World Development Vol. 64, pp. 448–459, 2014 0305-750X/Ó 2014 Elsevier Ltd. All rights reserved. www.elsevier.com/locate/worlddev
http://dx.doi.org/10.1016/j.worlddev.2014.06.022
Are Poor People Less Happy? Findings from Melanesia SIMON FEENY, LACHLAN MCDONALD and ALBERTO POSSO* RMIT University, Melbourne, Australia Summary. — Measures of happiness are increasingly being used to inform development policy. This is particularly true in Melanesia where linkages between income and life satisfaction can be weak due to the dominance of semi-subsistence lifestyles. This paper examines the happiness of households in two Melanesian countries: Solomon Islands and Vanuatu. The focus is on whether the poor are less happy. Findings indicate that wealth, increases in earnings, relative wealth, employment, and living on communally owned land are positively associated with happiness. Household size and food insecurity have a negative association. There is also strong support for poor households being less happy. Ó 2014 Elsevier Ltd. All rights reserved. Key words — happiness, poverty, Melanesia, Solomon Islands, Vanuatu
1. INTRODUCTION
destitution (Bayliss-Smith & Feachem, 1977; Lam, 1982). A similar view is that subsistence affluence prevails in these societies. The term “subsistence affluence” was originally coined by Fisk (1971) and relates to households being able to satisfy their needs using local environmental resources with very little labor input. 2 It is for these reasons that people have been reluctant to refer to poverty when evaluating human wellbeing in these countries. Yet lifestyles in these societies are changing and many households face a number of challenges akin to poor households across the world. Malnutrition and hunger exist (MICS., 2007), and households have been observed relying on cheap, sometimes poor quality imported food and even skipping meals during food price shocks (Feeny, McDonald, Posso, & Donahue, 2013). Monetization is increasing the importance of income to enable the payment of school fess and the purchase of basic household goods and services in order to meet the basic needs of the family (Abbott & Pollard, 2004; Regenvanu, 2009). Urbanization combined with very few employment opportunities has given rise to squatter settlements, high rates of unemployment, social tensions, and crime (Parks, Abbott, & Wilkinson, 2009). Moreover, climate change threatens to exacerbate the problems faced by these countries. Unfortunately, only a very limited number of surveys are conducted to provide insights into these issues. Assessing well-being in these contexts is therefore timely and important and can provide crucial information for policymakers and their social protection policies. The challenges facing Solomon Islands and Vanuatu are borne out by some of their official development indicators. While the World Bank classifies both as middle income countries, rates of income poverty and human development indicators reveal relatively low levels of development. Using basic
Measures and analyses of happiness have received great attention by policymakers in recent years. In 2011, the United Nations General Assembly adopted a resolution entitled “Happiness: towards a holistic approach to development” and the 2012 World Happiness Report advocates for selfreported well-being and happiness to take precedence over GDP in policymaking (Helliwell, Layard, & Sachs, 2012). 1 Throughout the Pacific, happiness gained much interest when Vanuatu topped the New Economics Foundations’ Happy Planet Index (HPI) in 2006 (NEF, 2006). Moreover, the Malvatumauri National Council of Chiefs (MNCC., 2012) focused on happiness in developing alternative indicators of well-being for Vanuatu. The report found that people living on customary land, that participate in traditional ceremonial activities and who are active members of their community are, on average, happier. This paper seeks to identify the main determinants of happiness in two Melanesian countries: Solomon Islands and Vanuatu. Communities in these countries are distinct, with the majority of households living semi-subsistence lifestyles on communally owned, customary land in rural areas. Most households have access to a garden on which to grow their food and systems of exchange, reciprocity, and social networks are known to be very strong (Regenvanu, 2009). A specific objective of the paper is to examine whether there is a relationship between poverty and happiness. While the link between income and happiness has been examined extensively, there are far fewer studies which have examined the relationship between poverty and happiness using household-level data. Household surveys rarely collect data on the multidimensional aspects of poverty in addition to data on happiness. Given the focus of the international community on poverty and the preference of policymakers in the region (and elsewhere) to find alternative indicators, it is important to determine whether the two are related. Defining and measuring poverty in Melanesian countries such as Solomon Islands and Vanuatu is a challenging task (Morris, 2011). One view is that poverty does not exist in these countries since communally owned land systems, subsistence or semi-subsistence lifestyles, and strong social support networks prevent extreme hunger, homelessness, and outright
* This paper is part of a project funded by the Australian Agency for International Development (AusAID) through its Australian Development Research Award (ADRA) scheme. These views expressed in the paper are those of the authors and not necessarily those of the Commonwealth of Australia. The Commonwealth of Australia accepts no responsibility for any loss, damage, or injury resulting from reliance on any of the information or views contained in the publication. Final revision accepted: June 20, 2014. 448
ARE POOR PEOPLE LESS HAPPY? FINDINGS FROM MELANESIA
needs poverty lines, in 2006, 23% of the population of Solomon Islands and 16% of the population of Vanuatu lived in poverty (AusAID, 2009). 3 According to the Human Development Index (HDI), Solomon Islands and Vanuatu rank 143 and 124 respectively, out of 186 countries (UNDP, 2013). Assessing progress toward the United Nations Millennium Development Goals (MDGs) is difficult due to a paucity of data but where data do exist they suggest that Solomon Islands is unlikely to achieve any of the goals while Vanuatu is on track to achieve just two: reducing child mortality and combatting HIV/AIDS and other diseases (PIFS, 2012). Improving child health remains a challenge in both countries, with 32% of children under the age of five in Solomon Islands and 26% of children in Vanuatu found to be stunted (AusAID., 2012). Given the distinctive lifestyles of Melanesian communities, defining poverty in the region is difficult. Questions arise as to what indicators should be used to measure poverty, how they should be weighted if they are employed in a composite poverty index and below what threshold should a household be deemed poor (Morris, 2011). There is, however, a consensus that conventional measures of poverty based on income or consumption are considered inappropriate (ADB, 2001; UNDP, 1999; Yari, 2004). Further, participatory poverty assessments conducted in these and other Pacific countries found that households preferred the term hardship over poverty and reported suffering from a lack of access to basic services, income earning opportunities, and good governance (Abbott & Pollard, 2004). In recognition that any measure of poverty in Solomon Islands and Vanuatu is likely to be contentious, the focus of this paper is on the well-known Multidimensional Poverty Index (MPI) devised by Alkire and Foster (2011). The index is arguably well-suited to Melanesian countries since it directly measures households’ deprivations in ten separate indicators across three non-monetary dimensions of well-being: health; education; and living standards. Eight of the ten indicators in the index are also directly based on the MDGs (Alkire & Santos, 2013). Additionally, the adaptability of the index allows for an investigation of the links between, not only, happiness and poverty, but also the links between happiness and vulnerability and more extreme forms of poverty. The remainder of this paper is structured as follows. Section 2 summarizes the existing literature which has sought to explain the determinants of happiness. Section 3 describes the data and methodology employed by the paper. Section 4 presents the findings from the empirical analysis and finally, Section 5 concludes.
