Structure agency nexus and resilience

The application of the concept of resilience in some subject areas such as ecology is considered to be neglecting the role of human agency (Martin and Sunley, Reference Martin and Sunley2014). In contrast to this, the debates in social sciences may ignore the role of structural conditions in the resilience of social subjects and may present them as autonomous and therefore unaffected by such influences. Narratives of resilience as a unique individual response to hardship may accentuate personal characteristics and disregard the social conditions in which they emerge.

Hence, the long-standing debate on the structure‒agency nexus (Dolfsma and Verbürg, Reference Dolfsma and Verbürg2008) is highly relevant to any discussion on resilience. Let us recapture this debate briefly. In a nutshell, while most schools of thought recognised the interaction between structure and agency, some stressed the prominence of one in forming the other. This is true of the classical Marxist theory in spite of the criticism levied against it for having a highly deterministic view of social entities.

In other words, while Marx recognised the role of individuals in bringing about radical change, the latter cannot be entirely separated from the developments of the past.Footnote ³ Parsons has been considered a structural determinist but the importance he attributed to voluntary action, individual achievement and personal qualities, especially in his earlier works, show the role he attributed to agency in social systems (1937). The disparate approaches pursued by Durkheim (Reference Durkheim1964, Reference Durkheim1982) and Weber (Reference Weber1978) in analysing the role of structure and agency can be viewed in the same way.

Giddens has been credited for stressing the existence of a ‘false-dichotomy’ in the structure‒agency debate. Instead, he argued for a continuous interaction between the two (Giddens, Reference Giddens1979, see also Sewell Reference Sewell1992). Structure as social roles, rules and resources is considered to be produced and reproduced in what people do, which in turn is constrained and enabled by structure through the use of power, application of sanctions and communication of meaning (1984). However, Giddens implicitly discredited his early writings in his more recent work in which agent-centric explanations explicitly supersede structure.Footnote ⁴ Another landmark contribution to the debate is in Bourdieu's work on practices, habitus, class-status and distinction (Reference Bourdieu1977, Reference Bourdieu1984). Not only did he recognise the temporality of structures and the capacity of agents to be autonomous but he also perceived agency to be conditioned by social rules, cultural values and conventions. Hence, the limits on agents’ autonomy are recognised in their tendency to reproduce existing structures.

Overall, it is well established in the literature that social structure (including norms and traditions, class, gender, ethnic divisions, power relations, colonial influence and their economic and political manifestations) shapes individual behaviour and social relations. On the other hand, although humans are not dupes and have the capacity to show autonomy, their influence on structure is likely to be gradual, involving discontent, isolation and struggle(s) with progression through setbacks.

Let us consider all this in relation to the topic of this paper, that is, resilience of individuals, families or households in times of economic crises. However, before doing so, it may be necessary to attribute some practical meaning to the term. Economic hardship during crisis periods takes the form of income or consumption losses through, say, unemployment, business failures and cuts in public spending. Assume for a moment that resilience in such circumstances is reflected in stemming or recovering losses relatively quickly by finding new lines of income-generating activities or developing strategies to manage and meet the necessary expenses rather than enduring the crisis-induced hardship on a long-term basis.

Going back then: In a given structure if some individuals or families remain resilient despite being subject to economic shock, while others are not, what are the conditions that give rise to their resilience? Is resilience a unique individual trait or is it socially produced, constructed and conditioned? The current perspectives on household resilience place a heavy emphasis on agency with very little recognition of structural conditions of resilience. As discussed in the previous section, resilience is often considered to reflect traits such as attitude and the abilities of individuals to deal with hardship. For example, resilience during and in the aftermath of the 2008 economic crisis is frequently discussed in connection with the depth and breadth of the ‘shock’, impacts of rising unemployment and declining incomes and how individuals or households respond in the face of changing living conditions. This obscures wider structural factors such as class, asset ownership, redistributive systems, cultural practices and so on. While resilience in times of economic hardship may be an occasional need for high-income households, it is likely to be more relevant for low income and poorer households as argued by Orthner et al. (Reference Orthner, Jones-Sanpei and Williamson2004) and Okech et al. (Reference Okech, Howard, Mauldin, Mimura and Kim2012) who view households’ vulnerability to shocks as associated with their wealth and income.

