Time to Mainstream the Environment into Migration Theory?

Abstract

As with all social processes, human migration is a dynamic process that requires regular theoretical reflection; this article offers such reflection as related to the role of the natural environment in contemporary migration research and theory. A growing body of evidence suggests that environmental contexts are increasingly shifting social and ecological realities in ways that are consequential to migration theory. We review some of this evidence, providing examples applicable to core migration theories, including neoclassical economic and migration systems perspectives, the “push-pull” framework, and the new economics of labor migration. We suggest that neglecting consideration of the natural environment may yield misspecified migration models that attribute migration too heavily to social and economic factors particularly in the context of contemporary climate change,. On the other hand, failure to consider migration theory in climate scenarios may lead to simplistic projections and understandings, as in the case of “climate refugees”. We conclude that migration researchers have an obligation to accurately reflect the complexity of migration’s drivers, including the environment, within migration scholarship especially in the context of global climate change.

A variety of disciplinary perspectives come together to inform migration theory, including economics, geography, political science, and sociology (e.g., White 2016; Wickramasinghe and Wimalaratana 2016). Together, these perspectives have provided the lenses through which migration scholars have examined and advanced our understanding of population movement for decades (e.g., O’Reilly 2015). Yet as Zolberg noted (1989a: 404), the most “stimulating” migration theory is that which pays “appropriate attention to the changing specificities of time and space,” and this “historicization of migration theory implies that theoretical concerns and emphases must be modified in the light of changing social realities.” Zolberg wrote these lines in a time of intensifying global inequalities, the politicization of borders, and exits from socialist states, among other substantial socioeconomic and political shifts. In this article, we suggest that contemporary climate change is yielding new socioeconomic, political, and ecological realities of relevance to migration theory.

An increasing body of evidence identifies an environmental ‘signal’ in migration as households engage migration as an adaptive strategy in the face of environmental stress (e.g., Antwi-Agyei and Nyantakyi-Frimpong 2021). Especially in the context of contemporary climate change, such findings suggest that migration scholars may be well served by regularly, and critically, considering aspects of the natural environment as context. Findings from such environmentally informed research are critical to further refine migration theory.¹

We build on Piguet’s (2013) useful articulation of the presence, absence, and re-emergence of environmental considerations¹ in migration research since the late 1880s to articulate the implications of failure to include environmental factors in migration theory. Of course, human history is replete with examples of difficult environmental conditions instigating population movement as an adaptive strategy in the face of stress (McLeman 2014). But as Piguet (2013) notes, over the past two decades, a body of research has emerged on the environmental dimensions of migration in more contemporary times. As recent examples of this emerging literature, temperature and rainfall extremes dampen urban outmigration across East Africa (Mueller, Sheriff et al. 2020) and shifts in the timing and amount of rainfall influence return migration in Thailand (Entwisle, Verdery, and Williams 2020). That said, while environmental factors emerge as important predictors of migration, they do not typically act upon migration separately from other socio-economic and political influences (e.g., Black et al. 2011). For instance, environmental factors reduce urban employment options in the case of East Africa (Mueller, Sheriff et al. 2020), or their influence may be filtered through social networks, as in the case of Thailand (Entwisle et al. 2020).

Related to these empirical findings, an important question emerges: If environmental factors play a role in migration processes but are excluded from migration theory and, subsequently, empirical models, do we risk omitted variable bias? Indeed, misspecified models can yield biased estimates and, therefore, an incomplete (or worse, incorrect) understanding of the processes under investigation. Important policy levers may be missed by such incomplete understanding. For example, since migration is often an adaptive strategy in the face of environmental stress such as rainfall shortage or intense heat (e.g., Carman and Zint 2020; McLeman and Smit 2006), understanding migration in the context of livelihood strategies can suggest actions to increase households’ adaptive capacity and, thereby, reshape migration patterns (e.g. Mueller et al. 2020).

The argument developed here extends an essay by Hunter and Menken (2015) that explores the possibility of climate change shifting demography’s “normal science” (Kuhn 1962). In his classic work, Kuhn describes normal science as a contemporary paradigm that legitimates current research questions and approaches. However, if sufficient anomalies arise within research findings – surprises not anticipated by the contemporary paradigm – a shift may occur within the paradigm, which then guides new scholarship (Kuhn 1962).²

Importantly, our aim is not to provide a comprehensive overview of migration theory or migration-environment research (for reviews, see Arango 2000, 2017; Hunter, Luna, and Norton 2015; Kaczan and Orgill-Meyer 2020). Instead, our goal is to critically explore the “environment” as related to classical migration theories by engaging research findings that have emerged in the past two decades. We begin with a broad discussion of “environment” as related to migration, including definitional issues, exploration of potential reasons underlying its historical exclusion from migration and demographic research more broadly, and the risks associated with climate determinism. Next, we offer brief introductions of six theoretical perspectives engaged within contemporary migration research, each introduction followed by research examples demonstrating environmental influences relevant to that particular theoretical lens.

We cover core functionalist theories, including neoclassical economics, migration systems, and “push-pull” approaches, followed by theories in the historical-structural paradigm, including political economic and world systems.³ As noted in a recent contribution by de Haas (2021:1), migration theory has remained under-theorized and “at an impasse for several decades,” even in the context of dramatic increases in empirical studies on migration. In an effort to move theorization forward, we encourage migration scholars to regularly and critically consider if/how the environment may be usefully integrated into the perspectives engaged within their research.

