Studies in Higher Education

ISSN: 0307-5079 (Print) 1470-174X (Online) Journal homepage: https://www.tandfonline.com/loi/cshe20

Limitations of human capital theory

Simon Marginson

To cite this article: Simon Marginson (2019) Limitations of human capital theory, Studies in
Higher Education, 44:2, 287-301, DOI: 10.1080/03075079.2017.1359823
To link to this article: https://doi.org/10.1080/03075079.2017.1359823

Published online: 08 Aug 2017.

Submit your article to this journal

Article views: 16935

View related articles

View Crossmark data

Citing articles: 81 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=cshe20
STUDIES IN HIGHER EDUCATION
2019, VOL. 44, NO. 2, 287–301
https://doi.org/10.1080/03075079.2017.1359823

Limitations of human capital theory*

Simon Marginson
UCL Institute of Education, University College London, London, UK; Melbourne Centre for the Study of Higher
Education, University of Melbourne, Melbourne, Australia

ABSTRACT KEYWORDS
Human capital theory assumes that education determines the marginal Human capital theory;
productivity of labour and this determines earnings. Since the 1960s, it educational policy; critical
has dominated the economics, and policy and public understanding, of realism; philosophy of
relations between education and work. It has become widely assumed education; sociology of
education
that intellectual formation constitutes a mode of economic capital,
higher education is preparation for work, and primarily education (not
social background) determines graduate outcomes. However, human
capital theory fails the test of realism, due to weaknesses of method: use
of a single theoretical lens and closed system modelling, inappropriate
application of mathematical tools, and multi-variate analysis of
interdependent variables. Human capital theory imposes a single linear
pathway on the complex passage between heterogeneous education
and work. It cannot explain how education augments productivity, or
why salaries have become more unequal, or the role of status. These
limitations are discussed with reference to research on social
stratification, work, earnings and education.

Human capital theory

Since its modern beginnings at the turn of the 1960s (Mincer 1958; Schultz 1959, 1960, 1961) and
fuller development by Gary Becker (1964), human capital theory has constituted a fecund research
programme in the economics of education, associated with many thousands of empirical studies.
In the foundational narrative of human capital theory, education drives the marginal productivity
of labour and marginal productivity drives earnings. Correspondingly, the value of investment in edu-
cation is defined by the lifetime earnings of educated labour. Education, work, productivity and earn-
ings are seen in linear continuum. When educated students acquire the embodied productivity (the
portable human capital) used by employers, graduate earnings follow. In the pure and original form
of the idea, higher education more or less automatically triggers private enrichment, career success
and national economic growth. The claim about the contribution to economic growth made by
aggregated investment in education as human capital, first asserted baldly by E.F. Denison (1962)
and developed with more econometric sophistication in endogenous growth theory’s account of
education and knowledge in technologically driven development (Romer 1990), is now a common
policy assumption (Psacharopoulos 1994; Keeley 2007). In some countries, data on private rates of
return to graduates are used to regulate the private/public split in education financing, between
fee payments and government subsidies (Chapman, Higgins, and Stiglitz 2014), though there is
debate about the respective roles of public and private investment.

CONTACT Simon Marginson [email protected], [email protected] UCL Institute of Education, University College
London, 20 Bedford Way, London WC1H 0AL, UK
*This is an extended version of the keynote address to the annual conference of the Society for Research into Higher Education,
Newport, Wales, UK, 9 December 2015. Thank you to Helen Perkins.
© 2017 Society for Research into Higher Education
288 S. MARGINSON

The dominance of human capital theory in the economics of education is matched by its authority
in the public and policy domains. Nevertheless, as discussed immediately below, there is a gap
between the world imagined in the theory and the real economic and social world in which it is
applied, and this gap may be growing. This article will argue that human capital theory’s failure to
meet the test of realism derives not from lack of sophistication – since its foundation, there have
been various innovations designed to increase its empirical purchase and utility – but from its
meta-method. The limitations in meta-method, which are discussed below, have led in turn to a
flawed and narrow understanding of education/work – and the first mover authority of human
capital theory has stymied alternative conceptions.