Likewise, considering resilience in isolation from power, politics and participation may be a futile exercise. This is acknowledged by various studies including Davidson (Reference Davidson2010), MacKinnon and Derickson (Reference MacKinnon and Derickson2013) and Keck and Sakdapolrak (Reference Keck and Sakdapolrak2013). Davidson discusses how concentration of power and privilege influences control over resources and decision making. This in turn determines the outcomes with respect to resilience or a lack of it, both at systemic as well as individual levels. Lack of focus on social conditions of resilience lead MacKinnon and Derickson to conclude that mainstream analyses in resilience studies are conservative in that they fail to incorporate the ‘critical role of the state’ and neglect inequalities.

Indeed, the degree to which households show resilience in times of ‘economic shock’ depends on the political choices made by the state regarding the operations and authority of the market. For example, promoting labour market flexibility, including the ease with which employees could be hired and fired, has been a policy choice by many countries globally. Increasing feelings of subjective insecurity are important to understand because, in a recession, more people are relying on precarious labour to survive (Harrison, Reference Harrison2013: 105–06). In relation to this:

Similarly, the presence or absence of systems that afford extensive protection to individuals, through redistribution of wealth and/or protection of income in times of economic adversity, is highly relevant to resilience. This is supported by Hossain et al. (Reference Hossain, Byrne, Campbell, Harrison, McKinley and Shah2011) who examined the changes in the living conditions of people in County Tyrone, Newhaven and Oldham in the UK and found that the potential negative effects of the deep recession of recent years had been counteracted somewhat by the welfare system which protected the people in these areas.

Overall, then, a more protective labour market or welfare system may institute a system of socialised resilience while more liberal varieties can expose the working classes as a whole to the risks of severe downturns and leave them to deal with hardship individually.

Amongst the many structural factors, one may also consider rural–urban differences. Resilience is likely to have a different character in rural as opposed to urban areas as rural residents may have reduced access to jobs, community organisations, social networks and so on for a number of reasons, thus leaving them socially excluded. Lack of infrastructure may compound what is potentially a poor employment situation (Philip and Shucksmith, Reference Philip and Shucksmith2003). Room (Reference Room and Room1995) summarises the issue of rural resilience and poverty succinctly: ‘Deprivation is caused not only by a lack of personal resources but also by insufficient or unsatisfactory community facilities’ (Room, Reference Room and Room1995: 238). Those living in urban areas are much more likely to be able to draw upon such resources, whereas those in rural areas must show that they are deserving (Sherman, Reference Sherman2006: 893; Milbourne, Reference Milbourne2004: 564; Cloke et al., Reference Cloke, Goodwin, Milbourne and Thomas1995: 353), be able to exploit networks sufficiently (Philip and Shucksmith, Reference Philip and Shucksmith2003: 476) and in some ways be much more creative with their existing resources in order to overcome the ‘dominant conservative ideologies’ (Philip and Shucksmith, Reference Philip and Shucksmith2003: 471) associated with the idyllised nature of rural living in the UK (Millbourne, Reference Milbourne2004: 569).

Moreover, resilience for men may look entirely different to resilience for women (MacLeavy, Reference MacLeavy2011). In rural areas, traditional stereotypes have a particularly strong hold, with employment opportunities to be given first to men (Philip and Shucksmith, Reference Philip and Shucksmith2003). This may affect the working composition of the household, as ‘[c]oping strategies that may be commonly used in the urban setting and seem financially rational, can be socially and economically irrational in the rural setting’ (Sherman, Reference Sherman2006: 892).

One can go on to extend the examples of structural influences on resilience to cultural practices, educational provision, class background and so on. In summary, while a ‘resilient’ reaction to economic hardship may be an individual response, the conditions that give rise to this reaction cannot be entirely separated from the structural forces that shape them in many ways.

