 to translate the attributes into indicators. Finally, the results and insights gained from our data 218 analysis complemented the quantitative results of our surveys. 219 In the third period (May 2010 to February 2011), we conducted 10 intermittent field visits, each 220 lasting 15 days, and deployed 196 questionnaires in the selected communities, with households 221 being selected randomly in both settings. The questionnaires were previously tested in both 222 communities and structured in five sections: (1) demographic structure; (2) land use; (3) income 223 and income-generating activities; (4) household expenditure; and (5) the time use of individual 224 household members. The basic unit of the time- and land-use survey was the household, since 225 our observations confirmed that the household was the key institution of the Q'eqchi' 226 communities where decision-making takes place (Grandia 2006). We interviewed both female 227 (N=98) and male heads of household (N= 98), collecting information on all household members. 228 Participant observations were used during the entire research process along with cross-checking 229 and validating the surveys' approaches and emerging results. 230 6 231 2.5. Data analysis 232 Q'eqchi' households shape and are shaped by the community, and a mutual dependency 233 between these two levels exists. On one level, households are the basic decision units for time- 234 and land-allocation issues (Mingorría et al. 2014. See also Netting, 1995; Chapter 2). Also, the 235 community imposes some constraints on the households, which then adjust their behavior to 236 cope with a reduced degree of freedom. For example, when the community develops a collective 237 project, the households are not completely free to decide how much time to allocate to these 238 activities: there is a minimum requirement from the community. In most of these cases, the rest 239 of the network (the community) is so strong that it is very difficult to undertake a significant 240 restructuring of the community (Giampietro, 2003). 241 In this context, we consider the household to be the unit of analysis. It is worth noting that the 242 households of a community do not all perform the same activities; nor do they have the same 243 metabolic pattern. However, the characteristics of the community heavily influence the 244 different production and reproduction strategies (the activities) of the households (Mingorría 245 and Gamboa 2010, Mingorría et al. 2014). In order to simplify this complex reality, we base our 246 analysis on households’ typologies. According to Giampietro (2003), a type is a simplification of 247 real entities, a representation based on expected relations between the components of the 248 entity, which gives rise to an expected behavior. The characteristics of a type are always 249 associated with the possibility of performing a given and expected function. 250 The definition of household typologies is achieved by means of a cluster analysis process. This 251 clustering process starts with the selection of a set of variables used to classify the households. 252 As mentioned previously, human time and agricultural land are the main production factors of 253 the peasant economies in which productive capital is scarce. They also constitute the main 254 constraints on the intensification of agrarian practices (i.e., on increasing yields by means of 255 increasing the workload involved in agricultural activities). The selection of these preliminary 256 variables is oriented by the objectives of the analysis and based upon the acquired knowledge 257 of the communities and their contexts. In this case, we chose twenty-nine variables describing: 258 i) the demographic structure of the households, ii) the agricultural system developed by the 259 households (land use), and iii) participation in the labor and food market (see Mingorria et al., 260 2014). 261 A Principal Component Analysis (PCA) was performed in order to identify the factors behind the 262 socio-economic differences among households. Subsequently, an agglomerative hierarchical 263 cluster analysis (HCA) employing Euclidean distance and Ward’s method was implemented. The 264 HCA was performed using those factors obtained from the PCA with an eigenvalue higher than 265 1 (Kaiser criterion). The number of clusters (i.e., household typologies) was determined to serve 266 the purposes of the analysis (Köbrich et al., 2003) and was based on the researchers’ experience 267 and the knowledge acquired through empirical observations (Garmendia and Gamboa, 2012). A 268 non-parametric Kruskal–Wallis test followed by Dunn’s multiple comparison tests were applied 269 to test differences among the household types for each of the indicators considered. 