Subscribe to Salvage and receive two print issues per year, plus digital access including audio and back issue PDFs
Class and Brexit
In the aftermath of the shock Brexit result, discussions on its causes and consequences have been frankly bizarre. So far, the have focused on bygone folk stories about sovereignty, migration, and globalisation, disconnected from any wider understanding of capitalism or history. There also seems to be no clear solution to the monumental fuck-up that has resulted from the absence of any clear plan for leaving the European Union which works, alongside our political system which is incapable of handling a depressingly English form of parliamentary populism. In this piece I want to try to untangle some of the causes and consequences of Brexit, and reflect on two key points; who actually voted for it and why; and what Europe actually is, and what its relationship to capitalism is.
These points are relatively hard to connect, but the purpose of this piece is to argue that Brexit was a consequence of a specific discourse meeting wider social change within capitalism. It therefore cannot be considered the working class reaction to globalisation, due to immigration and deindustrialisation, that for some reason everyone (both left and right) wants it to be. It was more constructed than that, and a consequence of class becoming culturally fractured after the decline of industrialisation, within an ideology which reinforces this by hiding economic change behind a false image of globalisation and of sovereignty as a cultural choice.
Who wanted to leave Europe According to Current Debates?
Analyses of Brexit have so far focused heavily on low income households, and on areas with a disproportionate leave vote. Although this has mainly concentrated on the disproportionate support of these areas rather than their aggregate contributions to the leave vote as a whole, it’s not entirely wrong to do so. Trying to push this deeper from who voted for Brexit to its causes, Matthew Goodwin extended his work on UKIP to highlight the role of wider educational factors, ethnicity, and low-income, specifically within underdeveloped and “left behind” areas. Goodwin’s work shows Brexit as being driven by what is seen as traditional English values and identity, as well as anti-immigrant sentiments and a disillusionment with mainstream politics. Arguing that it parallels UKIP voting preferences, and is generated by long term change within the UK’s socio-economic structures, rather than any immediate reaction to economic deprivation or immigration.
This is not a position I would broadly contend, however the problematic assumptions beneath it in regards to class need further work. This is especially so since both these implicit value and the “left behind” areas are explicitly linked to low skilled and older worker’s perspective in a deindustrialising economy. These are then treated as the working class as a whole due to their cultural outlook. This move then produces a view that it was the working class as such who voted for leave. This fails to provide any space for the increasing number of low-wage households in which the main income earner works in a high skill occupation; particularly in the income groups which traditionally dominated by the skilled working class and lower middle class, as well as the increasing capacity for lower skilled service workers and production workers to continually make up a significant proportion of all but the richest households.
Though a pattern clearly does still exist – with routine and semi-routine occupations still predominant in households with below average income, and lower professionals in the above average income households – the fact that lower professional occupations are now the largest group of the population marks a distinctive shift, one which undermines the idea that class can be associated solely with the industrial worker. By itself this doesn’t destroy the current literature’s understanding of the causes of the Brexit vote, but it does undermine the link between the causes it identifies and class.
Percentage of Occupation Group in Each Income Quartile, and in the Population as a Whole in 2014 Data
The roots of this lies in the understanding of class as a hangover from post-war identities, in which there was a stark divide between professional middle class and working class occupations, and the consequences of this for average income. In effect, the mainstream view sees the working class as literally vanishing with the rise of the service economy. To some extent this should be expected, as the mainstream literature draws on a fundamentally Weberian theory of class. In doing so its core focus is on the sociological construction of specific economic and political values deriving from occupationally divided communities. This doesn’t disconnect an analysis of class from the actual effect it has on an individual’s location in an economy, but neither does it define a class by its economic location. Instead, economic location alongside the resource, skills, values, sociability, and coherence of a class together determine its broad interests and discourse relative to other social actors by effectively binding it together as a cultural entity.
In itself this isn’t problematic, and the inclusion of wider economic characteristics is both welcome and well-founded. Occupations correlate well with a range of useful measures: their basis in varying skill defines inter-occupation mobility; their tendency to be defined by specific economic sectors means unemployment and wage stability is often shared; to some extent they do affect a cultural community, as entertainment and lifestyle habits are shared; voting patterns still exist though with education effecting values parallel and often in contrast to economic outlook; and they correlate well with health outcomes.