449
demonstrated the importance of resource access, culture and community vitality, necessitating the inclusion of such factors in the investigation carried out by this paper. The remainder of this section summarizes the determinants of happiness that have been identified by other studies. It comprises three parts. The first part examines the relationship between income and happiness. The second summarizes non-income determinants of happiness while the third considers the limited number of studies which have specifically examined the relationship between poverty and happiness. (a) Income and happiness The focus of many happiness studies has been the impact of income. Such investigations are motivated by the seminal work of Easterlin (1974) which presented a so-called paradox. He found that: (i) richer individuals in the US are happier than poorer individuals but; (ii) over time, as the US got richer, average happiness failed to increase. A number of explanations have been purported to explain this paradox. The first is that individuals’ relative incomes are important for happiness, rather than actual income levels (Jiang, Lu, & Sato, 2012). 4 Individuals compare themselves to others and feel happier if their relative circumstances improve. According to this theory, if all individuals’ incomes increase together, then their relative standing does not change and happiness remains unaltered. Secondly, as argued by Layard (2006), people adapt to higher incomes quite quickly, making it hard to secure permanent increases in happiness from increases in income. While a rise in income might therefore have an impact on happiness initially, this impact dissipates over the longer term. Thirdly, the World Happiness Report argues that higher incomes can come at a cost, such as environmental degradation, an increase in insecurity, a loss of trust and reduced confidence in government. At a country level, it might also be the case that additional incomes only flow to the rich, doing little to increase the average level of happiness (Helliwell et al., 2012). Another finding is that there is diminishing marginal utility of income with respect to happiness. In other words, the life satisfaction of the poor can be improved with small amounts of money while richer individuals require a much larger increase in absolute income to improve their happiness. At low levels of income, additional financial resources can secure basic needs such as food, clothing, housing, health care, water, and sanitation. At higher incomes, such needs have already been met (Helliwell et al., 2012). (b) Non-income determinants of happiness
2. LITERATURE REVIEW There is now widespread recognition of the inadequacy of income as measure of well-being and a search for more appropriate measures (Doyal & Gough, 1991; Greeley, 1994; Sen, 1999; Stiglitz, Sen, & Fitoussi, 2009). The Alternative Indicators of Well-being report for Vanuatu emanated from the 2008 Melanesian Spearhead Group (MSG) Trade and Economic Officials Meeting (TEOM) and the MSG Leaders’ Summit. The leaders agreed that governments in the region need to do more to account for and measure the noncash values that contribute to their peoples’ quality of life (MNCC, 2012). Findings suggest that people from the remote, northernmost region of Torba Province are the happiest in Vanuatu and this is despite them having low incomes and being a long way from a major market. The report also
There is generally a consensus on the non-income determinants of happiness (Appleton & Song, 2008; Dolan, Peasgood, & White, 2008; MacKerron, 2012). In particular, after basic needs are met, aspirations, relative income differences, and the security of gains in well-being (or the belief that gains in well-being will persist) are all found to be important (Graham, 2005). 5 Helliwell et al. (2012) categorize the main determinants of happiness into two groups. The first group includes external factors such as income, work, community and governance, and values and religion. Country-level data reveal that happiness is highest in countries with a sense of community, trust, and social equity—Denmark, Finland, Norway, and the Netherlands. The second group includes individual-level factors such as mental health, physical health, family experience, education, gender, and age.
450
WORLD DEVELOPMENT
Unemployment and job insecurity are found to impact negatively on happiness (Di Tella, MacCulloch, & Oswald, 2001; Frey & Stutzer, 2002) while trust, being part of a community, freedom, equality, and religion all have a positive association (Helliwell et al., 2012). Better mental and physical health is also found to be positively associated with happiness although these variables are often difficult to measure (Helliwell et al., 2012). Marriage is nearly always found to increase happiness but there is no strong evidence that having children is associated with happiness (Stutzer & Frey, 2006). The level of education is usually expected to increase an individuals’ happiness although there is relatively little evidence of this effect independent of income (Helliwell et al., 2012). With regard to gender, women are found to be happier than men in advanced economies but the reverse is sometimes found to be the case in developing countries (Blanchflower, 2008; Graham & Felton, 2005; Senik, 2004). The relationship between happiness and age is often found to be U-shaped with happiness declining for middle-aged individuals before rising in later life (Blanchflower & Oswald, 2004; Helliwell, 2003). These findings from the literature inform the specification of the empirical model specified in the proceeding section. (c) Poverty and happiness While many studies have focused on the relationship between income and poverty, there are very few studies which have examined whether poverty (and non-income poverty in particular) is related to happiness. It is sometimes argued that the poor are happy possibly due to a focus on relationships and community vitality rather than on money and materialism with pictures of smiling people in remote areas of Africa providing testament to this claim (Barford, 2011; Bundervoet, 2013; Graham, 2010). Conversely, it can be argued that such assertions are naı¨ve to the aspirations of the poor and their need for access to food and basic services for their survival. Amartya Sen notes that the some people can bear adversity cheerfully (although this does not mean there is no adversity and that we should ignore the depravity experienced by the poor) (see Barford, 2011). The empirical evidence for a link between poverty and happiness is somewhat mixed. Graham and Pettinato (2002a) find that in the cases of Peru and Russia, it is not typically the poorest who report being unhappy, identifying a “happy peasant and frustrated achiever” problem. However, a number of other studies suggest that a negative relationship between poverty and happiness exists. For example, Lever et al. (2005) confirms a link between poverty and subjective well-being in Mexico with the link being mediated by a number of psychosocial factors. Kingdon and Knight (2006) explore the inter-relationships between subjective well-being and income poverty in South Africa, finding that income is positively associated with subjective well-being along with a number of other non-monetary factors. Banerjee and Duflo (2007) find that the poor in India experience greater financial and psychological stress. Similar findings are reported by Case and Daeton (2005) for India, South Africa, and the US and by Rojas (2011) for the poor in Latin America. Camfield et al. (2009) also confirm a relationship between poverty (wealth) and happiness in Bangladesh but that the relationships play a more important role. Finally, White (2013) examines happiness and other measures of subjective well-being with respect to food security in rural India, finding many intervening variables and emphasizing the need to complement quantitative with qualitative data. This paper contributes to this small but growing literature.