270 7 271 2.6. Study area: the Polochic Valley 272 The Polochic Valley is located in the northeast of Guatemala in the Alta Verapaz and Izabal 273 departments. Geographically, the area is bounded by the Sierra de Santa Cruz mountain range, 274 the National Protected Areas of the Sierra de las Minas and Bocas of the Polochic. 275 We chose the Polochic case study as it is one of the Guatemalan regions with the highest rates 276 of poverty (ENCOVI 2006). It is also a territory in which poverty reduction policies have been 277 implemented by integrating the peasantry in the agribusiness market (Alonso-Fradejas et al 278 2012) and, at the same time, the indigenous communities have historically struggled to 279 overcome poverty without renouncing their traditional ways of living (Mingorría et al 2014). 280 Approximately 220,000 people inhabit the valley and rely on subsistence agriculture (INE 2002). 281 Of this population, 89% are indigenous Mayan Q'eqchi' and the rest are Mayan Poq´omchi´ and 282 mestizos (i.e., people of Spanish and indigenous origin) (ENCOVI 2006). 283 Since the 1980s, the valley has experienced an increasing expansion of an agro-export model. 284 At that time, coffee, cotton and banana were grown in large areas of land granted to German 285 families from the liberal government of that time. Maya Q’eqchi’ people were forced to migrate 286 or become mozos-colonos (i.e., people who worked for the landowner in exchange for a small 287 plot of land on which they could cultivate subsistence crops). Since then, the Q’eqchi’ have 288 claimed access to land, which has been ignored and often violently repressed (Grandia 2006). 289 As of the early 2000s, coffee farms and the rearing of livestock were affected by the coffee crisis 290 and a decrease in prices, respectively (Wagner 2001). Nowadays, most of the valley is dominated 291 by sugarcane and oil palm plantations and less so by cattle farms (Solano and Solís 2010). Since 292 1998, the valley has been covered by 8,500 ha of cultivated oil palm plantations, representing 293 almost three-quarters of the valley’s most fertile land, and, since 2005, by more than 5,000 ha 294 of sugarcane (Alonso-Fradejas et al., 2008; 2011; Mingorría and Gamboa 2010). 295 Throughout this time, the Q’eqchi’ people have maintained moral economies 2 based on 296 subsistence agriculture complemented by other sources of income (Grandia 2012). The majority 297 of Q’eqchi' communities produce maize for subsistence (INE 2002), but they differ in terms of 298 the degree and forms of market integration (Alonso-Fradejas et al., 2008, Mingorría and 299 Gamboa 2010). The main income-generating activity in the mountain area is the cultivation of 300 cardamom and coffee as traditional agro-export crops. The communities located in the valley 301 produce and sell surplus maize, and the money generated from this accounts for an important 302 part of their income; people may also work either for other farmers and/or on cattle ranches, 303 and oil palm plantations (Molina-Loza et al., 2009; Ronzon and Till 2004). 304 For this study we selected four communities that represent different degrees and forms 305 (individual and collective) of market participation promoted by policies oriented toward 306 reducing poverty in the Polochic Valley: a) two communities located in the mountains that 307 practice subsistence agriculture complemented by traditional export crops (cardamom) 2 According to Grandia (2006), "after saving enough for family food security, farmers can easily sell leftover corn to middlemen and use the proceeds to buy their basic household necessities, like medicines, shoes, clothes, tools, and daily comestibles (sugar, oil, coffee)." This confirms that household security is prioritized over cash accumulation, following a "safety first" principle in terms of behavior. 8 308 cultivated collectively; b) one community located in the valley whose households produce maize 309 and sell surplus in the market; and c) one community located in the valley whose members are 310 waged labor in plantations. 311 3. Results 312 We now show how different household types within these communities perform in relation to 313 the objectives of different poverty-reduction policies. To do so, we first describe the two most 314 contrasting policies for poverty eradication and outline the set of related narratives, attributes 315 and indicators that derives from these narratives for both policies. The indicators that we 316 developed are then used to evaluate household performance of different household types in 317 terms of poverty alleviation. 