Poverty is a wicked problem that has consistently defied attempts at reduction to simple causes or processes. In recent years, much effort has been put into analyzing diverse findings on poverty (and the related issues of deprivation and differences in socioeconomic status [SES]) and in developing theoretical perspectives to integrate these findings. The article by Pepper & Nettle (P&N) contributes to this effort by drawing together various research lines on SES differences and temporal discounting, describing what they call a “behavioral constellation of deprivation” (BCD). P&N present an interesting perspective on deprivation, especially in the application of evolutionary models on the effects of mortality risk to SES differences, yet we have doubts about the integrative value of the “behavioral constellations.” Like the observation that constellations in the night's sky are not used in contemporary astronomy because they exist more in the eye of the beholder than in systematic relationships between celestial bodies, we argue that P&N's B*CD (1) overestimates the coherence of the various behaviors associated with poverty and (2) underrepresents the range of behaviors that should be included in such a constellation.
First, as far as coherence in characteristics of poverty is concerned, the empirical record has proven to be rather stubborn. Various reviews have come to the conclusion that results are not consistent across methodologies (Duncan et al. Reference Duncan, Magnuson and Votruba-Drzal2017), that there is no conclusive support for any single explanation (Pampel et al. Reference Pampel, Krueger and Denney2010) and that there is no common solution to problems of poverty (Banerjee & Duflo Reference Banerjee and Duflo2011). Such empirical variation makes it hard to talk about a behavioral constellation or about exclusive psychological or environmental factors underlying such a constellation.
Studies on behavioral and psychological characteristics of low SES and poor samples rarely include the full range of measures representing a constellation. Rather, evidence for constellations mainly comes from narrative reviews like the one by P&N, drawing together findings from separate studies without clearly explaining criteria for their inclusion or exclusion. A risk of this strategy is selectively including only those studies that provide convergent evidence. There are, however, many divergent results. For example, correlational studies and (quasi-)experimental studies on the consequences of poverty have frequently yielded markedly different results (Duncan et al. Reference Duncan, Magnuson and Votruba-Drzal2017). In addition, the direction and magnitude of effects have been found to vary across behavioral phenomena associated with poverty. For example, poor people have been found to sometimes make worse decisions, because poverty “impedes cognitive function” (Mani et al. Reference Mani, Mullainathan, Shafir and Zhao2013), but at other times make better decisions, because “scarcity frames value” (Shah et al. Reference Shah, Shafir and Mullainathan2015). Other studies, linking poverty with decision making, do not show any consistent effects at all. For example, Carvalho et al. (Reference Carvalho, Meier and Wang2016) do not find any differences between before and after payday in relation to risk taking, quality of decision making, and cognitive function tasks. Sometimes, a lack of systematic differences can be explained by complex relationships underlying observations. For example, Callan et al. (Reference Callan, Kim, Gheorghiu and Matthews2016) have found that personal relative deprivation and subjective SES class acted as mutual suppressors, obscuring the relationship between SES status and prosocial behaviors. These examples illustrate the variability and complexity of the empirical record on the effects of poverty on behavior when looking for constellations of behavior.