The Place of Environment in Migration Theory: Considering Errors of Exclusion and Inclusion

Increasing interest in the environmental correlates of migration over the past two decades has been, in part, fueled by alarmist projections made in the 1990s about “climate refugees” (Hunter, Luna, and Norton 2015). Such projections reflected strong “environmental determinism,” or a position in which environmental factors are elevated to causal primacy within explanations of social processes (Hulme 2011; Piguet 2013). Subsequent research added nuance to simplistic projections, revealing several broad findings, including evidence of migration as a risk diversification strategy in the face of environmental pressures; the influence of household and individual factors in migration decision-making; and the importance of social networks and interactions between environmental factors and socio-economic-political processes (Hunter et al. 2015).

Environmental determinism is not, however, confined to overly simplistic projections of climate refugees. Hulme (2011) suggests it can also result from an over-emphasis on predictive science, resulting in disproportionate power being awarded to model-based understandings of climate’s influence on social life. In his essay, “Reducing the Future to Climate,” Hulme (2011) contends that the search for “simplistic chains of climatic cause-and-effect” neglects recognition of the complex entanglement of environment and society and can represent methodological reductionism (Hulme 2011: 253).⁴ Such reductionism holds the potential for errors associated with the inclusion of environmental factors in migration theory and analyses. That is, when incorporating environmental concerns into empirical investigations of migration, scholars should avoid attributing environmental factors as sole, deterministic drivers of social processes (Meyer and Guss 2017; Piguet 2013) – a problem which led to the alarmist predictions of climate refugees (e.g., Meyers 1993).

Yet, failure to consider environmental factors in migration theory and research, especially given contemporary climate change, may result in “environmental indeterminism,” whereby, the role of the environmental factors is minimized or entirely neglected. Such exclusion may ultimately lead to incomplete (or worse, incorrect) understanding of migration dynamics, especially as environmental conditions continue to change.

Given the potential political consequences of either extreme – environmental determinism and indeterminism – the stakes are quite high.⁵ Environmental determinism and alarmist projections of climate refugees may feed into the nationalism and anti-immigrant sentiment documented in many regions across the globe, especially in recent years (Henrich 2020; Kaufmann 2019; Polynczuk-Alenius 2020). Yet climate indeterminism, or lack of consideration of environment, prevents the generation of insights to inform climate-responsive policies and programs to help reduce or mitigate risk as well as to support climate-impacted households (e.g. Kuriakose et al. 2013).⁶ Ultimately, we advise migration scholars to seek an “adequate and creative tension” (Hulme 2011: 246) between the fallacies of environmental indeterminism and determinism. Mainstreaming environmental considerations into migration theory, we argue, will ensure their consideration and, in this way, a more holistic and realistic understanding of migration-environment associations.

Migration’s environmental influences

Prior to introducing migration theory, it is useful to define “environment” as used here. In general, we are addressing aspects of the physical, natural environment, including qualities of air, water, soil, as well as “natural” disasters.⁷ Overall, environmental stressors related to human migration are multi-faceted and can be seen along a continuum (Hugo 1996). For instance, migration is influenced by acute, short-term environmental stressors, such as typhoons (e.g., Gröger and Zylberberg 2016) and earthquakes (e.g., Thiri 2017), as well as by chronic, long-term environmental strain, such as drought (e.g., Debnath and Nayak 2020).⁸ Additionally, acute or short-term events can occur within the context of longer-term environmental change. As specific to climate change, climate researchers suggest that while it is challenging to associate a single event (e.g., hurricane, flood) with increasing greenhouse gases, the probability of such events is increasing in many parts of the globe (Seneviratne et al. 2012). To illustrate, estimates suggest that the probability of a very hot European summer like that of 2003 has doubled as a result of human influences (Seneviratne et al. 2012).

Across the globe, displacement due to natural disasters outnumbers displacement due to conflict and violence by three to one – of the 31.1 million new internal displacements in 2016, fully 24.2 million were attributed to disasters (IDMC 2017). Contemporary movement across international boundaries can also have environmental dimensions (e.g., Riosmena, Nawrotzki, and Hunter 2018). For instance, drought, heat, and natural resource scarcity have been empirically associated with international migration from settings as diverse as Bangladesh, Mexico, and Pakistan (Lu et al. 2016; Mueller, Gray, and Kosec 2014; Nawrotzki et al. 2015). In many cases, migration acts as an adaptive strategy, allowing for income diversification by sending a household member to a distant labor market (e.g, Atuoye et al. 2019). As noted above, however, environmental pressures are often indirect (e.g., Mueller, Gray, and Hopping 2020) and intersect with economic, social, and political contexts to ultimately influence migration (Black et al. 2011). In cases of acute events, however, the environmental event may be a strong and direct migration determinant (Ponserre and Ginnetti 2019).

The following review engages a variety of environmental stressors as related to international migration. We emphasize six theoretical perspectives that are likely familiar to migration scholars. Combined, they address core theoretical questions within migration, including the spatial patterns of migration flows, migration motivations, and related political economic issues (Pryor 1981). Each of the sections below provides a brief review of a particular perspective, followed by contemporary research that suggests the role of environmental factors within the highlighted perspective. Importantly, the research examples provided are not the product of a systematic or comprehensive review. Instead, they have been chosen as contemporary exemplars that represent different environmental stressors (e.g. heat stress, typhoons) and interactions with different socio-cultural, economic, and political forces. Together, we hope they motivate migration scholars to seek that “adequate and creative tension” (Hulme 2011:246) as they incorporate environmental factors into the theories that guide their research.