Path dependence and resilience

Acknowledging the importance of structure to understanding resilience, as argued in the previous subsection, highlights the fact that ‘crises are historical, social and political-economic products’ (Ribot, Reference Ribot2014: 671). As such, another relevant framework for examining the social conditions of resilience is ‘path dependence’. This approach developed in economics and was gradually adopted by other disciplines, including business studies, sociology and politics. Liebowitz and Margolis (Reference Liebowitz and Margolis1995) explain it as ‘a dynamic process [that] exhibits sensitive dependence to initial conditions’. The widespread use of the term to describe a range of circumstances popularised it as a way to emphasise that ‘history’ or ‘initial conditions’ matter. Studies in the literature highlight a number of tendencies giving rise to path dependence. One of these is ‘reactive sequences’ which describe how an initial or contingent condition generates a series of other outcomes, eventually resulting in a particular path of development. The other is called ‘self-reinforcing sequences’, a process in which an initial condition or a contingent development generates a reinforcement mechanism for the emergence of a particular path with particular outcomes, positive or negative (Mahoney, Reference Mahoney2000).

Path dependence was used to explain a wide range of phenomena such as why certain technologies dominate others despite their inefficiency (David Reference David1985); why industries agglomerate in certain regions as in the case of the manufacturing belt in the US (Krugman Reference Krugman1991); why business clusters emerge in areas such as Silicon Valley rather than Route 128 (Kenney and von Burg, Reference Kenney and von Burg1999); why capitalist development first took place in Europe not in China (Mahoney, 1999); why different regions show disparate economic performance (Martin, Reference Martin2006); why long-term or structural unemployment exists; and why certain organisational routines may persist (Schreyögg and Sydow, Reference Schreyögg and Sydow2011).

However, it is important to note that path dependency is a contested concept, and is certainly not the only option available in policy studies (for a discussion of the advantages and disadvantages of using path dependency in policy studies, see Kay, Reference Kay2005). Thelen (Reference Thelen, Mahoney and Rueschemeyer2003: 209) argues that the concept ‘tends to encourage a rather strict separation of the issues of institutional innovation and institutional reproduction’, which is premised on ‘a punctuated equilibrium model that emphasises moments of “openness” and rapid innovation followed by long periods of institutional stasis or “lock in”’. Yet, Thelen continues, this is not always the case. Some institutions, for example, show remarkable staying power even in the face of significant shocks. Others may continue to look the same, but undergo extensive change over a long period of time (Thelen, Reference Thelen, Mahoney and Rueschemeyer2003: 210). This being said, path dependence is not the dominance of one historical trajectory in the absence of alternatives. It should instead be understood as a (sometimes ongoing) restriction of choice rather than an absence of choice (North, Reference North1990: 98; Kay, Reference Kay2005: 553). Based on this, we argue that, for understanding the development of the concept of resilience to explain how households cope with hardship, path dependency is useful in explaining the influence of the social background of actors and the development of social support systems.

The relevance of path dependency in explaining resilience has already been discussed in economic geography. For example, Martin and Sunley (Reference Martin and Sunley2015: 34) view regional economic resilience as a:

In similar ways, resilience in times of economic hardship at the individual or household level is influenced by the general history of social, familial and cultural development, such as the wealth, health and knowledge/education of parents, close relatives and previous generations and the choices they made in their lives, as well as by the specific histories of regions and social groups. These in turn affect the inherited resources and abilities of current generations and their resilience as well as their need for developing it further.

We use a broader view of path dependence to emphasise the importance of the historical trajectory for the analysis of resilience against economic adversities. This means that strict assumptions used by some scholars have been ignored here. For example, a considerable proportion of the studies view path dependence as triggered by a historical ‘accident’. For the purposes of this paper, contingent events are not necessary conditions for the explanation of path dependency. This is crucial because, while past developments do influence the coping strategies of families in times of hardship, these are not always coincidental. Consider, for example, the role of wealth in resilience. Being rich or poor is hardly ever arbitrary or accidental. The socio-economic status of families has historical trails and is often built upon the wealth and education of previous generations. Even within low-income groups, income differences as well as social stratification may make a considerable difference to the ‘degrees of resilience’ they would demonstrate. Hence, one may in general find a high level of resilience in some communities or households precisely because they have the resources to develop resilience in the first place. On the other hand, the beginning of a harmful habit such as drug abuse may be considered as a contingent event creating a particular path dependency with negative consequences for the resilience of the person/people involved and their family. Likewise, a scholarship enabling someone to gain a university degree and enhancing her/his resilience could be viewed as an ‘accidental event’ but in another system where there is universal access to such opportunities it could just be a historical event.