In grasping these acute reference points for sociological analysis, however, the economic history of capitalism is lost, and its changing structure and effects on income – as historically the key determinant of class as a political reference point – is hidden. In effect, this reduces class to a range of cultural discourses and values, formed from the specific economic circumstances they developed under but existing independently from the generative processes that produce economic difference. This presents distinct problems for the narratives that focus on Brexit as a working class revolt. These assume that culturally working class occupations match neatly to a scale of exploitation and privilege that governs internal differences within the working class, and therefore assume that there is a neat divide between the increasingly skilled metropolitan workers and a peripheral working class in deindustrialised areas.
Beleaguered Marxist Attempt to Reinterpret the Causes for Leave
To determine who voted Brexit I want to focus directly on class as a derivative of wealth and income, discussing the role of education, geography, and community independently. This disaggregates the current model which assumes that specific combinations of these factors represent class, but which undermines income’s role. Economic location is key for a Marxist analysis, and is central for my reinterpretation of current literature on class voting’s role in Brexit, as it refocuses analysis on the role of economic location as a determinative cause in the construction of class positions. Highlighting that the notion of working class is produced by a specific relationship to the means of production in spite of the intrinsic sectoral, geographical, and skills based variation that exists within it, this will see identity as being formed from the ways these differences are understood within social discourse relative to these effects on economic location in a concrete moment of capitalist production. The working class is consequently a shared reference point of low-income and wealth, but with varied understandings of that position in the absence of a coherent mechanism for that to be experienced through.
For my analysis I will utilise the latest wave of the British Election Studies Internet Panel, which collected data just before the referendum, merged with geographical data from the 2011 census and area income data from the 2015 annual population and labour force survey sourced from NOMIS. The dependent variable simply asked people if they were going to vote leave or remain, removing those who didn’t know at the time. The dataset I am using is broadly similar to the final result, indicating a slight lead for leave but with error accounted for by remaining “don’t knows” and the inability to predetermine turnout.
Though the regressions I will present can only partially explain the vote as a whole, they can be utilised to compare the relevance of the included variables. Therefore, to avoid the Weberian model’s reductionism the independent variables I use will reflect four components of its broad categories which in turn can explain variations in income and wealth: sex, age and ethnicity as demographics which may alter income; education as a measure of skill; household income and home ownership status as indicators of income and wealth respectively; and, finally, the skill base where an individual lives measured by percentage of those with degrees and median yearly income, relative to the national median. Furthermore, following Goodwin’s focus on the varying impact of geography, I have also included several interaction variables which determine the impact of income and wealth relative to an area’s different income and educational characteristics.
Though this can provide us with information on peoples voting intentions, as well as their household income, key demographics, and details about their local area, it does not include a representative measure for occupation due to problems with the BES’s collection of this data. This is fine in so far as no other analysts have access to data which can capture this either and I will try to correct for it by utilising a second model which includes values derived from the left-right and liberal-authoritarian scales common in the literature, as well as a measure of anti-immigration sentiment which I will discuss in more depth below and analysed in reference to the literature’s understanding of their relationship to occupation.
My reference category is a 50 to 59 year old white British female with college education in the second income quartile, who is in the process of buying her home, and lives in an area which has the national median income as the majority categories within each variable, and national median number of graduates. Once values are introduced she will also have the nationally average positions.
Though I will be showing my regressions in full it’s worth noting that as they are logistic regressions, which measure the chance of selecting one answer over the other, they only indicate if a variable had a significant impact on the vote and in what direction, indicated by a * if its significant (0.05) and ** if it’s very significant (less than 0.001), and if they have a positive or negative effect, with a positive coefficients indicating an increased chance to vote leave, and a negative as an increased chance to vote remain. However, from this data alone the size of a variable’s impact isn’t readily visible. This will be calculated separately as odds which show the percentage change that the variable has on voting leave. It’s also worth noting that for the value and geography variables, their effect can be reversed. For example, the coefficient for the variable for a £1000 increase in area income relative to the national median is positive, however if an individual’s area is £1000 below the national median then that effect becomes negative though the scale of the changes effect is the same.
Looking at the significant variables in the first model, what is immediately apparent is the preponderance of skills relative to class, with no income variables having a significant effect either at the household or geographic scale. The exception to this is the role of social housing which increases the chance to vote leave, though interestingly it decreases the chance to vote leave if the social housing is in a high-income area. The role of skills is relatively straightforward, with lower education increasing the chance of an individual voting leave. In turn, although age and demographics do have an effect, with non-white British supporting remain, the effect of this seem one sided and shows little difference among those over the age of 30.