318 3.1. Narratives, attributes and indicators of two contrasting rural- 319 development policies 320 As mentioned in the Methodology section, we chose two policies with contrasting narratives on 321 poverty and poverty alleviation. These are the Competitiveness Agenda and the National Policy 322 of Integrated Rural Development, as presented in Table 1. 323 Table 1. Main characteristics of the analyzed policies for poverty alleviation PNDRI PRONACOM Government Agreement No. 196- Government Agreement No. 306- 2009 2004 Livelihood narrative Market narrative Objective Overcome poverty, inequality, social Improve the quality of life of the and political marginalization Guatemalan people and promote economic growth by means of fostering competitiveness Rural Development Advance toward a dignified and just Reduce the lack of employment and life in economic, social, political, opportunities to generate income; cultural, environmental and spiritual improve the precarious labor terms conditions, access to credit and the productive and basic service infrastructures (drinkable water, electricity, etc.) Priority sector Indigenous or peasant communities The entire population, especially the with insufficient land or without rural poor land, and seasonal or permanent paid workers Agriculture Improve efficiency and equity, Increase competitiveness by means of diversify and promote the participating in national and production of basic grains international markets, facilitating access to credit and through public and private investment Food Food sovereignty Food security (availability, access and (availability, access and consumption consumption of food that is of food) adequate in social and cultural terms) Land Transform the structure of land Regulate access to and through tenure and use, avoiding land market-led agrarian reform concentration 9 Employment Improve capabilities of the rural Increase employment by means of population in order to increase promoting both participation in employment and the quality of jobs national and international markets, access to credit and public-private investment 324 325 Based on these policies, two main narratives have been identified that we call the livelihood and 326 the market narratives. Both narratives aim to improve the quality of life of poor people. The 327 former focuses on facilitating poor people’s access to land by transforming the structure of land 328 use and tenure. It further puts forward the concept of food sovereignty by promoting the ability 329 of small peasants to produce culturally appropriate food by improving and diversifying 330 production techniques. In contrast, the market narrative proposes participation in national and 331 international markets, access to credit and the promotion of public-private investment as a 332 means of creating jobs and incomes. In this way, it is argued that poor families can improve their 333 quality of life and access to food and land. 334 We acknowledge that choosing these two policies, with the consequent identification of two 335 narratives, may seem an oversimplification of a long and contentious debate around poverty 336 conceptions and ways of measuring it. Guatemalan scholars have made important efforts in 337 measuring, analyzing and understanding rural poverty in a multidimensional way (e.g. Romero 338 and Zapil 2009, Romero 2015). However, choosing two policies and analyzing two competing 339 narratives about poverty reduction is an instrumental choice to achieve the purpose of this 340 article: to show how pre-analytical choices determine the relevant attributes to represent and 341 describe the system, the results of the analysis, and the design and implementation of public 342 policies. 343 Based on the content analysis of these contrasting policies, Table 2 presents the set of attributes 344 and indicators that we used to evaluate the performance of different household typologies in 345 terms of poverty. 