The theory and its policy contexts

Founding modern human capital theory was the product of a particular historical moment that
favoured its genesis and spread. It evolved amid the building of mass higher education in the US
(Kerr 2001). The theory provided a rationale for the government-sponsored expansion of higher edu-
cation, while also promising to efficiently regulate the pace and cost of expansion on the basis of the
measured economic returns to graduates. The main ideas were propagated internationally by the
United Nations Educational, Social and Cultural Organisation (UNESCO 1968) and later the Organis-
ation for Economic Cooperation and Development (OECD). They became general to economic
policy at the same time as another policy discourse, social rather than economic, that of equality
of opportunity through education. In the conjuncture, the two policy ideas were necessary to each
other (Marginson 2016a). The policy goal of equality of opportunity promised to optimise the econ-
omics of education by ensuring that all available productive talent would become educated. Human
capital theory provided an economic justification for investment in expanding educational opportu-
nity. In Capital in the Twenty-First Century (2014), Thomas Piketty shows that between the 1950s and
the 1970s, conditions in the US were unusually favourable for the reception of these ideas (Marginson
2016a, 3–4). The potential for upward social mobility via higher education was high. Professional jobs
were growing rapidly; income from inherited capital was at historically low levels; and to an extent
not seen before or since, income from work was the main source of wealth (Piketty 2014, 241).
Amid excess demand for educated labour, all graduates could obtain good jobs. This appeared to
confirm human capital theory in practice and also underpinned contemporary optimism about the
potential of higher education to create a fairer and more efficient society, in which educated merit
and hard work would determine success, rather than prior family position. Piketty notes that
Becker’s (1964) mathematisation of human capital theory is permeated by the belief that all forms
of capital other than human capital (that is, financial, social and cultural capital) have lost their deter-
mining importance (Piketty 2014, 385). The 1960s’ expansion of opportunity and social mobility
enabled human capital economists to imagine that the theory was not just necessary in explaining
the relationship between higher education and work, it was sufficient.
Half a century later, the context is different from that in which Becker published Human Capital. In
the 55 countries in which the higher education system includes 50% or more of the youth cohort
(Marginson 2016b), in variant and often fluctuating economies, not all graduates enter professional
jobs; while income inequality has dramatically increased in the US (Saez 2013; Piketty 2014, 265),
inheritance is more potent (393) and income from capital now outweighs income from labour as a
source of wealth (402). The power of family income, social and cultural capital in determining
access to both elite higher education and elite professional employment is attested repeatedly in
research (e.g. Soares 2007; Rivera 2015; Social Mobility Commission 2016). American social mobility
is at a lower ebb than in the 1960s/1970s (Corak 2012; Stiglitz 2013). Regardless, human capital theory
continues to shape understandings of relations between higher education and work. One reason is
that although equality of opportunity has faltered in societies becoming more unequal, the idea of
merit as learned and portable ability retains legitimating power. The notion of human capital, floating
free of other forms of capital, implies that those with social advantages succeed not because of their
 STUDIES IN HIGHER EDUCATION 289