In the second model, we add a control for people’s cultural values in an attempt to also include the characteristics of specific occupational groups which have been identified in political science literature. The effects of introducing these are dramatic as they remove the significance of all our class and demographic variables bar age, this should be expected as each value position does accord with certain demographic and economic characteristics and although it doesn’t invalidate our first model, as the focus here is compares influences on voters rather than explaining the entire vote, it does provide key insights into how combinations of specific factors drove the vote.
In the established literature on class and social values, a relationship is drawn between lower incomes and low-skill occupations and left-wing politics alongside a simultaneous tendency for those in low-skill occupations to have authoritarian values. The immigration scale sits parallel to authoritarianism, however, as indicated by the linear regressions below (in which the scale of effects can be directly compared through standardised coefficients rather than relying on odds as they measure a scale). Ethnicity lowers anti-immigration sentiment in contrast to increasing authoritarian sentiment, and while authoritarianism becomes stronger with age the effect of age on anti-immigrant is the same among all individuals over 30. Lastly, anti-immigrant sentiment is strongest in households with traditionally lower middle class and upper working class incomes, with the poorest and richest more supportive of immigration even when we control for ethnicity.
Combined, these models therefore present a far more complex image of Brexit than the Weberian models, though this is largely due to the way it orders the evidence rather than by undermining their results tout court. In my models the core results have largely shown the relevance of skill, represented by education, rather than class as such.
Though it’s worth remembering that occupation has become increasingly disconnected from income, a relationship between skill and increased wages still exist due to their capacity to increase social mobility. These models consider this by controlling for age and gender, therefore showing that income has little to do with the likelihood of voting for Brexit and where it does play a role it tends to show the poorest and richest supporting remain. Again, on aggregate this doesn’t disprove the idea that low income households voted disproportionally for Brexit. Instead it highlights the deficiencies of the narratives that assume income and skill simply move in parallel to each other, or that low skill could be interpreted as low income.
What Actually Drove People to Vote Leave?
This section will present the actual scale of the model’s significant variables relative to the constant we described above, and outline what drove Brexit using the standard assumption that the actual scale of the effect can be generalised to the entire population. The constant on each graph is represented by the black line and I have included all the variables in the same category as my significant variables to contextualise it.
The odds from my 1st model, which focused on my core variables, show the effect of educations, races and effects age on voting leave. The most significant effect here is having zero qualifications, which has almost double the effect of having GCSEs. The role of social housing should not be underestimated either and increases the chance of voting leave more than having GCSEs. Lastly, the effect of geographic area is relatively small – changing the chance to vote by less than a single percentage point for every £1000 from the average. Though social housing does reverse this effect, it is also small and means only those in social housing in the richest areas of the UK would have less than a 50% chance to vote remain.
In contrast the odds from my 2nd model which included values show that although being left wing does increase the chance of voting remain, and being more authoritarian does increase leave, the effects are small and cancel each other out. If an individual was therefore more left wing due to having a low income, but also more authoritarian due to having lower education we can assume it wouldn’t have much effect overall. Views on immigration is the real standout for increasing the leave vote, which from my regressions above indicate the importance of skill, ethnicity and a link with the middle two income quartiles. Similarly, being under 25 age was also key for decreasing support for leave, and although not as strong as immigration, it alters an individual’s propensity to vote remain below 50% by itself.
To finish my analysis of the data I also want to present a final linear regression. This utilises an alternative dependent variable which combines an individual’s vote with their conviction of their position, with the graph below showing coefficients which can be directly compared, and therefore show the scale of each variables effect. The seven independent variables I have utilised here show what participants believed the effects of leaving Europe would be. This instantly seems starkly different from my analysis of the actual causes for individuals to vote leave, as a chance to improve Britain’s economic and political status as a state was key.
In the end then the data shows Brexit was not a class vote. Instead, it reflects the various social positions that specific skills engender and the values they generate, particularly in regards to immigration. Though this matches well with the discourse of what it means to be working class, as an industrial worker who is strongly patriotic despite believing in a left wing management of the economy, I instead see class as peripheral to this. Brexit wasn’t driven by the disadvantaged but it represented, instead, an antipathy towards immigration by older white voters in low skill sectors, including both the aspirational low skilled home-owners and those in social housing, and supported more by the upper working class and lower middle class than the poorest sections of society. However, in spite of this being the structural cause, Brexit represented for many the chance for the UK to ensure greater economic growth and prosperity. Although this isn’t a class vote, there is a certain economic character to this, and despite the preponderance of anti-immigration sentiment driving leave it wasn’t necessarily a racist vote either.