Neoclassical Economics Perspectives on Migration Flows

Foundational to neoclassical arguments are assumptions about utility maximization and rational choice such that individuals are understood to be independent economic actors and migration is framed as informed by a cost-benefit calculation (DaVanzo 1981; Todaro 1976). According to the neoclassical perspective, such cost-benefit calculations emphasize investments in human capital, as well as wage differentials between origins and potential destinations (e.g., DaVanzo and Morrison 1981; Massey et al. 1993). At its core, the neoclassical economics perspective assumes perfect competition and information, and within this context, migration becomes a process through which equilibrium can be reached (e.g., de la Garza 1998).

At the macro scale, neoclassical perspectives emphasize spatial differences in labor supply and demand and resultant labor migration (e.g., Wickramasinghe and Wimalaratana 2016). An example is Lewis’s (1954) “dual economy models” that describe migration flows as emanating from differences in labor supply and demand across rural and urban sectors. Growing out of trade theory, Lewis argued that the positive wage differential in modern urban spaces, relative to rural spaces, resulted in the absorption of surplus rural agricultural labor through migration. In general, he argued that wage differentials across locations shape labor migration, which, then, moves toward labor-scarce regions offering greater opportunity and returns to capital (Lewis 1954). Yet application of such perspectives has moved far beyond only rural-urban dynamics and now interrogates the increasingly diverse migration flows resultant of a globalizing economy (e.g., Windzio 2018). Still, in this work, migration represents a central mechanism through which an equilibrium can be reached between labor supply and demand, with wage differentials, or expected income, representing the primary “push” and “pull” shaping migration’s spatial patterning (Pritchett and Hani 2020).

Through the years, there have been many critiques of neoclassical theories of migration (e.g., Porumbescu 2018). As an example, Canales (2019) contends that the neoclassical emphasis on wages and individual actors cannot explain migration’s complexities in today’s global, postmodern society. Instead, he argues, we need to begin with theories that explain the social and economic processes from which migration emerges (Canelas 2019). We further suggest that environmental processes are increasingly part of the context from which migration emerges and may deserve more consistent consideration.

Examples of environmental factors within neoclassical perspectives at the macro scale

One pathway through which environmental characteristics can shape migration within a neoclassical framework is through the impacts of environmental amenities on wages and market rents (e.g., Kerr 2017). An example would be the appeal of moderate climates (e.g., Cragg and Kahn 1999) and the adjustment of wages and market rents to such quality-of-life factors. As concisely explained by Albouy (2008:1), “workers will live in a place where real wages are low if their lower consumption of market goods is offset by a higher consumption of non-market amenities.”

Graves (1980) provided conceptual and empirical articulation of the potential for such non-economic values to become increasingly important in understanding migration. Three decades ago, Graves (1980) saw increasing migration flows to rural areas in the United States, improvements in information transfer, and overall increases in willingness to migrate as forces suggesting that income and unemployment differentials could no longer accurately capture overall utility differentials. He modeled gross migration flows at the scale of metropolitan areas with standard income and unemployment data, as well as integrated climate variables. Graves (1980) found that including climate variables (30-year “normal” for temperature, humidity, and wind) substantially increased explanatory power, compared to estimates including only municipal-scale income and unemployment. In a review of the historical development of interregional migration models, Biagi and Dotsel (2018:28) note the consideration of place-based amenities in migration decisions as “one of the most significant breakthroughs of this period.” over the subsequent two decades. In fact, in 2015, Beine and Parson (2015:725) lamented that “environmental determinants have remained wholly absent from the macroeconomic literature on the determinants of migration.” In their subsequent work, they identified an indirect effect of climate through widening the origin-destination gap in wages since environmental strain put downward pressure on wages at origin (Beine and Parsons 2017).

An increasing number of studies are integrating environmental factors into economic modeling of migration patterns (Feng, Partridge, and Rembert 2018). For instance, research is expanding on amenity values as related to migration (e.g., Fan, Fisher-Vanden, and Klaiber, 2018; Hjerpe, Hussain, and Holmes, 2020; Oliveira and Pereda 2020). A useful summary of the emergence of this literature is provided by Rudzitis, Graves, and Moss (2014).⁹

To the extent that continued climate pressures may exacerbate wage differentials, failure to explicitly integrate climate risks into migration modeling may result in missed opportunities to better understand the influence of contemporary climate change on macro-socioeconomic processes. The omission of environmental considerations would be an example of environmental indeterminism. The reverse is also true since it is possible to over-emphasize environmental effects. Fan et al. (2018; 644) offer an example, as they argue that “studies that do not account for labor market and housing market feedbacks likely overstate the economic impacts of climate-induced migration.”

Migration Systems as an Analytical Lens

Like neoclassical economics, migration systems approaches involve equilibrium concepts as they focus on stable groups of countries (or other geographic units) that send and receive migrants (e.g. Arango 2004). Early articulations of migration systems drew heavily upon functionalist social theory with emphasis on systems as important elements of societal structure (Bakewell, de Haas, and Kubal 2012). As compared to neoclassical perspectives, migration systems approaches broaden consideration through more direct integration of the cultural, social, and political aspects of flows and counterflows between two or more places (see Mabogunje 1970). Such flows include not only people but also goods, services, information and ideas, and the analytical foci within systems perspectives are the factors that stimulate, direct, and sustain these origin-destination flows (Massey et al. 1998). While systems perspectives are as “old as the scientific study of migration” (Fawcett 1989:671), data improvements in the 1980s increased scholarly interest in this approach, and substantial theoretical and empirical progress has been made since that time (DeWaard and Ha 2019; Hauer, Holloway, and Takashi 2020).