birth and connections, but because of their abilities and powers of application (Hennessy 2014). In a
‘hypermeritocratic’ parody of the original human capital idea (Piketty 2014, 264–265), the exceptional
salaries of American super-managers are legitimated by their prior selection into leading universities
(Rivera 2015) and within performance pay regimes by alleged super-productivity in the workplace
(Hanley 2011). In this curious backhand way, the core propositions at the heart of human capital
theory have ‘meritified’ self-reproducing privilege, though the normative commitment of economists
is often to equality of opportunity.
Many human capital economists have grappled with these problems. Since 1960s, the human
capital research programme (Blaug 1976) has taken on greater complexity and the foundational nar-
rative has been considerably supplemented. Becker’s later work (e.g. Becker and Murphy 2003) seeks
to account for the influence of the social setting on behaviours and choices, in the process extending
his 1964 conception of productivity-generating skills and knowledge beyond the education system.
Other scholars draw attention to the influence of non-educational factors in earnings. For example,
Glomm and Ravikumar (1992) combine choice-based investment in human capital with the capacity
of parents to pass on endowments to their children, which they define as another component of
human capital. This provides one possible reconciliation of human capital theory with unequal econ-
omic and social outcomes, suggesting that education may generate absolute economic gains
through productivity advance while leaving relative benefits unchanged. Delaney, Harmon, and
Redmond (2011) investigate parental education as causal in relation to graduate earnings expec-
tations. Britton et al. (2016) study UK graduates with 10 years in the labour market, investigating
the effects of variations in socio-economic background, gender, institution attended and field of
study. They note high dispersion in graduate outcomes (53–54), and find that ‘graduates’ family back-
ground – specifically whether they come from a lower or higher income household – continues to
influence graduate’s earnings long after graduation’ (55). Graduates from higher income households
earn at least 10% more at the median than graduates from low-income households after factoring
out other student characteristics, institution attended and field of study (55). High-income origins
protect graduates against low earnings and increase their prospects of very high earnings (56). In con-
sidering why these data modify expectations created by human capital theory, the authors suggest
‘students from wealthier families … might have access to financial, social and cultural capital’,
though they also pose a possible alternative, that students from wealthier families may have
greater ability (56).
A feature of human capital economics is that despite the complexities and qualifications intro-
duced by theorists, and notwithstanding marked changes and variations in real-world contexts,
the core 1960s’ propositions of the theory remain intact, at least as a partial truth. Crucially, they
still function as a form of default explanation. The other factors, which are often seen by economists
as social rather than economic, seem to be pasted on, framed as potential modifiers rather than the
basis of an alternative explanation of higher education, work, incomes, income distribution and social
outcomes. The human capital economist asks ‘why doesn’t human capital theory work as it should,
and what are the additional elements and modifications needed to make it work as it should?’ rather
than ‘what is a new and better explanation of the relationship between education and work?’ More-
over, the economists’ qualifications often drop away when human capital calculations are used
instrumentally in policy; and the complexification of the theory scarcely impacts the larger public dis-
cussion about education and work, in which it is the original and default explanation – rather than the
qualifications, complexities and contextual issues – that continues to have shaping effects.
This is because human capital theory has become influential in policy and public thinking not
simply as an applied research programme and a set of econometric techniques deployed one way
or another, but in the form of a widely understood metaphor for relations between work and edu-
cation, that is grounded in the foundational narrative of a linear continuum between education,
work, productivity and earnings. It could be argued that this metaphor has become so widespread
as to comprise part of the modern ‘social imaginary’ (Taylor 2007). The popularisation of human-
capital-as-metaphor helps to explain the ‘pervasive belief in the power of degrees to both allocate
290 S. MARGINSON

individuals in the labour market and to serve as job requirements throughout the occupational struc-
ture’ (Baker 2011, 62). That intellectual formation constitutes a mode of economic capital (Hodgson
2014); that in the first instance higher education can be primarily understood as preparation for work
and career; and that education itself, not family income or cultural attributes or social networks, is the
starting point for an explanation of career outcomes and earnings: all these notions have (arguably,
unduly) elevated education as a social and economic arbiter. For example, in the UK and Australia,
higher education institutions (HEIs) and their disciplines are held to account by government and
public on the basis of graduate earnings and/or employment rates in the early years of work, regard-
less of other elements that affect employment and earnings. Correspondingly, the idea of education
as self-investment in one’s own capital positions graduates (or their portable human capital) as
responsible for individual success/failure and weakens the obligation of government to steer a
more equal income distribution.
By no means all economists would endorse those conclusions. Few would agree with the simpli-
fied version of the relations between higher education and work current in much policy rhetoric and
public debate. Most professors of economics would firmly reject notions that the value of education
can be reduced to its measured effects in earnings or jobs. However, such positions are consistent
with the intellectual strategy of human capital theory, which is to protect the original ideas by ren-
dering them more complex and nuanced rather than call them into question. The founding paradigm
has not been declared obsolete – and like most social scientists, human capital economists are not
known for talking down their core idea. In this manner, human capital theory (buttressed by human-
capital-as-metaphor) tends to block from view alternative ideas, theories and measures about
relations between education and work.