How Did Brexit Take on this Character?
In establishing this position an obvious problem arises; how did this confluence of anti-immigration sentiment among low skilled workers translate into a distinct position on UK sovereignty? Trying to grasp the problems surrounding this is key, as it frames what the reaction to Brexit should actually mean. The current literature, in its focus on anti-immigration sentiments, has provided a justification for the new two main political parties which have moved to match a perceived working class anti-immigrant populism with a combination of reactionary policies, polite racism, and erudite traditionalism, while continuing with neoliberal economic policies. In effect, this has generalised the concerns of a culturally defined working class, though without any clear aims to support the working class as they actually exist, especially its significant non-white and immigrant components.
Though this response does seem to match the concerns of leave voters in my data, the idea that it will solve the problems of low skilled workers is problematic. It’s supposed source of change is sovereignty – which is seen as the power for distinct states to hold absolute power over their own territory, set against the geopolitical landscapes of mutually recognised states. Regaining this sovereignty was seen as essential for creating a capacity to end free movement and reorganise the economy to match British concerns, but without presenting any clear program of economic change.
Problematically though, sovereignty solely exists as an agreed up on legal institution of the organised inter-state system which itself is premised on the interaction of states. Though social organisation does have a territorial aspect, the necessity for inter-territorial relations for economic reproduction (even before capitalism) has ensured that state power is limited by its place within the world as a whole. Though this is recognised within international relations, there has been an abject failure to define the concept of an international system in liberal and realist debates, except as interacting sovereign states potentially alongside transnational actors. This reifies the legal form that power takes within the state, rather than the functional role of subnational actors have in constructing specific states and the mechanisms which power effects and is utilised through.
This reduces capitalist economics and the class construction of the state to effectively organic components of individual states interests, only including them as the foundation of state preferences or, in some cases, actors which can gain a form of sovereignty themselves and sit alongside states. It also fundamentally ignores the fact that states are constructions of economic systems which are generated beyond national units. Furthermore, it also ignores the extent to which states have only ever been fundamentally dependent on the interaction of classes across borders as the core source of their capacity to maintain power across wider units – defining not only their constituents, but also their revenue and key target for the application of power.
This doesn’t mean the state should be ignored: as sources of political power states are unparalleled. The concentration of resources they can provide in a stable manner means they are the core provider of infrastructure and an irreproachable support structure for the investment and income essential to modern capitalism. Instead, clear national coherence at the state level is problematic. Although states do have some capacity to help support their national economy, the role of localised sites of production and the organisation of specific class relations to support them are what defines the hierarchy and structures of global capitalism. It is between these sites that a capitalist global market emerges, rather than some sort of ether in which states sit when they reach a certain level of development in relation to some transhistorical market. By focusing on these underlying processes, the state instead becomes a tool for inter-capitalist competition, which is especially pertinent within the inter-state system where the state’s political power can allow strict demarcations for acceptable investment.
Ignoring this reduces the role of the EU to essentially a gatekeeper to a purely national self-interest, and a product of a desire to get trade from national and foreign influence. But once sovereignty is demystified, the EU’s role changes from a strict limit on sovereign power to simply an extended forum for the state as a tool of capitalist competition between subnational actors to formalise their relationships. Consequently, European integration can be better understood as a reaction to the changing organisation of global production, rather than as a cause. In particular, helping manage the transition from the Keynesian era focus on national economies, competing over final commodities produced in a national site’s distinct sectors, to the division of production into low skill manufacturing and raw material components centred in the global South and high skill manufacturing and services in the global North.
This has produced aggregate gains for the global North, as it has become the centre of global demand and of protected and high value sectors and services, while the economies of the global South are forced to maintain low wages if they wish to remain competitive in sectors with relatively low entry levels. This has supported increasing living standards, as capital becomes more readily available for debt expansion, whilst the cost of consumer goods has fallen despite stagnating wages. But the economic restructuring this requires is not painless; low skilled workers who were previously employed in sectors which are now more competitive in low wage economies have lost their access to their key source of income growth, as remaining production roles are increasingly low-employed, high-capital, and often higher-skill. Particularly in the UK the core replacement has been the retail sector and its infrastructure, which, although growing, relies on low-profit and wholesale-led cost decreases and discounting pricing led high-turnover. This has provided few opportunities for income advancement, as it relies on the expansion of core demand particularly through debt for its turnover.