Social networks have arguably received the most attention in examination of migration systems, especially regarding the perpetuation of migration flows (e.g., Massey et al. 1998). Networks represent structured sets of social relationships that act as pathways through which information and resources move in both directions (Gurak and Caces 1992). This information- and resource-sharing reduces the costs associated with migration and facilitates movement by, for example, easing the search for jobs and housing and providing information on policies and useful programs (Massey 1988). Importantly, the flows representative of migrant systems develop momentum over time, allowing them to function independently of their originating forces (Haug 2008).

Like neoclassical theories of migration, the systems approach is not without critiques (e.g., de Haas 2010). The critiques emphasize lack of investigation on the emergence of systems, with most investigations instead focused on their perpetuation (Bakewell 2010). Critics also argue that little is known about why systems sometimes do not emerge following the moves of pioneer migrants (de Haas 2010). In addition, Bakewell (2010) argues that little attention is paid to the shifts in the initial social and economic context following the development of a migration system, while the factors that shape system decline are also underexplored (de Haas 2010). We build upon these critiques by arguing that environmental factors represent an underexamined aspect of migration systems, potentially playing a role in system emergence, perpetuation, and decline.

Examples of environmental factors within migration systems

As expansions to migration systems research, some scholars have explicitly examined the environmental dimensions of reciprocal migration flows between places (e.g., Curtis, Fussell, and DeWaard 2015; Hauer 2017). Such dimensions include environmental amenities, natural disasters, and different aspects of climate change (e.g., Hauer 2017; Suckall et al. 2015). The research in this area suggests that environmental factors have a role in stimulating, directing, sustaining, and eventually changing origin-destination flows (e.g. Fussell, Curtis, and DeWaard 2014; McLeman 2006). For example, international amenity-related migration systems have emerged as retiree flows from cooler climates to Costa Rica (e.g. Matarrita-Cascante 2017) and Italy (e.g. King et al. 2021).

Environmental factors can not only generate new migration systems but also reshape existing ones, as research on Hurricane Katrina shows (Curtis et al. 2015; DeWaard, Curtis, and Fussell 2016). Katrina was a disastrous storm that struck the US Gulf coast in August 2005. Nearly 80 percent of New Orleans’ residents evacuated prior to the hurricane’s landfall (Fussell et al. 2014). Tracking origins and destinations with Internal Revenue Service county-to-county migration flow data, Fussell and colleagues (2014) examined “recovery migration” reflecting both returning residents and new in-migrants during the region’s recovery period (2007-2009). They found that the hurricane shifted the region’s existing migration system in that, as compared to previous in-migration to New Orleans, post-disaster recovery migration was more likely to be from nearby counties. On the other hand, outmigration ties between New Orleans and particular destinations dramatically decreased in the recovery phase (Fussell, Curtis, and DeWaard 2014).

Systems can also develop when migration is used as an adaptive strategy in the face of chronic environmental stressors. For instance, social networks support Mexico-US migration in the face of drought-like conditions, where international migration is most likely from communities with particularly strong existing systems of transnational networks (Riosmena, Nawrotzki and Hunter 2018). Additionally, migration systems have been documented following acute environmental events, as Hauer, Holloway, and Takashi (2020) identify two distinct migration systems following the Great East Japan Earthquake and Tsunami in 2011 – a short distance one for evacuees and a longer-distance one for permanent migrants.

Since migration may be seen as a necessary response to some natural disasters, there may be a temptation toward climate determinism when evaluating environmentally related displacement. However, such an interpretation would belie more complex forces. As an example in contemporary times, sea level rise may be seen as the ultimate “push” as a result of deeming coastlines uninhabitable. Yet recent work clearly positions migration in response to sea level rise as embedded within political, demographic, economic and social factors (Hauer et al. 2020). Further, these factors are themselves embedded within policy incentives that shape migration by encouraging (e.g., managing retreat) or lessening movement (e.g., constructing a sea wall) (Hauer et al. 2020). Of course, in such situations, the exclusion of environmental context in examination of coastal migration would yield a short-sighted understanding of population movement, and, in this way, migration systems represent another example where an “adequate and creative tension” (Hulme 2011:246) should be sought.

“Push-Pull” Perspectives on Migration

The above review of migration through the perspectives of macro-scale neoclassical economics and migration systems delved into the factors that instigate and perpetuate aggregate migrant flows. Moving now to a micro-perspective, we consider the factors that shape people’s migration decision-making and, ultimately, migration decisions. Lee’s (1966) “push-pull” framework offers one theoretical perspective at the micro level. Simply put, push factors characterize origins and can include low employment opportunities, wage levels, and high cost of living, all of which are commonly noted by migrants as primary in decision-making (Black et al. 2013). Other factors can be political and include the effectiveness of governance, presence of conflict, and/or the state of immigration policy (Alvarado and Massey 2010). Simultaneously, characteristics of destinations can pull would-be migrants, such as better employment prospects and political stability.