Furthermore, path dependence can be studied in relation to the consequences of ‘small initial changes, decisions or random shocks’ that change the direction of developments (Page, Reference Page2006). While these are important in our context, cumulative developments also matter. Consider, for example, the role of migration as a strategy against economic hardship. Enhanced resilience of current generations may be an outcome of the migration of their parents from areas with limited prospects for education and employment. Quite often, first generation migrants engage in a multitude of struggles to make a living in a new country. The cumulative positive changes through generations may eliminate these difficulties for their children and grandchildren.

Finally, the ‘lock-in’ property is presented as an essential feature of path dependence in that it is considered to create a tendency for its subjects to follow a particular trajectory. This may be a plausible prediction for some cases such as business performance due to firms’ purposive functioning, say, for profit maximization. For household resilience in economic hardship however, this property of path dependence is likely to have some degree of flexibility. This is due to multidimensional influences and complexities in the determination of individuals’ life paths. While particular cumulative or contingent developments may shape the path within which different forms of resilience may be located, these may be modified or radically changed in the process due to factors internal or external to the households. For example, scholars such as Martin and Sunley (Reference Martin, Sunley, Boschma and Martin2012) explicitly discuss the forces that may lock out the subjects of study of an existing path. The authors do this by taking various considerations into account such as external shocks, or incentives for change offered by ‘enabling new paths’ in comparison to ‘constraining existing paths’. In fact, as Mahoney (Reference Mahoney2000) argues, the theory of path dependence is often conjured to explain deviant developments (e.g., why the industrial revolution took place in England not elsewhere). Deviation from long-term and persistent patterns may open up new paths and hence break the locking-in properties of existing manifestations of path dependence. A longitudinal study by Emmy Werner (Reference Werner1995) and her colleagues in Kauai (Hawaii) provides compelling evidence for lock-outs from path dependence. The study follows 698 individuals born in 1955, from the prenatal period to the age of thirty-two, all with high-risk backgrounds involving chronic poverty, family break-ups and parental psychopathology. In the majority of these cases (two-thirds of all participants), the lock-in property of path dependence works, i.e. the off-springs remain in chronic poverty and at high risk. However, the remaining one-third, despite being subject to poverty and trauma, develops a capacity to break away from the predicted life path and show sustained competence and recovery from trauma. The study identified various factors that may help individuals to break away from expected trajectories and classified them as those within the individual (such as a positive self-concept, ability to ask for help, special interests and hobbies), the family (such as the ability to recruit surrogate parents, reliable support from carers) and the community (for example the presence of role models, and programmes that give second chances to individuals such as adult education schemes).

Overall, social and familial history shapes the environment within which individuals lead their lives as well as their development. Hence, it has consequences for the resilience of individuals and households despite some possibility for deviance. The exploration of the causes of deviance from path dependence is the subject of another paper. In this paper, we continue with a discussion of the scientific hazards that may result from treating resilience as an individual attribute that could be replicated while overlooking the social conditions within which it emerges.

The problem of identification

Definitions of resilience in the context of economic hardship mentioned earlier highlight the abilities, skills and attitudes of individuals to deal with unemployment, low income or lack of income, debt, housing problems, transport, etc. The considerations in Section 3 raise issues with such individually focused interpretations of resilience. One of these could be termed as the ‘identification problem’. This is the possibility that what we identify as resilience in individuals or households may just be a manifestation of structural factors or path dependence reflected in their behaviour. This problem is evident in the questions raised by Burchardt and Huerta (Reference Burchardt and Huerta2009): Is resilience an attribute of an individual or a product of interaction with his/her environment? Is it a distinct attribute or does it merely reflect the absence of an unobservable risk factor? More questions can be added. If social and structural influences are also at play, how do we disentangle the individual from the social? Are social scientists well-equipped to analyse unique individual characteristics such as mental strength, positive attitude, etc.?

The implication of all this is that what may appear as lack of resilience in times of economic hardship in one household compared to another may reflect a culmination of social exclusion, highly restrictive institutional support, various social barriers and so on. Examples can be drawn from new generation migrant households that experience language barriers, a lack of social connections, inadequate knowledge of rights and entitlements in a new setting and limitations in access to welfare provision. All these factors are likely to constrain their chances of substantial recovery in times of economic hardship. A different set of disadvantages exists, for instance, for households with longer term health issues, such as severe disability or mental health problems. Their chances of ‘beating the odds’ against economic adversity through individual action is likely to be circumscribed by the disadvantages they confront on a daily basis. How should resilience or a lack of it be judged in households like this? Moreover, there may be cases where a display of resilience in the form of speedy recovery from unemployment and poverty is a manifestation of a middle-class ability to navigate through the system. For example, Mullin and Arce (Reference Mullin and Arce2008) found that in their study area in Massachusetts, the most resilient families knew how to access support structures outside the family unit, highlighting the role of knowledge and existence of supportive institutions.