Bhutan’s well-known Gross National Happiness (GNH) measure incorporates large amounts of information on an individual’s well-being into a single index. In some ways, households in Bhutan are similar to those in the countries under consideration in the current paper, with close family ties, a strong sense of community, culture and tradition, and important links to the environment. In its latest version, the GNH index incorporates information on the following nine domains of well-being: psychological wellbeing; health; education; time use; cultural diversity and resilience; good governance; community vitality; ecological diversity and resilience; and living standards. These domains are, in turn, based on 33 indicators and 124 variables (Ura, Alkire, Tangmo, & Wangdi, 2012). However, “happiness” here relates to the creation of enabling conditions where people are able to pursue well-being in sustainable ways (UN., 2013). It is (deliberately) not a single measure of subjective wellbeing. 6 This paper uses a multidimensional poverty index to examine whether those that are deprived in a number of well-being dimensions in Melanesian countries, also consider themselves to be unhappy. A related issue is whether there is a relationship between vulnerability and happiness. Vulnerability refers to the likelihood of being poor or becoming poor in the future. Graham and Pettinato (2002b) find that the self-reported well-being of those who have escaped poverty can be lower than that of the poor due to an insecurity or risk of falling back into poverty. This paper contributes to the existing literature by examining these issues in greater detail. 3. DATA AND METHODOLOGY (a) Data This paper employs data from a unique household survey conducted in Solomon Islands and Vanuatu in 2012–13. A total of 619 households were surveyed (302 households in Solomon Islands and 317 households in Vanuatu). Ten communities were selected that would provide the research with a sufficiently rich data set on the different ways that households live. They were selected based on criteria that sought to reflect diversities in remoteness, economic activity, and environmental differences. It was also intended that the aggregate distribution of communities be not too dissimilar to Census data in terms of the distribution of urban and rural households. Four communities in the capital cities and six communities outside the capitals were therefore targeted. In Solomon Islands, the research was limited to the two largest and most populous islands: Guadalcanal, which is home to the capital city Honiara, and Malaita. In Honiara two communities were visited: White River and Burns Creek. The former is a multi-ethnic settlement located in the west of the city while the latter is situated in the east and consists mainly of Malaitans that were displaced during an ethnic conflict from 1999 to 2003. Two communities: Oa and Marauipa, were surveyed on the Weather Coast of Guadalcanal—a region renowned for its extreme geographical remoteness and its exposure to harsh climactic conditions. On the island of Malaita households were surveyed in the densely populated rural center of Malu’u, which is located about 80 kilometers north of Auki, the country’s second-largest urban area. In Vanuatu the research was also limited to the two most populous islands: Efate, home to the capital city Port Vila, and Espirito Santo. In Port Vila two communities were surveyed: Ohlen in the north of the city and Blacksands in the south west. Both communities are settlements for migrants
ARE POOR PEOPLE LESS HAPPY? FINDINGS FROM MELANESIA
from outer islands. In Santo, the two migrant communities of Pepsi and Sarakata were surveyed in Luganville, the country’s second-largest urban area. While only separated by the Sarakata River, these two communities had very different access to government services since, at the time of the survey, only Sarakata was under the auspices of the Luganville Municipal Council. Also surveyed in Santo was the rural village of Hog Harbour, located about 50 kilometers north of Luganville, which is well serviced by a tar-sealed road to Luganville, and located close to the well-renowned tourist destination of Champagne Beach. The survey was designed following a literature review of poverty and happiness and their drivers in developing countries, particularly in the Pacific Islands. In order to capture potential gender differences, the research teams aimed for a 45–55% gender balance in survey respondents. The survey gathers information on household demographics, physical household characteristics, productive assets, income and expenditure, social assets, water and sanitation, and respondents’ perceptions of their own well-being. To measure happiness, household members were asked (in their local language) “On a scale of 1–10 (with 10 being very happy and 1 being not happy at all) how happy are you?” 7 The happiness scores by community for each country are provided in Figure 1 below. (b) Methodology As outlined in Graham (2005), empirical models of happiness are usually specified in the same way. We follow this specification: W i ¼ a þ bxi þ ei ;
ð1Þ
where Wi is self-reported happiness of household member i, xi is a vector of explanatory variables including socio-demographic and socio-economic characteristics. a is a constant and b is a vector of coefficients. Unobserved characteristics and measurement errors are captured in the error term (e). As outlined above, the dependent variable, self-reported happiness, is an ordinal variable taking values of 1 to 10. Higher values correspond to a higher level of happiness.
The inclusion of the variables in vector xi is motivated by the literature review undertaken in Section 2 as well as by the characteristics, culture, and livelihoods of households in Solomon Islands and Vanuatu, discussed in the introduction. The variables include gender (taking the value of one is the respondent is male and zero otherwise), the age of the respondent (in years), employment (a dummy variable taking the value of one if the respondent works for money and zero otherwise), the number of people living in the household, income (defined by seven income brackets with a higher number indicating a higher income) and three binary variables indicating: whether the household lives on communally owned (customary) land; whether the household has access to a garden; and whether the household worried that their food would run out before they got money to purchase more (as a measure of food security). 8 Five remaining explanatory variables warrant further discussion: perceived relative wealth; wealth, multidimensional poverty, severe multidimensional poverty, and vulnerability. Following Carroll, Overland, and Weil (1997) and McBride (2001) perceived “relative wealth” variables are used to capture wealth relative to a reference group (external) and relative to past experience (internal). External perceived relative wealth was obtained by asking households about their relative financial situation: “Compared with the rest of your country do you feel poor or rich?” and respondents were provided with a five point Likert scale from one (for very poor) to five (for very rich). The same question was asked in relation to their village or community. 9 Internal perceived relative wealth was obtained by asking respondents whether they think that “the total amount of money that comes into the house has gone up or down in the past two years?” Similarly, respondents were provided with a five point Likert scale from one (down a lot) to five (up a lot), with the midpoint (three) representing no change. Wealth is measured using an index based on the approach pioneered in Filmer and Pritchett (2001). It uses principal components analysis (PCA) to construct the index using multiple indicators of durable assets and dwelling characteristics. 10 A total of 14 separate variables are used to construct a wealth index for households in Solomon Islands and Vanuatu (see
Vanuatu
Figure 1. Happiness scores by community in Solomon Islands and Vanuatu.
Pepsi
Sarakata
Black Sands
Hog Harbour
Ohlen
Oa
Burns Creek
Malu'u
Maruiapa
White River
0
0
2
2
4
4
6
6
8
8
10
10
Solomon Islands
Happiness, 1-10
451
452
WORLD DEVELOPMENT
Table 2 in the Appendix A). Since these variables are binary in nature and standard PCA is designed for continuous variables, the analysis uses the polychoric PCA methodology outlined in Kolenikov and Angeles (2009) to construct the wealth index. As per the approach in Filmer and Pritchett (2001) the mean of the resultant index is centered on zero implying that the index can take on negative as well as positive values. 11 Careful consideration was given to the measure of poverty used in the study. For the reasons discussed above, this paper replicates the MPI developed by Alkire and Foster (2011). The index has become a widely accepted and used measure of poverty and is currently reported in the annual United Nations Development Program’s Human Development Reports. A case is also being made for reductions in the MPI to be used as a headline goal in the new round of international development targets in the post-2015 development agenda (Alkire & Sumner, 2013). 12 The MPI considers three equally weighted dimensions of poverty (health, education, and living standards) captured by ten indicators. The MPI therefore conveys additional information not captured in single-dimensional measures. It identifies those who are poor through a two-step process involving identifying cut-offs of deprivation. “The first is the traditional dimension-specific cut-off, which identifies whether a person is deprived with respect to that dimension. The second delineates how widely deprived a person must be in order to be considered poor” (Alkire & Foster, 2011, p. 477). In this way, the MPI simultaneously concerns itself with how many people are experiencing poverty as well as how much (or depth of) the deprivation is being experienced. The MPI is calculated using the following formula: MPI ¼ H A;