Second, poverty, SES, and deprivation are such broad constructs that one would expect them to relate to a broad set of behaviors. Indeed, the literature on these constructs is diverse, ranging from health-related behaviors, to emotional experiences, to social and moral behaviors. Likewise, the range of associated environmental factors and psychological processes explaining such behaviors is much broader than those proposed by P&N. It includes, for example, reduced cognitive bandwidth (Mullainathan & Shafir Reference Mullainathan and Shafir2013); stress and negative affect (Haushofer & Fehr Reference Haushofer and Fehr2014); experienced societal rank and increased contextualism (Kraus et al. Reference Kraus, Piff, Mendoza-Denton, Rheinschmidt and Keltner2012); childhood economic conditions, impulsivity, and risk (Griskevicius et al. Reference Griskevicius, Ackerman, Cantú, Delton, Robertson, Simpson, Thompson and Tybur2013); culture and inheritance of dysfunctional beliefs, values, and behaviors (Lewis Reference Lewis1966); shame and stigma (Walker Reference Walker2014); and generalized trust (Hamamura Reference Hamamura2012). P&N choose to be rather restrictive in their inclusion of processes and behaviors, focusing on extrinsic mortality risk, lower environmental control, and increased temporal discounting. Because these factors have also been included in previous overviews of the link between poverty and decision making, such as by Mullainathan and Shafir (Reference Mullainathan and Shafir2013) and Haushofer and Fehr (Reference Haushofer and Fehr2014), the question is what such a restrictive constellation adds to our understanding of poverty. One possibility may be the application of models from evolutionary biology, answering the questions of why behavioral constellations should be observed in the first place and how they can be seen as contextually appropriate responses. However, the question remains: Why do P&N not apply this reasoning to a wider range of behaviors? Perhaps the most explicit omission is that of risk, which is assumed to be directly related to wealth in classical economic models and has been explicitly related to poverty by Griskevicius et al., who argue that people who grew up in poverty are not only less likely to defer immediate rewards but also should be more risk seeking in times of stress and when exposed to mortality cues. Although the evidence on risk is mixed (Carvalho et al. Reference Carvalho, Meier and Wang2016), a behavioral constellation including a broader range of behaviors would clearly be of more heuristic value to researchers and practitioners dealing with poverty.
To conclude, we think that P&N contribute an interesting perspective on poverty and associated behavior that merits further study. However, at the same time, we believe that the diversity of the empirical record and the narrow focus of their paper clearly limits their claim for the existence of a BCD. In line with more situational analyses (Banerjee & Duflo Reference Banerjee and Duflo2011; Bertrand et al. Reference Bertrand, Mullainathan and Shafir2004), we believe that problems of deprivation and poverty for the moment benefit more from specific, tailor-made analyses and solutions than from broad constellations that might exist more in the eye of the beholder than in the empirical record.
You have
Access
Poverty is a wicked problem that has consistently defied attempts at reduction to simple causes or processes. In recent years, much effort has been put into analyzing diverse findings on poverty (and the related issues of deprivation and differences in socioeconomic status [SES]) and in developing theoretical perspectives to integrate these findings. The article by Pepper & Nettle (P&N) contributes to this effort by drawing together various research lines on SES differences and temporal discounting, describing what they call a “behavioral constellation of deprivation” (BCD). P&N present an interesting perspective on deprivation, especially in the application of evolutionary models on the effects of mortality risk to SES differences, yet we have doubts about the integrative value of the “behavioral constellations.” Like the observation that constellations in the night's sky are not used in contemporary astronomy because they exist more in the eye of the beholder than in systematic relationships between celestial bodies, we argue that P&N's B*CD (1) overestimates the coherence of the various behaviors associated with poverty and (2) underrepresents the range of behaviors that should be included in such a constellation.
First, as far as coherence in characteristics of poverty is concerned, the empirical record has proven to be rather stubborn. Various reviews have come to the conclusion that results are not consistent across methodologies (Duncan et al. Reference Duncan, Magnuson and Votruba-Drzal2017), that there is no conclusive support for any single explanation (Pampel et al. Reference Pampel, Krueger and Denney2010) and that there is no common solution to problems of poverty (Banerjee & Duflo Reference Banerjee and Duflo2011). Such empirical variation makes it hard to talk about a behavioral constellation or about exclusive psychological or environmental factors underlying such a constellation.