346 Table 2. Attributes and indicators used to evaluate household performance according to 347 different narratives (source: own elaboration) Narrative Attribute Indicator Description Livelihood Access to land Land use: maize Surface land cultivated with maize Self-sufficiency Self-supply of maize Share of maize consumption from own production Workload Saturation of work Share of disposable human activity allocated to paid and unpaid work. Disposable human activity is the total amount of hours in a day minus time allocated to physiological overheads (i.e., sleeping, eating and personal care) Market Income Net Income Net income per capita Economic labor productivity Return to labor Net income per hour worked Consumption of food Maize consumption Amount of maize consumed per equivalent consumer 10 348 349 3.2. Performance of households 350 Five household typologies were noted in the sample. The Mix-traditional group of households 351 comprise the communities living in the mountain area. They undertake subsistence agriculture, 352 complemented with cardamom cultivation as a cash crop. We also noted two household 353 typologies within each of the other two communities: two household typologies producing 354 maize surplus for the market (Cash crop-Land and Cash crop-No land) and two household 355 typologies providing labor to the oil palm plantations (Oil palm-Land and Oil palm-No land). As 356 their names reflect, in both cases land tenure is the main characteristic that differentiates the 357 typologies of communities located in the valley. 358 Figure 1 illustrates the evaluation of these five household typologies according to the different 359 indicators of the market and livelihoods narratives. 360 361 Figure 1. Performance of household typologies according to the market and livelihoods narratives. 362 Note: Arrows next to indicator’s name denote whether the indicator is for maximizing or for minimizing. 363 Source: own elaboration based on Table A1, Appendix. 364 According to the indicators selected in the market narrative the household typology “Oil palm- 365 No land” presents higher income per capita3, followed by households of the “Oil palm-Land” 366 typology. The main differences between the two are that the former have no land and are 367 smaller households made up of younger people. These households can allocate a larger amount 368 of their available time to work in oil palm plantations and obtain a higher monetary income per 369 capita than the rest. Oil palm-Land are followed by the households that focus on producing 370 maize for the market. The same difference that we noted in the previous case also applies here. 371 Smaller households (i.e., “Cash crop-No land”) are able to obtain a higher net income per capita 372 by producing maize surplus for the market. 3 Monetary indicators are measured in Quetzales (Q). As of 2010, 1 US dollar was equal to 8 Quetzales. Therefore, Net income per capita ranges between 80 US$/year p.c. and 465 US$/year p.c. On the other hand, Return to labor ranges between 0.5 US$/h and 1 US$/h. 11 373 In terms of return to labor, the “Cash crop-No land” households generate almost double the 374 income per hour of human activity allocated to cash crop production compared to those 375 households whose members work in oil palm plantations. This reflects the fact that these 376 households prioritize participating in the market by selling large proportions of their maize to 377 the market (see Self-sufficiency below). 378 Finally, under the market narrative, we can see that all household typologies, except the “Oil 379 Palm-No land” type, consume more or less the same amount of maize per capita. As mentioned 380 previously, these households obtain higher income per capita than the rest, which enables them 381 to buy larger amounts of maize per capita on the market that complements their own 382 production. 383 According to the indicators selected in the livelihood narrative, the results are very different. For 384 “Mix-traditional” households, medium to high levels of land that were allocated to maize 385 production were noted, compared to the rest of the household typologies. For “Oil palm-Land” 386 households, a higher level of land use was noted, but this is due to the larger amount of owned 387 land. In fact, these households have lower productivities (by about half and one-third) than the 388 Cash crop households, owing to the lower levels of human time allocated to this activity 389 (Mingorría et al. 2014). 390 The prioritization of participating in the market is also reflected in the indicator self-sufficiency. 391 “Mix traditional” households perform better in this regard, followed by “Cash crop-Land” 392 households. The latter are larger households with access to land: They try to find a balance 393 between producing maize for the market and for their own consumption. Smaller households of 394 younger people (i.e. “Cash crop-No Land” and “Oil palm-No land”) present lower figures for self- 395 sufficiency, which reflects the lack of access to land and their prioritization of obtaining income 396 from the market in order to survive. 