ð2Þ
where H is the headcount or the percentage of people who are identified as multidimensionally poor and A (intensity) is the percentage of dimensions in which the average poor person is deprived. Following Alkire and Foster (2011), a household is deemed poor if it is deprived in at least 33% of the weighted indicators. From this discussion it is clear that a number of decisions or choices need to be made in the calculation of the MPI for Melanesia. They include the choice of well-being dimensions, the weighting system for those dimensions, and the threshold used to determine who is poor. 13 In the absence of a strong theoretical and conceptual framework to guide these choices for Melanesian countries, the paper largely follows the choices made and justified by Alkire and Foster (2011) which also allows for a cross-country comparison of the incidence of poverty. However, sensitivity analysis is conducted to test the robustness and validity of the choices made in this paper. Firstly, an alternative MPI that has been tailored for Melanesian countries is employed in the analysis (Clarke, Feeny, & McDonald, 2014). This involves augmenting the existing MPI with an additional dimension of well-being, relating to access. The access dimension, in turn, includes information from three equally weighted indicators: (i) access to a garden (which takes the value of one if a household does not have a garden and zero otherwise); (ii) access to services (which takes the value of one if it takes household members more than 30 minutes to travel to a health clinic, secondary school, or central market); and (iii) access to social support (which takes the value of one if a household indicates it has no one to rely upon in a time of financial difficulty). Ideally the choice of dimensions emanates from a participatory process (Alkire, 2008) and the
inclusion of the access dimension is justified upon previous Participatory Poverty Assessments (PPAs) noting that access to land, services, and social support are particularly important for Melanesian communities (Abbott & Pollard, 2004; ADB, 2001). 14 Secondly, alternative poverty thresholds are introduced. A looser threshold that classifies households as poor if they are deprived in 20% or more of indicators is employed in addition to the 33% threshold used by Alkire and Foster (2011). To explore the sensitivity of results further, additional classifications of Melanesian households are employed in the empirical model. Specifically, a category of vulnerable households is included (that is, households that are close to being poor). These households are deprived in 20–33% of all indicators. Severely poor households are also separately specified (defined as those households deprived in 50% or more of the indicators). Table 3 in the Appendix A provides the dimensions, indicators, weights and summary statistics for the MPI. Table 4 provides definitions for the other variables included in the model and summary statistics are provided in Table 5. In the absence of additional information on the relative importance of the well-being dimensions to Melanesian households, the dimensions of the index are equally weighted. While this could be viewed as a shortcoming, Alkire and Santos (2013) demonstrate that MPI results are robust to different indicator weights (and threshold levels). Given the ordinal nature of the dependent variable, OLS would yield biased and inefficient parameter estimates and an ordered probit model is therefore estimated. 15 All models include community fixed effects to account for unobserved characteristics that vary between communities and could potentially affect individual levels of happiness, such as the proximity to urban centers or the coast. In the econometric models presented in the following section we estimate Eqn. (1) using a number of independent variables interchangeably in order to avoid the problem of multicollinearity. Multidimensional poverty, food security, income, and wealth are found to be particularly highly correlated. Columns (1) to (10) present the results from estimating Eqn. (1) using alternative multidimensional poverty indicators. These models do not include the income or wealth variables and food security is also excluded in even-numbered columns to examine the sensitivity of results to the inclusion of this variable. Columns (11) and (12) present results from the model including income and wealth respectively but excluding measures of multidimensional poverty. Finally, it has been suggested that wealth is endogenous in models explaining happiness (see Stutzer, 2004). The same might also apply to the poverty variables. We tested for the endogeneity of these variables using a Durbin–Wu–Hausman test, using education and the number of children as instruments, noting that they are not found to be significant determinants of happiness. These tests indicated that all of the variables are exogenous at the 1% level of statistical significance. 4. RESULTS AND INTERPRETATION This section presents the results from the empirical model explaining the variation in happiness across the communities in Solomon Islands and Vanuatu. Many findings from this model are consistent with a priori expectations. Turning first to the poverty indicators, Columns (1) and (2) depict a negative and statistically significant coefficient attached to the MPI variable, suggesting that multidimensionally poor house-
ARE POOR PEOPLE LESS HAPPY? FINDINGS FROM MELANESIA
holds are less happy. 16 Results from Column (3) show a negative, albeit statistically insignificant, relationship between happiness and households that are classified as poor according to the Melanesian MPI. However, Column (4) indicates that once food insecurity is omitted from the model, the negative relationship between happiness and this measure of multidimensional poverty becomes statistically significant. The variable capturing access to a garden is excluded from these regressions since it is included in the Melanesian MPI. 17 Very similar findings are presented in Columns (5) and (6) where the looser (20%) threshold for MPI poverty is used. Results from Columns (7) and (8) indicate a negative and statistically significant association between extreme multidimensional poverty and happiness regardless of whether or not food insecurity is included. However, Columns (9) and (10) indicate that there is no statistically significant relationship between vulnerability and happiness. These results suggest that the negative relationship between multidimensional poverty and happiness is limited to those households already
453
living in poverty. In contrast, there is no discernible relationship between poverty and happiness for near-poor households facing the prospect of being tipped into poverty in the future. The results also suggest that food insecure households are less happy, with the coefficient attached to this variable being consistently negative and statistically significant when it is included in the model. The relationship between poverty, food insecurity, and happiness supports Maslow’s (1954) “hierarchy of needs” hypothesis, where meeting individuals’ most basic needs, such as obtaining nourishment, is fundamental to welfare. However, as in Tay and Diener (2011), our results also show that although meeting basic needs is an important determinant of wellbeing, other factors, such as access to education and ownership of assets, which make up the MPI, also matter. This, in turn, suggests that there may not be a “hierarchy” of needs per se, but a number of different factors that combine to determine happiness.
Table 1. The determinants of happiness in Solomon Islands and Vanuatu
Male Age Age squared Poverty (MPI)
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
0.15 [1.57] 0.022 [1.22] 0.00027 [1.37] 0.20** [1.96]
0.13 [1.30] 0.018 [1.07] 0.00022 [1.19] 0.26** [2.45]
0.14 [1.47] 0.022 [1.25] 0.00027 [1.38]
0.11 [1.17] 0.019 [1.10] 0.00022 [1.20]
0.16 [1.61] 0.022 [1.23] 0.00026 [1.36]
0.13 [1.37] 0.018 [1.08] 0.00022 [1.18]
0.14 [1.45] 0.021 [1.18] 0.00025 [1.33]
0.11 [1.16] 0.017 [1.00] 0.00020 [1.11]
0.15 [1.52] 0.020 [1.14] 0.00024 [1.26]
0.12 [1.24] 0.016 [0.96] 0.00019 [1.02]
0.098 [0.93] 0.012 [0.58] 0.00017 [0.78]
0.12 [1.28] 0.023 [1.30] 0.00028 [1.48]
0.17 [1.63]
0.21** [2.07] 0.14 [1.38]
0.19* [1.90] 0.25* [1.70]
0.29** [2.04] 0.053 [0.53]
0.051 [0.51]
Melanesian MPI Poverty (20% threshold) Extreme poverty Vulnerability Income
0.11 [1.63]
0.19*** [3.25] 0.53***
0.14*** [3.15] 0.14*** [2.75] 0.43***
[6.76] 0.41 [1.62] 0.026
[5.96] 0.38 [1.60] 0.041***
[1.29] 0.17 [1.21] 0.24**
[2.70] 0.093 [0.73] 0.26**
Wealth Internal relative wealth External relative wealth (Country) Employed Total people in household Garden Communally owned land Food insecure Community FE? Observations
0.14*** [2.80] 0.46***
0.15*** [2.95] 0.50***
0.14*** [2.86] 0.47***
0.15*** [3.02] 0.51***
0.14*** [2.83] 0.46***
0.15*** [2.98] 0.50***
0.14*** [2.87] 0.47***
0.15*** [3.04] 0.51***
0.14*** [2.88] 0.48***
0.15*** [3.05] 0.53***
[6.42] [7.05] [6.62] [7.31] [6.45] [7.06] [6.50] [7.18] [6.85] [7.61] 0.40* 0.51** 0.41* 0.52** 0.38 0.48** 0.41* 0.53** 0.41* 0.52** [1.70] [2.14] [1.73] [2.18] [1.62] [2.03] [1.77] [2.24] [1.77] [2.24] 0.031** 0.033** 0.034** 0.036** 0.032** 0.033** 0.032** 0.033** 0.033** 0.035** [2.08] 0.085 [0.68] 0.24**
[2.13] 0.12 [0.97] 0.28***
[2.26]
[2.36]
0.23**
0.27***
[2.27] 0.43*** [4.45] Yes 619
[2.66]
[2.21] 0.44*** [4.57] Yes 619
[2.58]
Yes 619
Yes 619
[2.13] 0.086 [0.68] 0.25**
[2.19] 0.12 [0.98] 0.30***
[2.10] 0.075 [0.60] 0.25**
[2.18] 0.11 [0.88] 0.29***
[2.22] 0.085 [0.68] 0.25**
[2.31] 0.12 [0.99] 0.29***
[2.38] 0.44*** [4.52] Yes 619
[2.81]
[2.35] 0.45*** [4.56] Yes 619
[2.78]
[2.33] 0.46*** [4.68] Yes 619
[2.77]
Notes: Dependent variable is happiness. Robust z-statistics in brackets. * Denotes statistical significance at the 10%. ** Denotes statistical significance at the 5%. *** Denotes statistical significance at the 1%.