Studies on behavioral and psychological characteristics of low SES and poor samples rarely include the full range of measures representing a constellation. Rather, evidence for constellations mainly comes from narrative reviews like the one by P&N, drawing together findings from separate studies without clearly explaining criteria for their inclusion or exclusion. A risk of this strategy is selectively including only those studies that provide convergent evidence. There are, however, many divergent results. For example, correlational studies and (quasi-)experimental studies on the consequences of poverty have frequently yielded markedly different results (Duncan et al. Reference Duncan, Magnuson and Votruba-Drzal2017). In addition, the direction and magnitude of effects have been found to vary across behavioral phenomena associated with poverty. For example, poor people have been found to sometimes make worse decisions, because poverty “impedes cognitive function” (Mani et al. Reference Mani, Mullainathan, Shafir and Zhao2013), but at other times make better decisions, because “scarcity frames value” (Shah et al. Reference Shah, Shafir and Mullainathan2015). Other studies, linking poverty with decision making, do not show any consistent effects at all. For example, Carvalho et al. (Reference Carvalho, Meier and Wang2016) do not find any differences between before and after payday in relation to risk taking, quality of decision making, and cognitive function tasks. Sometimes, a lack of systematic differences can be explained by complex relationships underlying observations. For example, Callan et al. (Reference Callan, Kim, Gheorghiu and Matthews2016) have found that personal relative deprivation and subjective SES class acted as mutual suppressors, obscuring the relationship between SES status and prosocial behaviors. These examples illustrate the variability and complexity of the empirical record on the effects of poverty on behavior when looking for constellations of behavior.
Second, poverty, SES, and deprivation are such broad constructs that one would expect them to relate to a broad set of behaviors. Indeed, the literature on these constructs is diverse, ranging from health-related behaviors, to emotional experiences, to social and moral behaviors. Likewise, the range of associated environmental factors and psychological processes explaining such behaviors is much broader than those proposed by P&N. It includes, for example, reduced cognitive bandwidth (Mullainathan & Shafir Reference Mullainathan and Shafir2013); stress and negative affect (Haushofer & Fehr Reference Haushofer and Fehr2014); experienced societal rank and increased contextualism (Kraus et al. Reference Kraus, Piff, Mendoza-Denton, Rheinschmidt and Keltner2012); childhood economic conditions, impulsivity, and risk (Griskevicius et al. Reference Griskevicius, Ackerman, Cantú, Delton, Robertson, Simpson, Thompson and Tybur2013); culture and inheritance of dysfunctional beliefs, values, and behaviors (Lewis Reference Lewis1966); shame and stigma (Walker Reference Walker2014); and generalized trust (Hamamura Reference Hamamura2012). P&N choose to be rather restrictive in their inclusion of processes and behaviors, focusing on extrinsic mortality risk, lower environmental control, and increased temporal discounting. Because these factors have also been included in previous overviews of the link between poverty and decision making, such as by Mullainathan and Shafir (Reference Mullainathan and Shafir2013) and Haushofer and Fehr (Reference Haushofer and Fehr2014), the question is what such a restrictive constellation adds to our understanding of poverty. One possibility may be the application of models from evolutionary biology, answering the questions of why behavioral constellations should be observed in the first place and how they can be seen as contextually appropriate responses. However, the question remains: Why do P&N not apply this reasoning to a wider range of behaviors? Perhaps the most explicit omission is that of risk, which is assumed to be directly related to wealth in classical economic models and has been explicitly related to poverty by Griskevicius et al., who argue that people who grew up in poverty are not only less likely to defer immediate rewards but also should be more risk seeking in times of stress and when exposed to mortality cues. Although the evidence on risk is mixed (Carvalho et al. Reference Carvalho, Meier and Wang2016), a behavioral constellation including a broader range of behaviors would clearly be of more heuristic value to researchers and practitioners dealing with poverty.
To conclude, we think that P&N contribute an interesting perspective on poverty and associated behavior that merits further study. However, at the same time, we believe that the diversity of the empirical record and the narrow focus of their paper clearly limits their claim for the existence of a BCD. In line with more situational analyses (Banerjee & Duflo Reference Banerjee and Duflo2011; Bertrand et al. Reference Bertrand, Mullainathan and Shafir2004), we believe that problems of deprivation and poverty for the moment benefit more from specific, tailor-made analyses and solutions than from broad constellations that might exist more in the eye of the beholder than in the empirical record.
ACKNOWLEDGMENTS
Support from the COFUND EU Marie Sklodowska-Curie program and from the Basic Research Program of the National Research University Higher School of Economics is gratefully acknowledged.