397 Finally, the saturation of work indicators demonstrates that the members of the “Mix- 398 traditional” households have a smaller workload than the others, which gives them more time 399 for communitarian work and organization (Mingorria et al. 2014). “Oil palm–No land” 400 households stand out in this regard, since they allocate three-quarters of their available time to 401 paid work activities. Furthermore, one can evaluate the degree of integration into the market 402 by calculating the saturation of paid work, which is the share of disposable time allocated to 403 paid work activities. At one extreme we have the “Mix-traditional” households, which allocate 404 6% of their disposable time to market activities; and at the other extreme we have “Oil palm-No 405 land” households, which allocate 21% of their disposable time to market activities. 406 In summary, we can posit that households participating in policies aimed at incorporating 407 peasants into the market (i.e., “Cash crop-No land,” “Oil palm-Land” and “Oil palm-No land” 408 households) obtained greater flows of money and food from the market. On the other hand, 409 households that tried to find a balance between subsistence agriculture and participation in the 410 market were able to keep the workload at half of their disposable time (i.e., take care of 411 themselves), and self-supply larger amounts of maize because they had access to enough land. 412 12 413 4. Discussion 414 The previous section has demonstrated that depending on which poverty narrative is used, 415 household types perform very differently. In this final section, we discuss two relevant 416 implications that arise from conceptualizing poverty as multidimensional: first, the existence of 417 trade-offs between different poverty dimensions; and second, the difference between weak and 418 strong poverty reduction (Scheidel, 2013). 419 We have noted that household types that exhibited positive performance under a market 420 narrative (i.e. PRONACOM policy), exhibited a comparatively bad performance under the 421 livelihood narrative (i.e. PNDRI policy). For instance, for “Oil palm-No land” households 422 approximately double the poverty threshold of 234 US$/year per capita. However, these 423 households allocated 75% of their disposable time to work (dedicating little time to community 424 activities and to producing food) and had very limited access to land, meaning that they 425 depended on maintaining their work in oil palm plantations to ensure they could access food 426 and cover basic needs. Moreover, age is an important limitation to work in oil palm plantations; 427 men older than 30 have little possibilities to keep working there and maintaining income from 428 this activity (Mingorría et al 2014). Both, limited access to wage work and land may hinder the 429 livelihoods of “Oil palm-No land” households in the near future. Hence, trade-offs between 430 poverty dimensions exist and need to be carefully considered in the design of policies and 431 programs, in order to avoid them becoming counterproductive. 432 Then, as evidenced by the case above, we can say that PRONACOM is a weak poverty-alleviation 433 policy, whilst the PNDRI would be a strong poverty-alleviation policy. The former focuses on 434 increasing the income of poor families by increasing their competitiveness, facilitating access to 435 credit and fostering their participation in national and international markets. Higher incomes 436 allow families to access food through purchase on the market, compensating thus for declines 437 in production for self-supply. Access to land may be obtained through credit, however, this also 438 becomes subject to the conditions as well as social and economic consequences of credit and 439 debt (Gerber, 2013). In these situations, access to monetary flows and capital thus substitutes 440 for, or conditions, direct control over other flows and assets. On the other hand, the PNDRI aims 441 to change the structure of land tenure, avoiding land-concentration processes and supporting 442 poor families’ access to land through land reform and land redistribution. Access to food would 443 be fostered through diversification in agricultural production and improvements in productivity 444 on the basis of technical advice given. 