In response to critiques of push-pull as overly determinist (e.g. de Haas 2010; Skeldon 1990), Van Hear, Bakewell and Long (2018) present a useful framework dubbed “push-pull plus.’ To better characterize interactions across drivers at all scales, they outline migration’s predisposing, proximate, precipitating, and mediating drivers. Migration’s predisposing drivers shape the context from which migration may occur (Van Hear et al. 2018). Economic disparities between places represent a predisposing factor (e.g., Abel et al. 2019; Cohen 2018). As compared to these broad contextual factors, proximate drivers are more direct, such as a localized economic downturn (Bank, Fröhlich, and Schneiker 2017). Even more direct are precipitating drivers, such as a factory closure embedded within that local economic downturn (Van Hear et al. 2018). The influence of predisposing, proximate, and precipitating factors on migration is further shaped by mediating factors, which enable or constrain migration, such as transportation, communications, and resource infrastructures, as well as household capitals (Garip 2008; Palloni et al. 2001). Lack of such mediating factors can constrain movement even in the face of severe precipitating drivers (Van Hear et al. 2018).

Examples of environmental factors in “push-pull” migration research

Rapid onset environmental disasters, such as hurricanes, floods and landslides, can act as strong push factors and increasingly do so across the globe (Dun 2011; Giannelli and Canessa 2021; Koubi et al. 2016). For instance, since 2008, nearly 230 million people have been displaced by sudden-onset hazards such as typhoons, floods, and wildfires (IDMC 2018). Much of this movement is shorter distance, typically within the same nation, and residents may generally have the intention of returning home (Findlay 2011). In this way, such displacement may not actually be considered “migration,” since there may be no intention to settle elsewhere for any length of time. Yet environmental stressors intersect with migration’s other drivers in ways consequential for the likelihood of permanent outmigration as well. The international relocation of Pacific Islanders in the face of sea level rise offers one example (Yamamoto and Esteban 2017).

Chronic environmental stress has also been associated with international migration in a wide variety of settings (Hunter et al. 2015). However, this association is most often identified in rural, agricultural settings where climate variability can impact agricultural productivity, as well as the natural resources harvested by households to meet daily needs (Falco, Donzelli, and Olper 2018). Indeed, excessive precipitation is associated with increased international migration from Senegal (Nawrotzki and Bakhtsiyarava 2019), while high temperature warm spells increased international migration from Mexico (Nawrotzki et al. 2016). In such cases, migration may be seen as a household adaptive strategy to diversify income sources and minimize risk (details below). Additionally, slow-onset environmental stressors can also drive migration by laying the groundwork for acute, rapid-onset events. Desertification can lead to wildfires, hot spells to heat waves, sea level rise to floods (IDMC 2017). At the same time, these slow-onset stressors can erode local populations’ socioeconomic capacity to cope, positioning a region for grave impact and potentially triggering displacement (Huq et al. 2015).

In all, this section demonstrates that dynamics within the natural environment can interact with a wide variety of migration drivers to ultimately influence migration. For example, environmental stressors can underlie economic downturns, food insecurities, and conflict, particularly as they are shaped by migration’s predisposing and proximate drivers, which themselves lay the foundation within which environmental stressors are experienced. Migration can also be seen as an adaptive response to environmental strain (e.g., Adams and Kay 2019; McLeman 2018), although migrants may not themselves report climate as a primary migration driver. Thus, failure to consider environmental factors in both theory and research has the potential to produce findings missing an important driver of migration (i.e., climate indeterminism as developed above).

New Economics of Labor Migration

At least in part a result of the critiques of the neoclassical theories of migration (described above), migration scholars, including those analyzing environmental dimensions of migration, have employed an alternative theoretical model, the new economics of labor migration, which expands understandings of who migrates and why (e.g., Stark and Bloom 1985). The new economics of labor migration (NELM) acknowledges that wage differentials are not the sole factor motivating individual migration (Stark and Bloom 1985). In fact, in contrast to neoclassical theories NELM focuses not on individuals but on households and families as the primary unit of decision-making (Stark and Bloom 1985). Much of the work using this framework has focused on international migration, illustrating why households in lower-income settings may want to send a migrant abroad (e.g., Gray 2009). Households may use temporary and circular migration to diversify income sources and minimize the risks associated with fluctuations and uncertainties in local markets (e.g., Bystander 2015). Given households’ limited access to credit and capital in many developing countries, remittances from a migrant abroad may also serve as supplemental income that can be put toward large household purchases or (re)investments in agricultural- or business-related improvements (Lindstrom and Lauster 2001; Massey 1999).

When a household decides to send a migrant to diversify income and minimize risks, it makes intuitive sense that they may prioritize the household member with the highest likelihood of success in both reaching the desired destination and securing employment (Moran and Taylor 2006). It is through this lens that we consider the question of who migrates. This question is inherently about migrant characteristics and the selective nature of the migration process. Indeed, migration scholars have demonstrated that migrants are often different from their non-migrant counterparts, finding some consistent selection patterns by age, sex, race/ethnicity, education, and health, as well as considerable variation in these relationships across settings and types of migration (e.g., Griga and Hadjar 2014; Morey et al. 2020; Riosmena, Wong, and Palloni 2013). To fully summarize these findings is beyond the scope of this article, so for this section, we focus on Mexico-US migration. Stark and Taylor’s (1991) analysis of rural households in Michoacán, Mexico, highlights many of these relationships. For example, when compared to internal migrants and non-migrants in the area, international migrants to the United States were more likely to be male, to come from larger, wealthier households, and to have family or friends already in the United States (Stark and Taylor 1991). These findings are robust in other areas of the country too. Indeed, Lindstrom and Lauster (2001) illustrate that both internal and international migrants came from more marginalized municipalities with fewer wage-earning opportunities, thereby fueling the need for income diversification and risk reduction among households in these areas.