Yes 619
Yes 619
Yes 619
[1.99] [2.49] 0.46*** 0.42*** [4.17] [4.33] Yes Yes 500 619
454
WORLD DEVELOPMENT
Results from Column (11) reveal a positive but statistically insignificant relationship between income and happiness for households in Solomon Islands and Vanuatu. This confirms, as suggested in the introduction, that income is not a good predictor of well-being in Melanesia, at least when measured using happiness. Results from Column (12) indicate that there is a positive association between household wealth and happiness consistent with Camfield et al. (2009). Turning to the other explanatory variables, results from across Table 1 also suggest that variation in happiness in the two Melanesian countries is partially explained by relative wealth. Specifically, internal relative wealth (an increase in earnings over the last two years) has a positive and statistically significant association with happiness. These results are consistent with the intuition in Carroll et al. (1997), where individuals gain utility by ever-increasing their current consumption relative to their past levels of consumption. Further, results indicate that external relative wealth (if individuals believe themselves to be wealthier relative to others in their country), is also positively associated with happiness. External relative wealth and its effect on happiness, therefore, seems to suggest that status, or how individuals feel to be perceived by others, are playing an important role in determining happiness in Melanesia. These findings are clearly consistent with the extant literature (Diener, 1984). However, there may also be Melanesian-specific factors that are driving these results, specifically wantokism and bigmen (Nanau, 2011). Wantok refers to the networks of various, tribal, ethnic, linguistic, and geographic groupings in Melanesia (Nanau, 2011). The wantok system is an identity concept that defines a set of obligations, mainly cooperation and reciprocal support, between individuals of the same community (de Renzio, 2000). As such the wantok generates a traditional custom (kastom in the local languages of Pidgin and Bislama) economy with redistribution and traditional social protection as central pillars. This system manifests itself in various forms, such as cash contributions to ceremonies and festivities, as well as sharing food and clothing at community fundraising events (McDonald, Naidu, & Mohanty, 2013). Higher perceived external wealth may allow households to meet their customary obligations more easily, thereby raising their happiness. A bigman culture prevalent in these societies also underlies the importance of status. A “big man” is a highly influential member of a localized community in Melanesia and often equated with the chief (Sahlins, 1963). Similarly, Table 1 reveals that individuals living on communally owned land are happier. Individuals with access to this land may feel happier because of strong cultural connections to land in these countries and a sense of belonging stemming from the fact that these lands have been passed from generation to generation through a variety of traditional tenure systems (MNCC, 2012). Results from Table 1 also indicate that respondents who are employed are generally happier. This is a common finding in this literature, suggesting that employment provides psychological benefits associated with social contact, being part of a collective purpose, engagement in meaningful activities, and social status (Jahoda, 1982). Table 1 also shows that larger household size is associated with lower happiness in these communities. To put this into perspective, an increase in the number of people living in a household by one lowers the probability that an individual will attain a
score of 10 in the happiness scale by 1%. This finding is consistent with anecdotal evidence from Melanesia that extended family members are moving to areas with better employment opportunities and placing stress on the households which host them (Maebuta & Maebuta, 2009). Increasing numbers of mouths to feed with limited employment opportunities and sometimes a shortage of land for gardens could be reducing the happiness of household members. Finally, results from the analysis indicate that gender, age, and having access to a garden are not important in explaining the variation in happiness in the countries. 18 5. CONCLUSION, POLICY IMPLICATIONS, AND LIMITATIONS Using statistical modeling of unique household survey data, this paper identifies the main determinants of happiness in both Solomon Islands and Vanuatu. An important finding is that there is little evidence of income explaining the variation in happiness in these two countries. The quest for alternative indicators of well-being in the region is therefore well justified. So too is a focus on reducing multidimensional poverty. Findings indicate that the poor in these countries are, on average, less happy. This has important implications. Improving the health, education, and living standards indicators of the MPI will improve happiness in the region, as well as being worthwhile in their own right. Results from a tailored MPI for Melanesia also suggests that improving access to markets, basic services, and strengthening communities will improve happiness. While improving access to markets and basic services are already a focus of policy makers, strengthening communities is very challenging, particularly in urban areas—and particularly so in the migrant settlements, which formed a major part of this research. Governments and donors should focus on capacity building and community empowerment schemes and involving community networks in the delivery of services. Regenvanu (2009) contends that continued participation in custom ceremonies can strengthen communities. He also argues that disputes be resolved within communities by traditional leaders using traditional dispute mechanisms and that customary land ownership be maintained. Other findings reveal that wealth, increases in income, and relative income are all positively associated with happiness. In contrast, household size has a negative association. These findings are generally consistent with many previous studies explaining the variation in happiness (usually across country). Two other findings from the analysis are unique to this study, with further implications for policymakers in Melanesia. The first is that people on communally owned land are found to be happier. This is likely to reflect the close ties that communities in these countries have to their land as well as the importance of land for gardens. Secondly, food security is also found to be important for happiness. Policymakers should therefore find ways to strengthen access to land and improve food security. In urban areas this might involve support for urban gardens and land segregation schemes, as well as programs that encourage food cultivation in the available space around homes, such as potted and hanging gardens. Improving agricultural productivity through education, access to finance, and the provision of tools will also improve food security.
ARE POOR PEOPLE LESS HAPPY? FINDINGS FROM MELANESIA
Finally, it is important to note the limitations of the study. Firstly, there is potential selection bias since participants were not selected at random. While the distribution of communities in the sample is not dissimilar to Census data in terms of the distribution of capital and non-capital city households, there is a question as to whether the findings from this paper can be generalized for all of Solomon Islands and Vanuatu. Secondly, while a case is made for the choice of happiness as a measure of well-being, this choice can always be contested (White, Gaines, & Jha, 2012). Further research is required to examine how sensitive findings are to the measure of
455
well-being employed. Thirdly, the role of cultural factors in Melanesia such as wantokism in explaining happiness also warrants further attention. More in-depth qualitative analysis into how cultural norms and institutions affect the happiness of both men and women in Melanesian communities is required. Finally, with regard to the measure of multidimensional poverty, relatively arbitrary choices have to be made regarding the inclusion of different well-being dimensions and indicators and their weights as well as the threshold below which households are deemed poor. Stronger theoretical and conceptual frameworks to guide these choices are required.