445 Participation in weak poverty-alleviation policies has fostered structural changes in the cultural 446 and productive systems of Q'eqchi' communities. Household typologies based on cash-crop 447 cultivation and the provision of labor to oil palm plantations must allocate a larger proportion 448 of their time to the new economic activities compared to households practicing more traditional 449 activities. This results in a lack of time for maintaining social and community structures, such as 450 those mechanisms which serve to help people confront and solve conflicts, manage the 451 commons, or represent the community in official institutions, among others (Mingorría et al 452 2014, Mingorría 2016). These changes in social and community structures have also influenced 453 how these households have invested in additional funds relevant for the long-term sustainability 454 of their livelihoods. For example, the “Oil palm” households have mainly invested in domestic 13 455 appliances and motorbikes4, while the “Cash crop” households have not invested in any 456 additional capital fund. The communities in the mountains, with mainly “Mix-traditional” 457 households, have invested in a cardamom drier and in a communitarian stock-breeding project. 458 In this way, these communities are investing in productive funds in order to improve their 459 livelihoods and quality of life over the long-term, thus engaging in strong poverty reduction. 460 5. Conclusions 461 Nowadays, most scholars agree that poverty is a multidimensional phenomenon. Yet few 462 policies and programs take an integrated approach to multidimensional poverty reduction, but 463 rather focus on different aspects separately, which are based on different understandings of, 464 and narratives on, poverty. This however, has important implications for the design of rural- 465 development policies, as well as for the related actions that affect rural communities. 466 Based on an empirical case study of Q’eqchi’ communities in Guatemala, this paper has 467 illustrated how two policies with the same broad aim of poverty reduction can lead to very 468 different assessments of how rural households and communities perform. Rural households 469 participating in weak poverty alleviation policies (i.e. PRONACOM) have better performance in 470 terms of income but are dependent on maintaining monetary flows to access food and cover 471 basic needs. On the other side, households participating in strong poverty alleviation policies 472 are less dependent on the market to access food and have more available time to maintain social 473 and community structures. 474 The paper has also illustrated that these assessments are related to differences in how the 475 communities organize their livelihood systems. While these policies may achieve their particular 476 goals, they have also produced a series of trade-offs across other poverty dimensions. Some 477 trade-offs may be acceptable for rural communities. Other trade-offs, however, need to be 478 carefully considered, particularly when communities may lose access to important funds (such 479 as fertile land) that would allow them to access a dignified life in the long-term. 480 Hence, the pre-analytical choices for defining a narrative on poverty, which were adopted during 481 the process of policy making, have crucial implications for rural communities. In order to deal 482 seriously with multidimensional poverty reduction, it is not only necessary to set up several 483 programs targeting different dimensions, but to also seek their integration by carefully 484 considering the possible trade-offs as well as new poverties that may be created. 485 486 6. References 487 1. Allen T.F.H. and Giampietro M. Narratives and transdisciplines for a post-industrial 488 world. Systems Research and Behavioral Science 2006, 23, 1-21 489 2. Alonso-Fradejas A., Alonzo F., Dürr J., 2008. Caña de azúcar y palma africana: 490 combustibles para un nuevo ciclo de acumulación y dominio en Guatemala. Guatemala: 491 Magna Terra Editores 4A motorbike can also be a strong livelihood asset, or fund, enabling quite a lot of new livelihood activities. However, the motorbike also depends on flows of money and gasoline to work. 14 492 3. Alonso-Fradejas A., Caal J. L., Chinchilla T., 2011. Plantaciones agroindustriales, 493 dominación y despojo indígena-campesino en la Guatemala del s. XXI. Guatemala: 494 Magna Terra Editores 495 4. Barham B., Carter M. R., and Sigelko W. 1995. Agro-export production and peasant land 496 access: Examining the dynamic between adoption and accumulation. Journal of 497 Development Economics, 46(1), 85-107. 