Critiques of human capital theory

Since its inception, human capital theory has been subject to repeated and often devastating cri-
tiques. Few scholars from outside mainstream economics with a close research knowledge of edu-
cation have endorsed human capital theory. Many scholars in the political economy of education
and labour have challenged the core narrative, from Bowles and Gintis (1976) to Spring (2015). On
the economics/sociology border, screening theory sees higher education not as a site of self-invest-
ment in cognitive formation that delivers economic returns, but a system for signalling a competitive
position that delivers economic returns – an alternative narrative to human capital theory using much
the same evidence (e.g. the early study by Berg 1971). Sociologists including Trow (1973), Collins
(1979), Teichler (2009) and Baker (2011) provide very different accounts of work and education. In
his work on social reproduction in education, Pierre Bourdieu (1984, 1988) highlights positional com-
petition and status, which human capital theory cannot encompass, and introduces family cultural
capital and social capital networks as central to the explanation, rather than dispensable add-ons.
The OECD (2014a) treats social background effects on vocational outcomes, and human capital
effects, as intermeshed, without giving priority to one over the other. A large literature explains
socially differentiated educational outcomes more as a function of prior inequalities and institutional
stratification in education, than individual choices about self-investment in education, pointing to
ways in which social inequalities affect aspirations (e.g. Hoxby and Avery 2013) and are reproductive
(Boliver 2011, 2013). In The Global Auction (2012), Philip Brown, Hugh Lauder and David Ashton
describe declining private returns and dispersion of graduate outcomes amid unequal and exploita-
tive societies, again a different world from that suggested by human capital theory.
However, most critical scholars are at cross-purposes with those they criticise. After all, any theory
can be criticised from the standpoint of a different theory; and any discipline can be interrogated
from the perspective of another in several ways. But sociological critiques have limited potential
to persuade economists or change the minds of economic policy-makers for whom economics is
the master social science. Rather than posing an alternate theory or discipline as the basis of critique,
it is more fruitful to go the roots of human capital theory – to interrogate the default narrative in
 STUDIES IN HIGHER EDUCATION 291

relation to its own purpose, which is to provide a universal explanation of relations between edu-
cation and work. In this article, the basis of critique is historicisation, which tests a theory against
the empirical terrain it purports to explain. The emergent weakness of human capital theory is
lack of realism. Friedman (1953) argued that economic theory does not need to be realistic to
make viable predictions and secure normative effects. Yet, for many economists, and more policy-
makers, a key idea that lacks realism does have a problem. Lack of realism undermines the scope
of government to understand and to act.
The present article argues that human capital theory lacks realism in at least four areas. First,
human capital theory uses a closed analytical system and independent variables but neither external
effects nor co-dependence can be eliminated from the problems it addresses. Second, a linear theory
is applied to material that is non-homogeneous in space and time. Third, human capital theory unifies
two heterogeneous domains, education and work, as if they are a single domain. Fourth, it eliminates
other possible explanations of education/work relations, of which there are many. It is further argued
that these weaknesses at the base of human capital theory derive from the underlying meta-method
of its social science (Dow 1990), which blocks the possibility of realistic explanations. The problems of
meta-method, not unique to human capital theory, are: (1) the theory’s claim to a universal theoris-
ation based on a single lens, and its closed system modelling of social relations; and (2) the mis-appli-
cation of mathematical tools, and in particular, the use of multi-variate analysis of social relations in
contexts in which the variables are irretrievably interdependent. The article begins with discussion of
these problems of meta-method and then moves to the four points above.
The discussion of meta-method draws on critical realism (e.g. Sayer 2000; Bhaskar 2008) and het-
erodox economics, realist and historicised bodies of thought that work across multiple disciplines and
theorisations and use multiple methods. The limitations in human capital theory’s understanding of
relations between education and work are discussed with reference to selected research on social
stratification, work, earnings and higher education.