In a nutshell, ignoring social conditions of resilience may give rise to an identification problem. For example, when the vulnerability or resilience of households may just be reflecting the differences in the ways the social system works for individuals or the history of their development rather than individual traits like mental strength, resourcefulness, adaptability and flexibility. The methodological difficulties of measuring resilience and distinguishing its causes from its consequences imply that making a comparative assessment of individuals in isolation from social structure and history is very difficult indeed. Given this difficulty, there are serious implications for misclassifying individuals. Any conceptualisation of resilience necessarily implies the existence of its opposite (i.e. defining resilience would reveal what is non-resilience). Hence, a misidentification is bound to stigmatise people who by design or default fall into the ‘non-resilient’ camp. What would be more unfortunate than inflicting such a stigma with the stamp of ‘science’?

Intermittence and continuity in resilience or the lack of it

Another potential problem for conceptualising resilience in isolation from structure, social and personal history would be the difficulty of explaining intermittence and continuities in resilience. The emerging findings from various research studies show that resilience or coping with hardship can be temporary as well as continuous. Background to hardship (e.g., recession) is important. In deindustrialised and deprived areas in the UK, for example, the recent recession is found to be less of a ‘shock’ and more of a process of compounding long-standing deprivation and vulnerability, especially considering that 40 per cent of unemployment is considered to be long-term (Harrison, Reference Harrison2013: 101). This sort of continuity in dealing with hardship highlights the importance of structural factors rather than showing resilience as a response to a one-off adversity.

More interestingly, other studies found that those who manage to come out of poverty are not immune from returning to poverty. For example, Seccombe (Reference Seccombe2002: 386) argued that many longitudinal studies indicate that most periods of poverty in the US are less than two years long but a single crisis such as a loss of work or a serious illness can re-expose people to poverty. Another study concluded that of those who managed to escape poverty, half would fall back into it within five years (Stevens, Reference Stevens1994). Philip and Shucksmith (Reference Philip and Shucksmith2003: 467) show that the tendency to fall back into poverty is higher in urban than rural areas. Canvin et al. (Reference Canvin, Marttila, Burstrom and Whitehead2009: 241) report similar observations about people living in poverty: ‘They move on, but they fall back, but it's not a circle, more of an upwards moving spiral.’ This sort of intermittence cannot be substantiated if resilience is conceptualised in isolation from structural factors as well as path dependency.

The crucial question is then whether resilience should be perceived as a permanent attribute of individuals, i.e. once resilient is always resilient. Intuition as well as the anecdotal evidence from the literature suggests otherwise. It would be difficult to explain such irregularities without reference to social conditions of resilience.

Neglecting the detrimental consequences of ‘resilience’

Resilience when considered as an individual attribute has positive connotations and may evoke admiration, although it may also hide the detrimental consequences of being resilient. It is important to recognise the costs at which the positive outcomes associated with resilience are achieved. For example, holding several jobs to make ends meet may be considered as a form of resilience, enabling individuals and families to survive through hardship, remain independent and avoid the stigma associated with being dependent on others. But it may also cause negative health effects (due to the physical and mental strain associated with long working hours), a deterioration in families’ work‒life balance, a sacrifice from essential leisure time, social activities and ultimately social isolation. The ability to survive on a very low income is also a reflection of resilience and yet again it is unlikely to be free of some undesirable consequences. Support from family and friends in times of hardship can be a contributing factor of resilience, but it may also mean transferring risk of hardship onto others.Footnote ⁵

Indeed, a number of scholars such as Harrison (Reference Harrison2013) and Batty and Cole (Reference Batty and Cole2010) recognise the costs of resilience and the risk of ‘burnout’. There is some research supporting the relevance of these concerns. For example, a report by Hossain et al. (Reference Hossain, Byrne, Campbell, Harrison, McKinley and Shah2011) found that in the recession, people made changes to their everyday routines. It highlighted that although people were cutting back on socialising, support from family, friends and neighbours mattered greatly. Coping with the rising cost of living meant less choice, and more time and effort went into the everyday work of feeding families on low incomes; for some it also meant poorer diets, colder homes and less physical mobility (Hossain et al., Reference Hossain, Byrne, Campbell, Harrison, McKinley and Shah2011: 6).