When compared to other theoretical models, NELM has proven to be one of the best at predicting who migrates from Mexico (Massey and Espinosa 1997). However, Garip (2012) illustrates that migrants’ socio-demographic characteristics vary by period and cohort. In addition, gender scholars have critiqued the lack of consideration of gendered norms and unequal power relations as they negatively impact women’s’ ability to migrate (e.g., Oishi, 2005). Hughes (2021:388) suggests that the NELM framework is essentially too simplistic in its reliance on “straightforward, optimized, microeconomic calculations” while neglecting the complexity of cultural, political, and institutional forces. Thus, the answer to who migrates from within a household in Mexico is largely contingent on context – both time and place – which begs the question how the environment may influence these migrant decision-making processes in combination with other complex contextual forces related to culture and the political economy.

Examples of environmental factors within NELM

Given the strong evidence suggesting that Mexican livelihood strategies are often generated through household processes (e.g., Massey et al. 1993), the NELM framework has proven useful for integrating environmental factors into migration decision-making (for review, see Simon 2018). Remittances from a migrant can serve as an ex-ante risk mitigation strategy as households send a migrant to help mitigate potential future losses from crop failure or climate extremes (Quiñones et al. 2018). Alternatively, remittances may take the form of an ex-post means of coping following challenging environmental exposures like drought or heat stress (Maharjan et al. 2021).

Whether ex ante or ex post, remittances are an important income source in Mexico, as rural households are particularly susceptible to environmental stressors, with agricultural production contributing as much as two-thirds to household income portfolios (de Janvry and Sadoulet 2001). Thus, failure to consider environmental factors in neoclassical and NELM frameworks would neglect the important influence of the natural environment on Mexican households’ income and livelihoods in the first place. In fact, nearly 80 percent of all economic losses between 1980 and 2005 were estimated to be a result of extreme weather events (Saldana-Zorilla and Sandberg 2009), and reductions in crop yields resulting from climate variability are shown to increase emigration from rural states (Feng and Oppenheimer 2012).

In their study of drought and migration from 12 Mexican states, Hunter et al. (2013) demonstrate that measures of climate (i.e., precipitation) interact with existing social networks to shape the network’s influence. Specifically, emigration rates increased two years following drought, but only for households in communities with strong migration networks (Hunter et al. 2013). Other work demonstrates how strong community networks can also help rural households adapt in place by channeling remittances to affected areas and reducing the need for climate-related migration (Nawrotzki et al. 2015). Although directionality differs in these examples, key for our purposes is that these studies suggest that environmental factors influence the role of social networks within household migration decision-making.

A more recent study comparing migration-environment associations during a period of increasing (1995-1999) and decreasing (2005-2009) US migration flows finds that climate might actually “trap” people and households in place (Riosmena, Nawrotzki, and Hunter 2018). For example, the most rural households were less likely to migrate to US locations during hot or dry environmental conditions. But highlighting the complex ways in which climate interacts with known correlates of migration, Riosmena et al. (2018) also found that during hot and dry conditions, emigration only increased for households in less vulnerable places as measured using community international migration prevalence and degree of socioeconomic marginalization. This finding suggests that climate and the natural environment influence the type of household that is able to send a migrant at the baseline.

Taken together, this section illustrates several examples of research employing a NELM framework to understanding the environmental correlates of migration. Socio-demographic factors such as age, race, income, and education are important in distinguishing migrants from non-migrants, but scholars studying migration environment in Mexico have demonstrated that the natural environment influences both who migrates and from where.

Political Economic Perspectives

Returning to macro-structural considerations, some scholars consider international migration to be reflective of existing global inequalities (Faist 2016). As noted above, migration is indeed influenced by structural forces underlying cross-national income differentials such as state and governance failures and violence (de Haas et al. 2019). To this end, and as related to political, policy, and human rights issues, we discuss political economy and world-systems theories, including the political factors that may propel migration.

World-systems theory is one macro-level economic model that can be used to understand international migration flows or, at the very least, the factors that encourage international migration. This approach grew out of previous dependency theories and was originally elucidated by Wallerstein (1974, 1982), who classified nation-states as core, meaning the dominant capitalist economies, semi-peripheral, or peripheral. Although world systems theory is not inherently a migration theory, scholars have made connections between it and international migration, as some flows originate in peripheral nations, directed to the Global North (Castles 2013; Massey et al. 1993).

World systems theory and other political economic approaches argue that global capitalist penetration into emerging economies of the Global South create and maintain inequalities that spur migration (e.g., Alvarado and Massey 2010; Babb 2005). To illustrate, world systems theorists suggest that there is a selective migration process whereby the underdevelopment of peripheral nations encourages the highly skilled and educated to migrate to core nations, further undermining development (Morawska 2007). Additionally, the always-expanding search for new markets further enables conditions for internal migration in developing countries, as firms from core nations enter peripheral countries seeking lower wages and more limited regulations (Froebel, Heinrichs, and Kreye 1977). This dynamic is observed in Mexico, where internal migrants take low-paying maquiladora jobs in the export industries near the US border (Villarreal and Hamilton 2012).

Others point to the emergence of structural adjustment policies as an example of core nations encouraging underdevelopment in the Global South and, therefore, the conditions that necessitate migration (Alvarado and Massey 2010; Massey 1999). Structural adjustment policies were ostensibly designed to help countries pay off considerable debts, but many argue that these adjustments simply forced countries to overhaul their economies at the direction of elite capitalist firms and states and that the policies favored free market reforms, reductions to public services, trade liberalization, elimination of subsidies, and emphasis on crop exports (Alvarado and Massey 2010; Massey 1999). Given the new emphasis placed on export crops under liberalized trade regimes, investors bought farmland and encouraged monoculture practices, promoting deforestation, degrading the land, and increasing vulnerability and the potential for displacement in peripheral nations (McLeman, Schade, and Faist 2016).