Problems of meta-method
Universal lens and closed system
Human capital theory operates as a single and universal lens. The use of the single exclusive lens rests
on the dualistic proposition that there is only one possible truth about social phenomena, and that
particular truth has absolute authority (Dow 1990, 2012). In this kind of social science, the researcher
applies a fixed theoretical framework and linked methodology to a succession of empirical obser-
vations in different sites. The theory is seen as universally applicable to all sites. Obversely, the
only phenomena that can be recognised in observation are those nominated in the template of
the theory. It is as if one objective of each act of research is to affirm the theoretical components
by identifying and codifying them. The weight of successive papers seems to ‘prove’ the master
theory, but it is a test that tends to guarantee its own result. The possibility that the master theory
is more applicable to some social sites than others is not considered. However, the succession of
similar narratives has diminishing returns, in that they are increasingly less likely to create new
knowledge.
Two lacunae follow from the use of a single exclusive lens. First, as suggested, observation is
stymied in sites where the single lens does not readily apply to the material. Second, other possible
explanations, arising from the use of the many other lenses, are obscured. Researchers using a single
lens might acknowledge limitations of their study but rarely question the capacity of the master
theory to address any possible problem.
For universal explanations to work, they need closed systems with limiting premises. However,
critical realism argues that social structures are always partly open, to other structures and agents,
and historical contingency (Sayer 2000). While a temporary partial closure is necessary in any research
and analysis, the problem arises when analytical closure is placed beyond interrogation and has the
292 S. MARGINSON

force of a fixed and permanent law, as with human capital theory. This creates conditions for fallacies.
Tony Lawson critiques neoclassical economics on the grounds that it imagines the economy as a
closed system operating by deductive logic. ‘Deductivism’ is ‘the thesis that closed systems are essen-
tial to social scientific explanation (whether the event regularities, correlations, uniformities, laws, etc.
are either a prior constructions or a posterior observations)’ (Lawson 2012, 3–4).
By deductivism I mean a type of explanation in which regularities of the form ‘whenever event x than event y (or
stochastic near equivalents) are a necessary condition. Such regularities are held to persist, and are often treated,
in effect, as laws, allowing the deductive generation of consequences, or predictions, when accompanied with the
specification of initial conditions. Systems in which such regularities occur are said to be closed … If mathematical
methods of the sort economists mostly fall back on are to be employed, closures are required (or presupposed).
(Lawson 2003, 5, emphasis in original)

If mathematical sets in economics are universally relevant, strict ‘event regularities’ must be ubi-
quitous in the real world. However, when deductivism is used in real-life contexts, ‘social event regu-
larities of the requisite kind are hard to come by’ (Lawson 2003, 13). The alternative is to imagine the
economy/education as a partly open system without strict ‘event regularities’, to acknowledge the
partial character of the truth about that system obtained through any one lens, and to open up
‘the possibility of a range of approaches’ (Dow 2012, 82). Theories ‘can vary according to changed
times and circumstances’ (Carabelli and Cedrini 2014, 44). This is also true of the policy applications
of theory. Hence, human capital theory is closer to realism under full employment than high unem-
ployment, and more explanatory of investment in financial management education than investment
in a music or drama programme with negative rates of return. If no single discipline, theory or meth-
odology has universal reach, by the same token, no one explanation excludes, cancels out or invali-
dates all other explanations. This means that in each research site and problem, it is necessary to
identify the appropriate theoretical lens or combine and match the appropriate lenses.