498 5. Caizhen L., 2010. Who is poor in China? A comparison of alternative approaches to 499 poverty assessment in Rural Yunnan. J. Peasant Stud. 37, 407–428 500 6. Daly H.E., 1990. Sustainable development: from concept and theory to operational 501 principles. Population and Development Review 16, 25–43 502 7. Ellis F., Biggs S., 2001. Evolving Themes in Rural Development 1950s-2000s. 503 Development Policy Review 19, 437-448 504 8. Encuesta Nacional de Condiciones de Vida – ENCOVI, 2006. Instituto Nacional de 505 Estadísticas, Gobierno de Guatemala 506 9. Gallopín G., 1997. Indicators and Their Use: Information for Decision-making. In: Moldan 507 B. & Billharz S. (Eds) Sustainability Indicators. Report of the project on Indicators of 508 Sustainable Development. Scope 58. 509 10. Gamboa G., Kovacic Z., Di Masso M., Mingorría S., Gomiero T., Rivera-Ferré M., 510 Giampietro M. 2016. The Complexity of Food Systems: Defining Relevant Attributes and 511 Indicators for the Evaluation of Food Supply Chains in Spain. Sustainability 8, 515. doi: 512 10.3390/su8060515 513 11. Garmendia E., Gamboa G. Weighting social preferences in participatory multi-criteria 514 evaluations: A case study on sustainable natural resource management. Ecological 515 Economics 84, 110–120 516 12. Georgescu-Roegen N., 1971. The Entropy Law and the Economic Process. Harvard 517 University Press, Cambridge, MA 518 13. Gerber, J.-F. 2013. The Hidden Consequences of Credit: An Illustration from Rural 519 Indonesia. Development and Change, 44(4), 839–860. 520 https://doi.org/10.1111/dech.12045 521 14. Giampietro M., 2003. Multi-scale integrated analysis of agroecosystems. CRC Press LLC, 522 Florida 523 15. Giampietro M., Mayumi K., Ramos-Martin J., 2009. Multi-scale integrated analysis of 524 societal and ecosystem metabolism (MuSIASEM): Theoretical concepts and basic 525 rationale. Energy 34, 313–322 526 16. Grandia L. (2006). Land dispossession and enduring inequity for the Q’eqchi’ Maya in 527 the Guatemalan and Belizean frontier colonization process reactions. Berkeley, CA: 528 University of California-Berkeley 529 17. Grandia L. 2012. Enclosed: conservation, cattle, and commerce among the 530 Q’eqchi’Maya lowlanders. University of Washington Press. 531 18. Grünbühel C.M., Schandl H., 2005. Using land-time-budgets to analyze farming systems 532 and poverty alleviation policies in the Lao PDR. International Journal of Global 533 Environmental Issues, Vol. 5, Nos. 3/4. 534 19. Haveman R., Wolff E., 2005. The concept and measurement of asset poverty: Levels, 535 trends and composition for the U.S., 1983–2001. J. Econ. Inequal. 2, 145–169 15 536 20. INE. 2002. XI Censo Nacional de Poblacio´n and VI de Habitación. Guatemala 537 Government 538 21. INE. 2006. Encuenta Nacional Sobre Condiciones de Vida. Instituto Nacional de 539 Estadística. Guatemala. 540 22. Isakson S. R. 2014. Maize diversity and the political economy of agrarian restructuring 541 in Guatemala. Journal of Agrarian Change, 14(3), 347-379. 542 23. Isakson S. R. 2011. Market provisioning and the conservation of crop biodiversity: An 543 analysis of peasant livelihoods and maize diversity in the Guatemalan highlands. World 544 Development, 39(8), 1444-1459. 545 24. Kakwani N., Silber J., 2008. Introduction: Multidimensional Poverty Analysis: Conceptual 546 Issues, Empirical Illustrations and Policy Implications. World Development 36, 987-991 547 25. Köbrich C., Rehman T., Khan M., 2003. Typification of farming systems for constructing 548 representative farm models: two illustrations of the application of multi-variate 549 analyses in Chile and Pakistan. Agricultural Systems 76, 141–157. 550 26. Konkel R., 2014. The monetization of global poverty: the concept of poverty in World 551 Bank history, 1944–90. Journal of Global History 9(2). pp 276 – 300. DOI: 552 10.1017/S1740022814000072, 553 27. Kovacic Z., Giampietro M., 2015a. Empty promises or promising futures? The case of 554 smart grids. Energy, 93:67-74. 555 28. Kovacic and Giampietro, 2015b. Beyond “beyond GDP indicators:” The need for 556 reflexivity in science for governance. Ecological Complexity, 21: 53-61. DOI: 557 10.1016/j.ecocom.2014.11.007 558 29. Laderchi C.R., Saith R., Stewart F., 2003. Does it Matter that we do not Agree on the 559 Definition of Poverty? A Comparison of Four Approaches. Oxford Development Studies 560 31, 243 - 274 561 30. Lakoff G. Why it Matters How We Frame the Environment. Environmental 562 Communication 2010 4 (1), 70-81. DOI: 10.1080/17524030903529749 563 31. Magrini T. 2016. Ballad and Gender: Reconsidering Narrative Singing in Northern Italy 564 Ethnomusicology Online 1995. Available at: 565 http://www.umbc.edu/eol/magrini/magrini.html (Accessed: 15-January-2016) 566 32. Max-Neef M., Elizalde A., Hopenhayn M. 1989. Human Scale Development: An Option 567 for the Future. Development dialogue: a journal of international development 568 cooperation 1, 7-81 569 33. Mingorría S., Gamboa G. 2010. Metabolismo socio-ecológico de comunidades 570 campesinas Q’eqchi’ y la expansión de la agro-industria de caña de azúcar y palma 571 africana: Valle del Río Polochic, Guatemala. Instituto de Ciencia y Tecnología 572 Ambientales – UAB, Instituto de Estudios Agrarios y Rurales – CONG COO P. Magna Terra 573 editores. 