This process has occurred away from the EU as an institution, though it was certainly supported by it. In this regard, the EU was a tool which focused on protecting the region against excessive internal competition whilst supporting the region against global competition. The specific legal form of this has been central to its strength. However, what is key is that its function is not to produce restructuring as a shared process across all states, but to provide a coherent and robust tool for national economies to manage the changes that stem from this wider shift in capitalism, without really ending continued competition between them. For example, the organisation of the European Central Bank simultaneously supports regional financial stability and is prefaced on equalised voting rights rather than allowing larger European economies to dominate. However, rather than this undermining sovereignty, its core effect has been to simply provide a secondary avenue through which advantages could be negotiated, and crucially in a way which maintains the a secure forum for this to occur within and in which agreed upon gains can be maintained.
Therefore, although the EU mimics a federal system in which sovereignty becomes shared, it functions as little more than as a forum for uneven development across Europe to be stabilised, creating strict exit costs to ensure compliance. In return, this reinforces the differences between member states, as industrial power remains centred in the more advanced, Northern economies while ensuring less advanced Southern economies remain relatively dependent on supporting them by producing inputs and agricultural goods. In effect the debate on the EU focuses on phenomenon with which it has relatively little causal influence over, and the debate on the changing economic structure of capitalism was ignored in favour of a debate which assumed the EU’s legal capacity was the sole fulcrum around which change functioned. This ends up privileging the legalistic form that capitalist competition has taken, as production has expanded globally, rather than its stable function which remained primarily dependent on wider changes in capitalism and which can only be changed when the actual value of specific sectors and the class compositions that support them changes, not by changing its institutions.
The Debate That Never Was on a Vote That Meant Little
Trying to delineate why the EU has been framed in this way is difficult. There are empirically visible between British and European capitalism, particularly due to the disproportionate influence of finance generating a need to maintain the pound. This creates clear reasons for UK capitalists to feel somewhat restricted by some of the potential rules that the EU supports and a logical desire to maintain a more arms-length relationship with the EU if they do not want to deal with the costs of circumventing legal obligations. Linked with this, immigration has become a potent reference point for popularising British political competition over these concerns, even if the UK does have a relatively weak anti-immigrant movement. Much has been done to associate the EU with free movement as a clear legal obligation, and in doing so interweaving wider opposition to specific directions of EU reform with a relatively broad populist concern. In the absence of a clear narrative to support a discussion on capitalism, the focus on sovereignty fitted well with concern across all workers in low skilled sectors to fixate on the EU as the source of an already established, though false, view that immigration caused the stagnation of low-skill sectors.
By failing to counter the focus on sovereignty, the debate on the EU therefore avoided the more serious ongoing concerns on the changing structures of capitalism and instead falsely promoted an answer which has no intrinsic effect on altering the structure of low skilled sectors, bar allowing the UK to formally leave a legal mechanism. This almost embarrassing feature of the discussion cuts to the core of the referendum, being generated at its starting point when the vote was called to head off right wing populism and to provide leverage in within EU negotiations. Consequently, as my data shows this allowed Brexit to occur as older white low skill workers grabbed an opportunity to alter the form through which changes in capitalism had been presented to them on the basis of an absurdly shallow debate on the actual EU’s role in their predicament or the limits its placed on solving it. This however, is not a class vote, the joint failure of income or wealth to remain significant once skills, demographics, geography, and (in my second model) values were controlled for Instead indicate that it was a cultural vote which although produced by economic circumstances didn’t reflect economic disadvantage.
In spite of this, I don’t want to under-emphasise low skilled worker’s tendency to be poorer. Instead I want to highlight that they are not necessarily the poorest section of the British working class. Instead they are simply the component of the working class which have the clearest cultural connection to the industrial working class, and consequently, due to the prevalence of Weberian theories of class, are therefore treated as the working class as a whole. Therefore, in contrast to most analysis of the vote my data instead shows that the class component is spurious, and although the vote was driven by clear disadvantages many people face in modern British capitalism disaggregating class and skill is essential to prevent the cultural conception of class being reified and setting in stone what the working classes influence in politics should look like, and what it requires to fix.
BELLOFIORE, R 2013. ‘Two or three things I know about her’: Europe in the global crisis and heterodox economics. Cambridge Journal of Economics, 37(3), pp.497-512.