Problems of multivariate modelling

The high standing of mathematical modelling in much of social science reflects a society-wide belief
that mathematics is fundamental to science, a conviction (or ideology) that derives not just from the
elegant simplification permitted by mathematics, but also from the success of mathematical pre-
cision in many domains (Lawson 2012, 16). However, the subject matter of the ‘social disciplines’
is often inappropriate for mathematical treatment (Carabelli and Cedrini 2014, 31), especially when
complex, holistic, synthetic accounts are required. ‘The fundamental problem of modern economics
is that methods are repeatedly applied in conditions for which they are not appropriate’ (Lawson
2012, 1) – mathematical methods are often applied to phenomena they cannot adequately compre-
hend and problems they are not competent to solve. Mathematical methods have potential in
research on education and work, as auxiliary tools in studying relations and comparisons. They can
be used to map proportions and changes in bounded sub-systems. But in themselves, these
methods do not explain, they illustrate. Sayer (2000, 22) states: ‘Statistical explanations are not expla-
nations in terms of mechanisms at all, merely quantitative descriptions of formal (not substantial)
associations’.
One heterodox line of thought in economics rejects the main path taken by methods of mathe-
matisation and statistical modelling in human capital theory and parallel domains, particularly multi-
variate analyses that impose arbitrary definitions on indeterminate social variables in complex sites in
which many variables are at play. Multivariate statistical analyses use probabilistic methods to dis-
tinguish nominal degrees of causality for each one of a set of variables. However, Alfred Marshall
argued that when the subject matter becomes more complex, rather than devising ways of reducing
that complexity, the economist should diminish the use of abstract reasoning and mathematics (Mar-
shall 1898, 39). Marshall argued that the problem with much of the use of mathematics in economics
is that the econometrician ‘takes no technical responsibility for the material, and is often unaware
 STUDIES IN HIGHER EDUCATION 293

how inadequate the material is to bear the strains of his [sic] powerful machinery’ (Marshall [1920]
1961, 781). Similarly, John Maynard Keynes noted that mathematical reasoning was formally rigorous
yet hostage to the quality of the initial assumptions (Keynes [1936] 1973, 297–298).
In reflecting on the limits of statistical inference, Keynes noted that statistical analysis depends on
the universal validity of assumptions, and is valid only when the variables used are wholly indepen-
dent of each other (Keynes [1921] 1973, 276–277; Carabelli and Cedrini 2014, 28–29; Lawson 2012, 1–
2) – a test many multivariate analyses fail to meet. As Keynes remarked,
we are faced at every turn with the problem of organic unity, of discreteness, of discontinuity – the whole is not
equal to the sum of the parts, comparison of quantity fails us, small changes produce large effects, the assump-
tions of a uniform and homogeneous continuum are not satisfied. (Keynes [1933] 1972, 262; Carabelli and Cedrini
2014, 36–37)

The ‘atomic hypothesis’, which justifies inductive reasoning and mathematical calculus, cannot be applied to
organic complex systems … Keynes is critical of the attempt to blindly apply mathematics and statistics, with
their assumptions of homogeneity, atomism and independence, to an economic material that is essentially
vague and indeterminate, not homogeneous, not divisible in homogeneous independent parts, not finite, and
is characterised by organic interdependence. (Carabelli and Cedrini 2014, 29–30)

In the same vein, Pierre Bourdieu and Jean-Claude Passeron remark on the interdependent and
organic character of the factors affecting social inequality:
It is the system of factors, acting as a system, which exerts the indivisible action of a structural causality on behav-
iour and attitudes … so that it would be absurd to try to isolate the influence of any one factor, or, a fortiori, to
credit it with a uniform, univocal influence at the different moments of the process or in the different structures of
factors. (Bourdieu and Passeron [1990]1977, 87)

A multi-variate analysis of relations between higher education and work would require that all rel-
evant variables are independent of each other, each separately interacts with the other variables, and
all interactions are governed by a common law. Such conditions hold only in closed systems gov-
erned by a single universal logic. They do not apply in the real world of education and work,
where many variables under consideration are impossible to conclusively separate from each
other, for example, family income, cultural capital, ‘ability’. Worse, in many (if not most) human
capital studies, the statistical correlation or coincidence between two variables is held to constitute
not a suggestive association between them, but a demonstration (or a strong suggestion) that they
are causally related. It is remarkable how often statistically based research papers about higher edu-
cation and earnings conclude with a statement equating correlation with causality, with weak or no
qualification, and with little regard for the limits imposed by the contextual conditions in which the
data were generated.
By no means all multi-variate analyses are careless of these difficulties. Multicollinearity and endo-
geneity have generated a large literature. However, efforts to compensate for their problems from
within the method cannot be conclusive; and if the limitations of the research are fully acknowl-
edged, its precision and its generic claims are fundamentally undermined. This suggests that the sol-
ution often lies in stepping outside the multivariate framework.