Other aspects of the political economy also shape migration such as governance breakdown, which can result in conflict and subsequent population redistribution as refugees and migrants seek to escape political discrimination or violence (Zolberg, Suhrke, and Aguayo 1989). In fact, as of 2009, every nation state in Africa had sent or received political refugees (Raleigh 2011). While genocide, civil war, and other forms of violence produce migration and refugee situations, political uncertainty more generally can also serve as a migration “push.” Additionally, states seek to control the volume and direction of international flows through policies that define and regulate entry, duration of stay, work permit requirements, and citizenship criteria (Morawska 2007). As such, citizenship is fundamentally a form of inclusion and exclusion that can be based on race, ethnicity, religion, education, and language and is reflective of inequalities at the global scale (Brubaker 2014; Shachar 2009).

Of course, world systems perspectives are not without critics. Hughes (2021), for example, argues that such perspectives inappropriately maintain distinctions between macro- and micro-analytical levels and, thereby, do not integrate global power dynamics in analyses beyond the nation-state. We argue that bringing in micro-scale understandings of migrant experiences would also yield more emphasis on external pressures, such as environmental degradation at the local scale, including that wrought by global economic processes.

Examples of environmental factors within political economic perspectives

Environmental pressures intersect with global structural inequalities to, in many cases, amplify the migration “push” factors noted above (McLeman et al. 2016). The impacts of climate change are already disproportionately burdening poor nations and poor people, who contributed least to greenhouse gas emissions (Padilla and Serrano 2006; Puaschunder 2020). In fact, the most recent Intergovernmental Panel on Climate Change (IPCC) report warns that populations at the highest risk of adverse consequences of climate change are those already disadvantaged, such as indigenous populations and vulnerable households dependent on agriculture or coastal livelihoods (IPCC 2018). For example, sea-level rise is already a human rights issue, as some estimates predict that a sea-level rise of one meter would displace over 20 million people in the coastal regions of Bangladesh, Egypt, and Nigeria alone (McMichael and Lindgren 2011).

The United States will face displacement, too (Hauer 2017), with impoverished coastal areas in Alaska and Louisiana especially vulnerable to sea-level rise, land erosion, and permafrost thaw due to climate change (Maldonado et al. 2012). Tribal communities are already having to relocate due to inadequate government support for adaptation strategies (Maldonado et al. 2012). Yet other populations may be trapped in place, due to lack of adequate social and financial capital to leave challenging conditions (DeWaard and Nawrotzki 2018). Planned resettlement, one potential solution, can come with considerable negative impacts, including loss of community and culture, mental and physical health effects of displacement, and economic decline (McMichael and Lindgren 2011). In addition, many facing forced displacement, such as those in low-lying Pacific Islands, wish to remain in their ancestral homelands due to strong place attachment (Mortreux and Barnett 2009).

While climate change presents new challenges to human migration patterns, the connection between the global political economy, environmental conditions, and migration is not new. Take the case of land degradation related to colonial histories which shaped settlement patterns, political borders, and agricultural systems. In Niger, mandated cash cropping in the early 1900s yielded substantial soil degradation, and the resulting food shortages necessitated regional migration and resulted in circular flows that still exist today (Afifi 2011). In Botswana, international policy such as the 1975 EU Lomé Convention – a trade agreement between European, African, Caribbean, and Pacific nations – altered agricultural policies and practices and, as a consequence, migration patterns (Sporton, Thomas, and Morrison 1999).

Taken together, this section outlines political economic explanations of migration with an emphasis on global inequalities. As mentioned above, Faist (2016) asserts that migration itself represents global inequalities in well-being, freedom, security, income, wealth, power, and life chances. We agree and would add that the environment often interacts with these forces, necessitating more regular, albeit critical, inclusion in political economic investigations of migration.

Where Now? The Future of Environment in Migration Theory and Research

The above review has engaged several core migration theories and provided exemplar research integrating environmental concerns within those perspectives. The review was motivated by the belief that human migration is a dynamic process requiring regular theoretical reflection “…in light of changing social realities” (Zolberg 1989). We do not contend that environmental influence on migration is a new social reality; it certainly is not (e.g., Schwindt et al. 2016). Instead, we argue that climate change holds the potential to increasingly influence contemporary social realities in ways that are consequential for migration theory and research. While environmental factors may not necessarily be expressed by migrants as front-and-center in their decision-making (Koubi et al. 2016), substantial evidence suggests that such factors act as indirect influences, especially as related to compromised livelihoods (e.g., Hunter et al. 2015). In some cases, extreme environmental events may even dominate as direct causes of involuntary population movements (e.g., Valdez-Pizzini 2020).

By structuring this review around theoretical perspectives, we have aimed to provide examples of the places where migration theories and contemporary migration-environment research findings align. Doing so provides the background for our contention that migration scholars should seek an “adequate and creative tension” (Hulme 2011:246) between the fallacies of environmental indeterminism and determinism. The former entails exclusion of environmental factors from migration theory, which risks inaccurate results and incomplete understandings. Yet inclusion of environmental factors also poses risks and must be undertaken carefully to minimize environmental determinism. In this way, scholars must work carefully to represent the complex interactions between the many contextual factors that influence migration decision-making and flows, environmental factors included.