BLOOM, N. DRACA, M & VAN REENEN, J 2016. Trade Induced Technical Change? The Impact of Chinese Imports on Innovation, IT and Productivity. The Review of Economic Studies, 83(1), pp.87-117.
BONEFELD, W 2000. The Spectre of Globalization: On the Form and Content of the World Market. In: W. BONEFELD & K. PSYCHOPEDIS (eds.) The Politics of Change: Globalization, Ideology and Critique. Hampshire: Palgrave. pp.31-68.
BUZAN, B & LITTLE, R 1994. The Idea of “International System”: Theory Meets History. International Political Science Review, 15(3), pp.231-255.
CHAN, T W & GOLDTHORPE, J H 2007. Class and Status: The Conceptual Distinction and its Empirical Relevance. American Sociological Review, 72, pp.512-532.
CLARK, G L 2015. The geography of the European Central Bank: form, functions and legitimacy. Journal of Economic Geography, 15(5), pp.855-881.
DIXON, A D 2014. The New Geography of Capitalism: Firms, Finance, and Society. Oxford: Oxford University Press.
FIELDHOUSE, E. GREEN, E. EVANS, G. SCHMITT, H. VAN DER EIJK, C. MELLON, J & PROSSER, C 2016. British Election Study Internet Panel Waves 1-8.
GODELIER, M 2011. The Mental and the Material. London: Verso.
GOLDTHORPE, J H & MCKNIGHT, A 2004. The Economic Basis of Social Class. CASEpaper, 80. London: Centre for Analysis of Social Exclusion.
GOODWIN, M & HEATH, O 2016a. Brexit vote explained: poverty, low skills and lack of opportunities. Joseph Rowntree Foundation. [online] Available at: <https://www.jrf.org.uk/report/brexit-vote-explained-poverty-low-skills-and-lack-opportunities>.
GOODWIN, M J & HEATH, O 2016b. The 2016 Referendum, Brexit and the Left Behind: An Aggregate-level Analysis of the Result. The Political Quarterly.
HALDON, J 1993. The State and the Tributary Mode of Production. London: Verso.
HEINRICH, M 2004. An Introduction to the Three Volumes of Karl Marx’s Capital. New York: Monthly Review Press.
HEINTZ, J 2005. Low-wage manufacturing and global commodity chains: a model in the unequal exchange tradition. Cambridge Journal of Economics, 30(4), pp.507-520.
LATOUR, B 2016. Onus Orbis Terrarum: About a Possible Shift in the Definition of Sovereignty. Millennium – Journal of International Studies, 44(3), pp.305-320.
MILBERG, W 2008. Shifting sources and uses of profits: sustaining US financialization with global value chains. Economy and Society, 37(3), pp.420-451.
NATCEN SOCIAL RESEARCH 2014. British Social Attitudes Survey 2012. UK Data Service. DOI:10.5255/UKDA-SN-7476-1.
NAYYAR, D 2005. Globalisation, history and development: a tale of two centuries. Cambridge Journal of Economics, 30(1), pp.137-159.
NOWELL, G 2009. Hilferding’s Finance Capital versus Wal-Mart World: Disaggregating the Dollar’s Hegemony. Perspectives on Global Development and Technology, 8(2-3), pp.315-346.
OESCH, D 2008. The Changing Shape of Class Voting. European Societies, 10(3), pp.329-355.
POWELL, R 1994. Anarchy in international relations theory: the neorealist-neoliberal debate. International Organization, 48(2), pp.313-344.
ROSENBERG, J 2006. Why is There No International Historical Sociology? European Journal of International Relations, 12(3), pp.307-340.
SCOTT, A J 2006. Geography and Economy. Oxford: Oxford University Press.
SERFATI, C 2016. EU Integration as Uneven and Combined Development. In: Analytical Gains of Geopolitical Economy (30. pp.255-294.
SHATZ, H J & VENABLES, A J 2000. The Geography of International Investment. In: G. L. CLARK, M. P. FELDMAN & M. S. GERTLER (eds.) The Oxford Handbook of Economic Geography. Oxford: Oxford University Press. pp.125-145.
SMITH, J 2016. Imperialism in the Twenty-First Century: The Globalization of Production, Super-Exploitation, and the Crisis of Capitalism. New York: Monthly Review Press.
VAN DER WAAL, J. ACHTERBERG, P & HOUTMAN, D 2007. Class Is Not Dead It Has Been Buried Alive: Class Voting and Cultural Voting in Postwar Western Societies (1956 1990). Politics & Society, 35(3), pp.403-426.