574 34. Mingorria S., 2009. “Investigando” en el Valle del Polochic. Guatemala. Desafíos en la 575 aplicación de la metodología del Análisis Integrado Multiescalar del Metabolismo 576 Ecológico y Social (MuSIASEM). Master’s thesis dissertation in Environmental Sciences, 577 Institute of Environmental Science and Technologies, Autonomous University of 578 Barcelona. 579 35. Mingorría S., Gamboa G., Martín-López B., Corbera E., 2014. The oil palm boom: socio- 580 economic implications for Q’eqchi’ households in the Polochic valley, Guatemala. 16 581 Environment Development and Sustainability 16: 841–871. DOI 10.1007/s10668-014- 582 9530-0 583 36. Mingorría S., 2016. The nadies weaving resistance. Oil palm and sugar cane conflicts in 584 the territory, communities and households of the Q’eqchi’, Polochic valley, Guatemala. 585 PhD thesis, Autonomous University of Barcelona, Cerdanyola del Vallès, Spain. Available 586 at: https://ddd.uab.cat/record/176186. Accessed: 27/10/2018 587 37. Netting. R., 1995. Smallholders, Householders. Farm families and the ecology of 588 intensive, sustainable agriculture. Stanford University Press. Stanford, California 589 38. Rivera Ferre M.G., Ortega-Cerdà M. 2011. Recognising ignorance in decision-making. 590 Strategies for a more sustainable agriculture. EMBO Reports, 12(5), 393-397 591 39. Robinson J., 2008. Being undisciplined. Transgressions and intersections in academia 592 and beyond. Futures 40, 70-86. 593 40. Rojas M., 2008. Experienced Poverty and Income Poverty in Mexico: A Subjective Well- 594 Being Approach. World Dev. 36, 1078–1093 595 41. Ronzon T., Tillie P., 2004. Diagnóstico agrario de la región de Telemán-Sierra de las 596 Minas, Valle del Polochic. Guatemala: entre frontera agrícola y grandes latifundios. 597 Institut Nacional Agronomique Paris-Grignon/Agrónomos y Veterinarios sin Fronteras 598 42. Scheidel A., 2013. Flows, funds and the complexity of deprivation: Using concepts from 599 ecological economics for the study of poverty. Ecological Economics 86, 28-36. 600 doi:10.1016/j.ecolecon.2012.10.019 601 43. Scheidel A., 2016. Tactics of land capture through claims of poverty reduction in 602 Cambodia. Geoforum, 75, 110–114. https://doi.org/10.1016/j.geoforum.2016.06.022 603 44. Sen A., 1999. Development as Freedom. Oxford University Press, Oxford New York 604 45. Solano L., Solís F., 2010. La Palma Africana: Agronegocio que se expande. Territorio: 605 Posiciones Contrapuestas. El Enfoque, 6. Guatemala: El Observador 606 46. Sumner A., 2007. Meaning versus Measurement: Why Do 'Economic' Indicators of 607 Poverty Still Predominate? Development in Practice 17, 4-13 608 47. Wagner R., 2001. Historia del café de Guatemala. Guatemala: Villegas Asociados. 609 48. Wiggins S., Kirsten J., Llambí L., 2010. The future of small farms. World Development 610 38(10), 1341–1348 611 49. Yin R., 2003. Case study research. Design and methods. Sage, Thausand Oaks, CA. 17 Appendix (1) Mix-traditional (2) Cash crop-Land (3) Cash crop-No Land (4) Oil palm-Land (5) Oil palm-No Land Kruskal-Wallis test Mean Std.dev. Mean Std.dev. Mean Std.dev. Mean Std.dev. Mean Std.dev. 2 p-value Net income 3839.3 A 2315.2 6127.6 A 3874.6 8014.5 A.B 6050.5 18477.2 B.C 19283.4 16100.9 C 5985.3 45.41 < 0.001 Net income p.c 639.3 A 432.6 1011.9A.B 613.3 1953.5 B.C 1327.8 2581.7 B.C 3037.9 3716.9 C 1662.0 45.02 < 0.001 Returns to labor 3.91 A 2.77 4.51 A.B 3.54 14.81 A.B 29.77 6.75 B 2.19 7.93 B 4.44 18.58 0.001 Self-sufficiency 0.894 C 0.180 0.77 B.C 0.239 0.492 A.B 0.483 0.541 A.B 0.206 0.393 A 0.139 37.27 < 0.001 Consumption of 422.6 247.0 476.0 261.2 474.6 334.8 443.1 256.0 662.1 338.0 6.1 0.191 maize (eq.c) Productivity of 990.0 A.B 501.4 1793.0 C 810.4 1523.2B.C 552.7 631.8 A 528.5 602.9 A 330.9 35.4 < 0.001 maize in dry season Saturation of work 0.49 A 0.11 0.57 A.B 0.14 0.54 A 0.17 0.54 A 0.15 0.75 B 0.14 18.16 0.001 Land used for maize 1.20 A.B 0.68 1.07 A.B 0.37 0.85 A 0.58 1.57 B 0.46 0.64 A 0.37 25.109 <0.001 in dry season Table A1. Mean values and statistical differences of calculated indicators, based on Kruskal-Wallis test 18