Piguet (2013:151) suggests that one reason behind the environment’s disappearance in migration research over the past century is “the Western idea that progress implies a decreasing impact of nature on human fate.” As climate change continues to manifest in extreme events (Melillo, Richmond, and Yohe 2014), might there be more recognition of the interrelated fates of nature and humans? As the world grapples with climate change and resulting displacement (Rigaud et al. 2018), migration scholars must be ready to provide insight on population dynamics, guided by theory that can engage environmental concerns when and where needed. Mainstreaming environmental context into migration theory will ensure its consideration and, in this way, a more holistic and realistic understanding of migration-environment associations.

There are several pathways through which the migration-environment research community might move forward. First, we find inspiration in intriguing new work by de Haas (2021) that illustrates the potential for integration and expansion of functionalist and historical-structural perspectives to generate updated, comprehensive theorizations on migration. Through articulation of a useful “aspirations and capabilities” framework, de Haas (2021) expands on prior work by Sen (1999) and Berlin (1969) to consider the role of structural forces such as global capitalism and inequalities as combined with individual perceptions, preferences, and options. This effort is indicative of the potential for new forms of migration theorization that integrate insights from a variety of perspectives and push through impasse. A similar endeavor would be useful for highlighting the environmental dimensions of migration, and we hope this article provides a useful foundation for such an effort.

Second, the migration research community could be more consistent in critically considering environmental dimensions within examination of demographic processes, including migration. Existing migration-environment scholarship demonstrates the many ways in which environmental factors can be empirically integrated. For example, scholars have developed a variety of quantitative measures of precipitation and temperature to explore the association between weather anomalies and migration, with much of this work considering migration as a livelihood adaptation strategy (Eklund et al. 2016). As an example, with a focus on rural Bangladesh, Call et al. (2017) used flood data from the Dartmouth Flood Observatory to create an indicator of a flood during the past month, year, or two years. To predict migration, flood indicators were combined with a variety of theoretically-informed controls. They find that migration declined immediately after flooding but quickly returned to normal, suggesting that environmental variability acted as a short-term disrupter of livelihood strategies (Call et al. 2017). Using a similar approach, Thiede and Gray (2017) find that monsoon timing influences migration in Indonesia. Merging information from the Indonesian Family Life Survey with rainfall data from NASA’s Modern-Era Retrospective Analysis for Research and Applications (MERRA), they show that climate variability influences short-distance population movements more so than long-distance moves. Adding important nuance, Theide and Gray (2017) also find that effects vary by gender, location, and whether the individual resided in a farming household.

Aspects of local land conditions and tenure can also be usefully incorporated into migration research. For instance, a mixed-method investigation in northwestern Nicaragua reveals complex interactions between land, labor migration, and climate change (Radel et al. 2017). Here, survey and interview data provided insight on land status, including whether it was rented, owned, or borrowed. The size of plots was also ascertained, as was the type of agricultural activity, and an historical understanding of shifts in land tenure regimes was crafted from information gleaned from elderly respondents. Altogether, these data allowed for a powerful story of vulnerability and migration’s essential role as an adaptive strategy to provide for household needs in the context of increasing land scarcity (Radel et al. 2017). Many other research examples can be found in recent reviews by Borderon et al. (2019), Hunter et al. (2015), and Kaczan and Orgill-Meyer (2020). Methodological reflections also provide insight into the important contributions to be made by migration scholars (e.g., Fussell et al. 2014), while useful overviews of data and data challenges have also recently emerged (see Vinke and Hoffman 2020).

In all, progress has been made with regard to migration-environment research, and it is now time for theorization to catch up. It is our hope that this article provides a foundation from which migration scholarship can evolve to reflect the environmental challenges faced by millions. In addition to the development of more informed research findings, the integration of environmental factors into migration theory can guide identification of important policy levers. Given that migration is often an adaptive strategy in the face of environmental stress (McLeman and Smit 2006), better understanding livelihood choices, including migration, can inform actions to increase households’ adaptive capacity and improve human well-being. Understanding such associations is the role of the migration scholar, and perhaps it is time our “normal science” (Kuhn 1962) expand to more regularly incorporate environmental concerns, particularly in the era of contemporary climate change.

Table I.

Illustrative Environmental Dimensions of Migration and Links to Theory.

Theory	Example	Link to Environment	Example
Neoclassical	Lee (1966); Harris and Todaro (1970)	wage & market rent impacts; amenities & disamenities	Graves (1980); Beine and Parsons (2015)
Migration Systems	Mabogunje (1970); Massey (1988); de Haas (2010)	snowbirds & sunbirds; recovery migration following Hurricane Katrina	Litwak and Longino (1987); Fussell, Curtis and DeWaard (2014a)
Push-Pull	Lee (1966); Van Hear, Bakewell and Long (2018)	natural disasters, drought, wildfire, conflict	Strobl (2011); Nawrotzki and Bakhtsiyarava (2017)
New-Economics of Labor Migration (NELM)	Stark and Bloom (1985); Massey and Espinosa (1997)	ex ante risk mitigation strategy or ex post means of coping with environmental stress	Gray (2010); Riosmena, Nawrotzki and Hunter (2018)
World Systems Theory and Political Economy	Castles (2013); Wallerstein (1974); Wallerstein and Tompkins (1982); Zolberg, Suhrke and Aguayo (1989)	disproportionate climate impacts to global South; land degradation and colonial histories	Afifi (2011); Afolayan and Adelekan (1999); IPCC (2018)