Ideological exploitation of the term ‘resilience’

The term resilience is prone to ideological exploitation when it is considered in isolation from its social conditions. Let us focus on a particular possibility to demonstrate the relevance of this argument. This is the possibility that the resilience discourse ‘glosses up the reality’ to legitimise the status quo or regressive changes in wider policy making. For example, the topic of this paper – the response of households to the ensuing deep recession following the 2008 crisis – can be viewed from different perspectives. We may provide a positive account of the survival strategies of the households and their agency within a framework of ‘resilience’ discourse. Alternatively, we may highlight the continuity in poverty (living with more of the same) and its intermittence (getting out of poverty only to fall back into it again), and emphasise insecurity of jobs/income and declining support systems. The difference between these two interpretations may not be a ‘thin line’ or simply different sides of the same coin. A positive account of the reality may just be a political tool for legitimising structural changes, for example in the welfare system, to the disadvantage of low-income groups who are classified as ‘resilient’. It may engender a discourse in which ideologically framed anticipation of ‘bouncing back’ justifies political negligence with regard to the negative consequences of such policy changes (e.g., higher poverty and destitution).

Indeed, as noted by Harrison (Reference Harrison2013: 99), the term resilience ‘depoliticizes and shifts responsibility for dealing with the crisis away from those in power. It also creates an expectation that people should “bounce back”’. Similarly, Hickman, et al. (Reference Hickman, Batty, Dayson and Muir2014: 6) argue that ‘an overemphasis on agency has potentially negative policy implications as it provides scope for policy makers to ‘blame’ low income households for their failure to escape poverty’. Intriguingly, the discourse on resilience has so far sidestepped ‘resistance’, involving collective responses to injustice, inequality and exploitation as a form of resilience because they do not conform to accepted notions of advancement that are usually centred on individualistic notions of entrepreneurship and ‘aspiration’ (e.g., Gregory, Reference Gregory2014; Hoggett, Reference Hoggett2001: 38). Developing Hickman et al.’s argument that an overemphasis on agency can result in blaming the poor for their situation, this overemphasis within the current resilience discourse effectively limits the ways in which one can acceptably escape poverty.

In the UK, the current political paradigm for dealing with the consequences of crisis could benefit from a resilience discourse with an agent-centric focus. For example, the Conservatives announced a series of public sector cuts (Smith, Reference Smith2010: 827), implemented them in coalition with the Liberal Democrats (Borges et al., Reference Borges, Clarke, Stewart, Sanders and Whiteley2013: 397) and argued that the impact of the cuts would be mitigated by an increase in civic-mindedness and voluntarism through the ‘Big Society’ (MacLeavy, Reference MacLeavy2011; Cameron, Reference Cameron2007, Reference Cameron2008, Reference Cameron2009). The Coalition Government utilised political and social nostalgia to legitimise these changes in an allusion to the period of enforced austerity of the 1940s, celebrated for the ‘blitz spirit’ that was cultivated at this time (MacLeavy, Reference MacLeavy2011: 357; see also Kynaston, Reference Kynaston2007; Calder, Reference Calder1992). Such imagery appeals to the mantra that ‘we're all in this together,’ legitimising further the welfare reforms and spending cuts, accelerating the trends towards greater individual responsibility (e.g., Jessop, Reference Jessop and Overbeek2003; MacLeavy, Reference MacLeavy2011: 359; see also Klein, Reference Klein2007). This legitimisation can also draw attention away from larger systemic issues, including the failings of the global economy and capitalism itself.

MacLeavy's analysis chimes with the findings of other scholars who argue that people are relying more heavily on informal networks of support. Indeed, the role of the third sector increased with the onset of the recession with government-led initiatives, such as Third Sector Action Plan (Alcock, Reference Alcock2012), that are justified through ideologically oriented narratives such as the Big Society.Footnote ⁶