Problems of human capital theory

These problems are now considered specifically in relation to human capital theory.

Bounded statistical analysis and organic realities

The OECD sees human capital theory as necessary but not sufficient, noting that ‘a host of education-
related and context-related factors’ other than learning itself ‘affect the returns to education’ (OECD
2014a, 151). Richard Arum and Josipa Roksa are more sceptical, arguing that ‘colleges have little
control over wage outcomes’ (Arum and Roksa 2014b, 125). As noted, there is a long literature on
294 S. MARGINSON

factors that affect earnings, additional to higher education per se. Graduate earnings vary by the
differential status and resources of HEIs (‘college quality’ in the US literature); family income
(Wolniak et al. 2008, 131); family life not mediated by education (Triventi 2013, 45) including
support for child development such as whether children are read to at a young age (Corak 2012,
6); measured ‘ability’; type of secondary school; and social and family networks at entry to higher edu-
cation, entry to work and later careers (Bingley, Corak, and Westergård-Nielsen 2011; Hallsten 2014,
20; Arum and Roksa 2014b, 14; Borgen 2015.). Earnings are affected by varying customs and hierar-
chies in professions and workplaces; by the wage determination system and the industrial balance of
power (Piketty 2014, 305); and the configurations and fluctuations of economies. Given these factors
– which are all constantly changing – it is delusional to seek to measure or compare the quantity,
quality or productivity of education programmes, institutions or systems, on the basis of the
private rates of return to, or the rate of employment of, those graduates.
Statistical methods design to eliminate the effects of factors other than higher education floun-
der given the number of variables, their interdependency, and the impossibility of isolating each
causal factor from all the others. This in turn leads to problems of selection effects. The economist
struggles to find causality in the face of multicollinearity problems but the comparisons are con-
taminated by hidden factors. It must be said the problem of selection effects is a non-problem
grounded in the assumption that elements are atomistically separable. Nevertheless, in research
premised on the assumption of atomism, the problem must be solved. Attempts to account for
selection effects generate diverse results. Reviewing research on graduate earnings in China,
Hongbin Li and colleagues note that while some researchers identify returns to college selectivity
after selection effects are accounted for, others find these returns disappear. Much of the variation
in findings is due to arbitrary assumptions about selection effects, not variations in the real world (Li
et al. 2012, 78–79).

Non-homogenous and non-linear material

Human capital theory also fails to deal effectively with real-world sites in which patterns are non-
linear and non-homogeneous. Nicolai Borgen remarks in relation to studies of graduate outcomes
that while averages create order from diversity, they do so ‘by masking important heterogeneity
across the wage distribution’ (Borgen 2015, 43). He also identifies non-linear economic returns associ-
ated with higher education. Family background effects seem greatest at the top end of the wage dis-
tribution. ‘The returns to college quality are five times larger at the 90th quantile compared to the
10th quantile’ (42). Gregory Wolniak and colleagues find that after graduation, education is associ-
ated with a growing impact on earnings, in non-linear fashion (Wolniak et al. 2008, 131). Paul
Bingley, Miles Corak, and Niels Westergård-Nielsen researched the ‘intergenerational transmission
of employers’ between fathers and sons. In both Canada and Denmark, 30–40% of young adults at
some time work for a firm that has employed their fathers. In both countries, the transmission of
employers was positively associated with paternal earnings, ‘rising distinctly and sharply at the
very top of the father’s earnings distribution’ (Bingley, Corak, and Westergård-Nielsen 2011, 3, 7
and 12.). Again at the top end on incomes, Iftikhar Hussain and colleagues find the apparent
income effects of selective institutions inflate, and returns associated with degrees are increasing
(Hussain, McNally, and Telhaj 2009, 12). Lemieux (2006) finds that in the US, over 30 years, ‘within-
group inequality grew substantially among college-educated workers, but changed little for most
other groups’ (195). ‘The median, the tenth and the ninetieth percentiles are remarkably stable for
up to 12 years of education’. However, ‘above 12 years of education … the return to education at
the ninetieth percentile increases much more than the return to education at the tenth percentile,
leading to a large increase in the 90-10 gap’ (196). Lemieux concludes that ‘changes in wage inequal-
ity are increasingly concentrated in the very top end of the wage distribution … ’ [and] ‘postsecond-
ary education plays a crucial role in explaining this phenomenon’ (199). The empirical data are
consistent with Bingley, Corak, and Westergård-Nielsen (2011) and Borgen (2015), but Lemieux’s
 STUDIES IN HIGHER EDUCATION 295