NOTES 1. The country of Bhutan is often viewed playing a lead role on this issue by devising its measure of Gross National Happiness. This is discussed further in Section 2. 2. While Fisk (1971) was referring to livelihoods in Papua New Guinea, the term has also been applied to other Melanesian countries (see PIFS., 2012). 3. Basic needs poverty lines measure the level of income required to meet certain thresholds for housing, food, clothing, healthcare, education, and to meet customary obligations (AusAID, 2009). If the household earns less currency than the amount deemed necessary to meet these needs, the household is classified as poor. Urban poverty is found to be far higher than rural poverty in these countries (AusAID, 2009). However, Gibson (2010) and Narsey (2011), Narsey (2012) argue that this is likely to be the result of errors in calculating comparable poverty lines between urban and rural areas. 4. Relative standing or achievement relative to peers and parents can also be important (Bookwalter & Dalenberg, 2010). 5. Helliwell (2003) finds that six variables can explain 80% of the variance in happiness across 50 countries: the divorce rate, the unemployment rate, the percentage of people reporting that “most people can be trusted;” membership in non-religious organizations; the percentage of citizens who “believe in God;” and the quality of government. 6. The satisfaction indicator combines individuals’ subjective assessments of their contentment levels with respect to health, occupation, family, standard of living, and work–life balance (Ura et al., 2012). 7. This measure is line with previous work in the region (MNCC, 2012) which, in turn, uses the Self-Anchoring Striving Scale of Cantril (1965). Discussions with local researchers and enumerators revealed that people would interpret the question as happiness in general or happiness with life as a whole rather than happiness today. 8. Following McBride (2001), the mid-point of each income bracket is taken and divided by the number of household members. The variable is then logged. Households with income in the final bracket are excluded since there is no mid-point. We thank an anonymous referee for this suggestion. The finding relating to income is consistent when income is kept as an ordinal variable and entered in the regression specification and the full sample is employed. 9. The regression output presented in Table 1 only reports external relative wealth relative to country. The coefficient estimates of external relative wealth relative to community are similar to those relative to country. These results are available upon request.
10. Filmer and Pritchett (2001) use principal components analysis (PCA) to construct a single, latent, variable (wealth index) from 21 different household assets in rural India. They find that their index is robust to the assets included, produces internally consistent results, and predicts school enrollment in India better than consumption. They argue that such a wealth variable may therefore be preferable to consumption or income as a proxy for long-run household wealth, since, unlike wealth, both durable assets and dwelling characteristics can be observed with precision. A substantial number of authors have since successfully adopted the approach to explain, inter alia, health outcomes, extreme poverty and inequality, as well as controlling for economic status in program evaluation when expenditures data are not available (Filmer & Scott, 2012). 11. As indicated in Table 5, mean wealth is not precisely zero because 41 households from the original survey sample of 660 (upon which the variable was calculated) were dropped for the analysis due to a lack of data. 12. It should also be recognized that the MPI has its critics. Ravallion (2011) criticizes multidimensional poverty indices for their arbitrary choice of dimensions and weights as well as a need to address issues relating to the aggregation of variables and trade-offs among them. Silber (2011) also makes a case for analyzing each dimension of the MPI separately. 13. See, for example, Mitra, Posarac, and Vick (2013) and Mitra et al. (2013) for alternative choices with respect to dimensions and weights. 14. The access dimension of well-being receives a weight of 25% together with the other standard dimensions of the MPI. See Clarke et al. (2014) for further details. 15. Ferrer-i-Carbonell and Frijters (2004) find that different empirical methods to modeling subjective well-being are unimportant to results. The findings of this paper are broadly consistent if OLS is used instead of an ordered probit model. 16. Following the suggestions of an anonymous referee, we examine the bivariate relationship between the MPI and the happiness. The pairwise correlation coefficient estimate is statistically significant with a value of 0.35 which is of similar magnitude to those found in Table 1. Furthermore, a simple Pearson’s chi-squared rejects the null hypothesis that the rows and columns in a two-way contingence table are independent confirming a statistical relationship between poverty and happiness. 17. To examine whether multicollinearity was affecting the results, Variance Inflation Factors (VIFs) were calculated for the explanatory variables. The VIFs indicated multicollinearity between the age and agesquared variables but not between any of the other variables. When the models were re-estimated excluding the age-squared variable, other
456
WORLD DEVELOPMENT
parameter estimates and their statistical significance changed very little. It is therefore possible to conclude that the findings from the paper are not the result of multicollinearity.
18. The latter finding is likely to be explained by a lack of variation in the garden access variable. Over 90% of households reported having access to the garden in the sample.
REFERENCES Abbott, D., & Pollard, S. (2004). Hardship and poverty in the Pacific: Strengthening poverty analysis and strategies in the Pacific. Manilla: Asian Development Bank. ADB (2001). Poverty: Is it an issue in the Pacific?. Manilla: Asian Development Bank. Alkire, S., & Santos, M. E. (2013). Measuring acute poverty in the developing world: Robustness and scope of the multidimensional poverty index. OPHI working paper No. 59. Oxford Poverty and Human Development Initiative. University of Oxford. Alkire, S., & Sumner, A. (2013). Multidimensional poverty and the post2015 MDGs. Oxford Poverty and Human Development Initiative Research Brief, University of Oxford, UK. Alkire, S., & Foster, J. (2011). Counting and multidimensional poverty measurement. Journal of Public Economics, 95, 476–487. Alkire, S. (2008). Choosing dimensions: The capability approach and multidimensional poverty. In N. Kakwani, & J. Silber (Eds.), The many dimensions of poverty (pp. 89–119). New York: Palgrave Macmillan. Appleton, S., & Song, L. (2008). Life satisfaction in Urban China: Components and determinants. World Development, 36(11), 2325–2340. AusAID (2009). Tracking development and governance in the Pacific. Canberra: Australian Agency for International Development. AusAID. (2012). Achieving education and health outcomes in Pacific Island Countries – Is there a role for social transfers? AusAID Pacific Social Protection Series: Poverty, vulnerability and social protection in the Pacific. Canberra: Australian Agency for International Development. Banerjee, A., & Duflo, E. (2007). The economic lives of the poor. Journal of Economic Perspectives, 21(1), 141–162. Barford, A. (2011). The myth of the happy poor. Retrieved from . Bayliss-Smith, T. P., & Feachem, R. G. (Eds.) (1977). Subsistence and survival: Rural ecology in the Pacific. London: Academic Press. Blanchflower, D. G. (2008). International evidence on well-being. Working paper No. 14318. Retrieved from National Bureau for Economic Research website: . Blanchflower, D., & Oswald, A. (2004). Well-being over time in Britain and the USA. Journal of Public Economics, 88, 1359–1387. Bookwalter, J. T., & Dalenberg, D. R. (2010). Relative to what or whom?: The importance of norms and relative standing to well-being in South Africa. World Development, 38(3), 345–355. Bundervoet, T. (2013, March 28). Poor but happy? Africa can end poverty [World Bank Blog]. Retrieved from . Camfield, L., Choudhury, K., & Devine, J. (2009). Well-being, happiness and why relationships matter: Evidence from Bangladesh. Journal of Happiness Studies, 10, 71–91. Cantril, H. (1965). The pattern of human concerns. New Brunswick, NJ: Rutgers University Press. Carroll, C. D., Overland, J., & Weil, D. N. (1997). Comparison utility in a growth model. Journal of Economic Growth, 2(4), 339–367. Case, A., & Daeton, A. (2005). Health and wealth among the poor: India and South Africa compared. American Economic Review Papers and Proceedings, 95(2), 223–229. Clarke, M., Feeny, S., & McDonald, L. (2014). Vulnerability to what?: Multidimensional poverty in Melanesia. In S. Feeny (Ed.), Household vulnerability and resilience to economic shocks: Evidence from Melanesia (pp. 83–106). UK: Ashgate. de Renzio, P. (2000). Bigmen and Wantoks: Social Capital and Group Behaviour in Papua New Guinea. Queen Elizabeth House working paper series, No. 27. Oxford: University of Oxford.
Di Tella, R., MacCulloch, R. J., & Oswald, A. J. (2001). Preferences over inflation and unemployment: Evidence from surveys of happiness. American Economic Review, 91, 335–341. Diener, E. (1984). Subjective well-being. Psychological Bulletin, 95(2), 542–575. Dolan, P., Peasgood, T., & White, M. (2008). Do we really know what makes us happy?: A review of the economic literature on the factors associated with subjective well-being. Journal of Economic Psychology, 29, 94–122. Doyal, L., & Gough, I. (1991). A theory of human need. London: Macmillan. Easterlin, R. A. (1974). Does economic growth improve the human lot? Some empirical evidence. In P. A. David, & M. W. Reder (Eds.), Nations and households in economic growth: Essays in Honour of Moses Abramovitz (pp. 89–125). New York: Academic Press. Feeny, S., McDonald, L., Posso, A., & Donahue, J. (2013). Household vulnerability and resilience to shocks: Findings from Solomon Islands and Vanuatu. SSGM discussion paper 2013/2, State Society and Governance in Melanesia Program. Canberra: Australian National University. Ferrer-i-Carbonell, A., & Frijters, P. (2004). How important is methodology for the estimates of the determinants of happiness?. The Economic Journal, 114, 641–659. Filmer, D., & Pritchett, L. (2001). Estimating wealth effects without expenditure data – Or tears: An application to educational enrolments in states of India. Demography, 38(1), 115–132. Filmer, D., & Scott, K. (2012). Assessing asset indices. Demography, 49(1), 359–392. Fisk, E. (1971). Labour absorption capacity of subsistence agriculture. Economic Record, 47(119), 366–378. Frey, B., & Stutzer, A. (2002). What can economists learn from happiness research?. Journal of Economic Literature, 40, 401–435. Gibson, J. (2010). Recent shocks and long-term change in the Samoan economy. Pacific Economic Bulletin, 25, 7–23. Graham, C. (2005). The economics of happiness: Insights on globalisation from a novel approach. World Economics, 6(3), 41–55. Graham, C. (2010). Happiness around the world: The paradox of happy peasants and miserable millionaires. Oxford: Oxford University Press. Graham, C., & Felton, A. (2005). Does inequality matter to individual welfare: An exploration based on happiness surveys in Latin America. Center on Social and Economic Dynamics working papers series No. 38. Washington, DC: The Brookings Institution. Graham, C., & Pettinato, S. (2002a). Frustrated achievers: Winners, losers and subjective well-being in new market economies. Journal of Development Studies, 38(4), 100–140. Graham, C., & Pettinato, S. (2002b). Happiness and hardship: Opportunity and insecurity in new market economies. Washington, DC: The Brookings Institution. Greeley, M. (1994). Measurement of poverty and poverty of measurement. IDS Bulletin, 25(2), 50–57. Helliwell, J. (2003). How’s Life? Combining individual and national variables to explain subjective well-being. Economic Modelling, 20, 331–360. Helliwell, J., Layard, R., & Sachs, J. (2012). World happiness report. New York: The Earth Institute, Columbia University. Jahoda, M. (1982). Employment and unemployment: A social–psychological analysis. London: Cambridge University Press. Jiang, S., Lu, M., & Sato, H. (2012). Identity, inequality and happiness: Evidence from Urban China. World Development, 40(6), 1190–1200. Kingdon, G. G., & Knight, J. (2006). Subjective well-being poverty vs. income poverty and capabilities poverty?. Journal of Development Studies, 42(7), 1199–1224.
ARE POOR PEOPLE LESS HAPPY? FINDINGS FROM MELANESIA Kolenikov, S., & Angeles, G. (2009). Socioeconomic status measurement with discrete proxy variables: Is principal component analysis a reliable answer?. Review of Income and Wealth, 55, 128–165. Lam, N. V. (1982). A note on the nature and extent of subsistence surplus in Papua New Guinea. Pacific Viewpoint, 23(2), 173–185. Layard, R. (2006). Happiness and public policy: A challenge to the profession. Economic Journal, 116, C24–C33. Lever, J. P., Pin˜ol, N. L., & Uralde, J. H. (2005). Poverty, psychological resources and subjective well-being. Social Indicators Research, 73(3), 75–408. MacKerron, G. (2012). Happiness economics from 35,000 feet. Journal of Economic Surveys, 26(4), 705–735. Maebuta, H., & Maebuta, J. (2009). Generating livelihoods: A study of urban squatter settlements in Solomon Islands. Pacific Economic Bulletin, 24(3), 118–131. Maslow, A. H. (1954). Motivation and personality. New York: Hoper & Row. McBride, M. (2001). Relative-income effects on subjective well-being in the cross-section. Journal of Economic Behavior and Organization, 45(3), 251–278. McDonald, L., Naidu, V., & Mohanty, M. (2013). Vulnerability, resilience and dynamism of the custom economy in Melanesia. In S. Feeny (Ed.), Household vulnerability and resilience to economic shocks: Evidence from Melanesia (pp. 107–128). UK: Ashgate. MICS. (2007). Multiple indicator cluster survey: Vanuatu. The United Nations Children’s Fund (UNICEF). Retrieved from . Mitra, S., Jones, K., Vick, B., Brown, D., McGinn, E., & Alexander, M. (2013). Implementing a multidimensional poverty measure using mixed methods and a participatory framework. Social Indicators Research, 110, 1061–1081. Mitra, S., Posarac, A., & Vick, B. (2013). Disability and poverty in developing countries: A multidimensional study. World Development, 41, 1–18. MNCC. (2012). Alternative indicators of well-being for Melanesia. Vanuatu Pilot Study Report. Port Vila, Vanuatu: Malvatumauri National Council of Chiefs. Morris, M. (2011). Measuring poverty in the Pacific, Development Policy Centre. Discussion Paper No. 9, Development Policy Centre. Canberra: Australian National University. Nanau, G. L. (2011). The Wantok System as a socio-economic and political network in Melanesia. OMNES: The Journal of Multicultural Society, 2(1), 31–55. Narsey, W. (2011). The incidence of poverty in Solomon Islands: The importance of methodology. Journal of Pacific Studies, 31(1), 31–58. Narsey, W. (2012). Poverty analysis in Vanuatu: A critical review and alternative formulation. South Pacific Studies, 33(1), 25–51. NEF (2006). The happy planet index: An index of human well-being and environmental impact. London: New Economics Foundation. Parks, W., Abbott, D., & Wilkinson, A. (2009). Protecting Pacific Island children and women during economic and food crises: A working document for advocacy, debate and guidance. Suva: UNICEF Pacific, UNDP Pacific Centre, UNESCAP Pacific Operations Centre. PIFS. (2012). Pacific regional MDGs tracking report 2012. Suva: Pacific Islands Forum Secretariat. Ravallion, M. (2011). On multidimensional indices of poverty. Journal of Economic Inequality, 9, 235–248. Regenvanu, R. (2009). The traditional economy as the source of resilience in Melanesia. Port Vila: Vanuatu Cultural Centre. Rojas, M. (2011). Poverty and psychological distress in Latin America. Journal of Economic Psychology, 32(2), 206–217. Sahlins, M. D. (1963). Poor man, rich man, big-man, chief: Political types in Melanesia and Polynesia. Comparative Studies in Society and History, 5(3), 285–303. Sen, A. (1999). Development as freedom. Delhi: Oxford University Press. Senik, C. (2004). When information dominates comparison. Learning from Russian subjective panel data. Journal of Public Economics, 88, 2099–2133. Silber, J. (2011). A comment on the MPI index. Journal of Economic Inequality, 9, 479–481. Stiglitz, J., Sen, A., & Fitoussi, J. P. (2009). Report by the Commission on the Measurement of Economic Performance and Social Progress.
457
Retrieved from . Stutzer, A. (2004). The role of income aspirations in individual happiness. Journal of Economic Behavior and Organization, 54(1), 89–109. Stutzer, A., & Frey, B. S. (2006). Does marriage make people happy, or do happy people get married?. Journal of Socio-Economics, 35(2), 326–347. Tay, L., & Diener, E. (2011). Needs and subjective well-being around the world. Journal of Personality and Social Psychology, 101(2), 354–365. UN. (2013). The Bhutanese voice – The future we want for all: Well-being and happiness. Retrieved from www.worldwewant2015.org/file/ 370993/download/404371. UNDP (1999). Pacific human development report 1999: Creating opportunities. Suva: United Nations Development Program. UNDP (2013). Human development report: The rise of the South. New York: United Nations Development Program. Ura, K., Alkire, S., Tangmo, T., & Wangdi, K. (2012). An extensive analysis of GNH index. Thimphu, Bhutan: Centre for Bhutan Studies. White, S. C. (2013). Beyond the grumpy rich man and the happy peasant: Subjective perspectives on wellbeing and food security in Rural India. BPIDW working paper No. 25, Bath Papers in International Development. UK: University of Bath. White, S. C., Gaines, S. O., & Jha, S. (2012). Beyond subjective well-being: A critical review of the Stiglitz report approach to subjective perspectives on quality of life. Journal of International Development, 24(6), 763–776. Yari, M. (2004). Beyond subsistence affluence: Poverty in Pacific Island Countries. Bulletin on Asia-Pacific perspectives. Bangkok: United Nations Economic and Social Commission for Asian and the Pacific.
APPENDIX A
Table 2. Variables included in the wealth index Variable Clock Bicycle Radio TV Computer Sewing machine Mobile phone Unimproved drinking water
Flush Toilet Electricity High-quality building materials High-quality roofing materials
Cook with biomass Inside kitchen
Takes the value of 1 if Household owns a clock Household owns a bicycle Household owns a radio Household owns a television Household owns a computer Household owns a sewing machine Household owns a mobile phone Household mainly sources its drinking water from an unimproved water source, including: unprotected well or spring or purchases water from a vendor Household has flush toilet Household has access to electricity Dwelling is predominantly made from concrete wood or tin Dwelling has a roof that is predominantly made from tin or tiles Household cooks with: firewood; charcoal; coconut husk or dung Household has a kitchen indoors
458
WORLD DEVELOPMENT Table 3. Dimensions, indicators, deprivation thresholds and weights for the multidimensional poverty index
Dimension (weight)
Indicator (weight)
Deprived if
Mean
sd
Min
Max
Health (1/3)
Mortality (1/6) Nutrition (1/6)
Any child has died in the family Any adult or child for whom there is nutritional information is malnourished*
0.113 0.115
0.317 0.319
0 0
1 1
Education (1/3)
Years of Schooling (1/6)
No household member has completed five years of schooling Any school-aged child is not attending school in years one to eight
0.115
0.319
0
1
0.079
0.270
0
1
0.397 0.373
0.490 0.484
0 0
1 1
0.693
0.462
0
1
0.512 0.876
0.500 0.330
0 0
1 1
0.338
0.473
0
1
School Attendance (1/6) Standard of Living (1/3)
Electricity (1/18) Sanitation (1/18)
Water (1/18)
Floor (1/18) Cooking Fuel (1/18) Assets (1/18)
The household has no electricity The household´s sanitation facility is not improved (according to the MDG guidelines), or it is improved but shared with other households The household does not have access to clean drinking water (according to the MDG guidelines) or clean water is more than 30 minutes walking from home The household has dirt, sand, or dung floor The household cooks with dung, wood, or charcoal The household does not own more than one of: radio, TV, telephone, bike, motorbike or refrigerator, and does not own a car or truck
Notes: A proxy measure was used for this indicator. A households is deprived if they answered in the affirmative to the question “Did you or any other adults in the house not eat food for an entire day because there wasn’t enough money to buy food” taken from the US Food Security module.
Table 4. Variable definitions Variable
Definition
Happiness
On a scale of 1–10 (with 10 being very happy and 1 being not happy at all) how happy are you Dummy variable taking the value of 1 for male respondents Age of respondent in years Household earnings (per adult equivalent) measured using a Likert scale (1–7) using seven income brackets with a higher number indicating a higher income. The variable is transformed by taking the mid-point of each bracket and dividing it by the number of householders (per adult equivalent). The variable is also logged Index of a households assets and access to basic services. Based on Filmer and Pritchett (2001). See Table 2 Binary dummy variable taking the value of 1 for poor households (deprived in 33% or more of MPI indicators). Binary dummy variable taking the value of 1 for poor households (deprived in 33% or more of the Melanesian MPI indicators (see Clarke et al., 2014)) Binary dummy variable taking the value of 1 for poor households (deprived in 20% or more of MPI indicators) Binary dummy variable taking the value of 1 for households that are close to the poverty threshold (deprived in 20–33% of MPI indicators) Binary dummy variable taking the value of 1 for extremely poor households (deprived in 50% or more of MPI indicators) Based on the question “During the past two years, how have your earnings changed?” Measured using a Likert scale (1–5) Responses to “Compared with the rest of your country do you feel poor or rich?” Measured using a Likert scale (1–5) Responses to “Compared with the rest of your community do you feel poor or rich?” Measured using a Likert scale (1–5) Binary dummy variable taking the value of 1 if the respondent works for money Number of people in the household Proportion of household members that have finished primary and secondary school Binary dummy variable taking the value of 1 if the house has access to a garden Binary dummy variable taking the value of 1 if the house is on communally owned land Binary dummy variable taking the value of 1 if the household worried that their food would run out before they got money to purchase more
Male Age Income
Wealth Poverty (MPI) Poverty (MMPI) Poverty (20% threshold) Vulnerability Extreme poverty Internal relative wealth External relative wealth (Country) External relative wealth (Community) Employed Household size Education Garden Communally owned land Food insecure
ARE POOR PEOPLE LESS HAPPY? FINDINGS FROM MELANESIA
459
Table 5. Summary statistics Variable
Obs
Mean
sd
Min
Max
Happiness (scale from 1 to 10) Male Age Income Wealth Poverty (MPI) Poverty (MMPI) Poverty (20% threshold) Vulnerable Extreme poverty Internal relative wealth External relative wealth (Country) External relative wealth (Community) Employed Household size Education Garden Communally owned land Food insecure
619 619 619 500 619 619 619 619 619 619 619 619 619 619 619 619 619 619 619
7.30 0.54 41.14 5.46 0.04 0.31 0.33 0.55 0.24 0.09 3.14 2.77 2.92 0.96 5.53 0.47 0.77 0.53 0.55
2.33 0.50 14.01 0.91 1.36 0.46 0.47 0.50 0.43 0.29 0.98 0.69 0.65 0.20 2.79 0.35 0.42 0.50 0.50
1 0 17 3.30 3.09 0 0 0 0 0 1 1 1 0 1 0 0 0 0
10 1 86 8.18 3.17 1 1 1 1 1 5 5 5 1 23 1 1 1 1
Available online at www.sciencedirect.com
ScienceDirect