interpretation is questionable. Is the concentration at the top end of incomes an effect of higher edu-
cation, as Lemieux suggests, or due to something else?
These empirical patterns are consistent with Piketty (2014) and Saez (2013) on income inequality
in the US. In 2012, the top 1% of income recipients received 22.5% of all income, second highest level
since 1928 (Saez 2013, 7–9). Non-linear returns at the top end of the income distribution cannot be
explained by human capital theory without introducing factors from outside the theory, eroding its
claim as a universal lens. For if returns to labour are simply driven by the chain of human capital and
marginal productivity, then income inequality must also derive from unequal skills and productivity.
Piketty comments wryly that while ‘US educational institutions … surely need to be improved and
made more accessible’, they ‘probably do not deserve such extravagant blame’ (Piketty 2014, 330):
This very sharp discontinuity at the top income levels is a problem for the theory of marginal productivity: when
we look at the changes in the skill levels of different groups in the income distribution, it is hard to see any dis-
continuity between ‘the 9 percent’ and ‘the 1 percent’, regardless of what criteria we use: years of education,
selectivity of educational institution, or professional experience. (Piketty 2014, 314)

The reality, however, is that higher education is largely decoupled from the surge in top incomes
(Piketty 2014, 315 and 330). The recent broad consensus among scholars studying work-based
incomes in the US is that the blow-out in managerial salaries is more a price effect than an education
effect (e.g. Autor, Katz, and Kearney 2008, 317–318; Mouw and Kalleberg 2010; Hanley 2011; Bentele
2013; OECD 2014b), one grounded in tax cuts for high-income earners and work-related practices
such as salary deregulation, de-unionisation and performance-pay. Human capital theory cannot
explain sharp variations in graduate incomes over time, nor differences in patterns of income distri-
bution, and top-end earnings, in countries with similar higher education. ‘A major problem’ facing
‘marginal productivity theory’ is that ‘the explosion of very high salaries occurred in some developed
countries but not others. This suggests that institutional differences between countries rather than
general and a priori universal causes such as technological change played a central role’ (Piketty
2014, 315, also 304, 308, 321). Nevertheless, the non-linear earnings pattern is suggestive, implying
that higher education affects American occupational outcomes less among high-income earners than
among middle-level earners. While this again undermines the universal claim of human capital
theory, it suggests a fruitful opening for further research.

Heterogeneity of higher education and work

The human capital equations unify higher education and work at the cost of suppressing much that
is distinctive about each. Arguably, work and higher education are different kinds of social site, each
with its distinctive history, requirements, daily practices, subject-positions, rhythms and drivers. This
does not mean that work and higher education are unconnected. Graduation is associated with
higher employability and earnings (OECD 2014a, 102–170), whether due to superior attributes of
graduates or their positioning in the labour markets. Some higher education is in continuum
with work (e.g. selective programmes that train professionals such as doctors), and many higher
education programmes have occupational contents. Students and graduates, HEIs, professions
and employers often make strenuous efforts to connect education and work. However, the linear
transition imagined in the human capital narrative does not describe higher education/work
relations. The transition is often fraught. The education/work alignment is partial and unclear.
Relations between the two domains are multiple, context-bound, fragmented, uneven and must
be continually worked on.
For many graduates, job allocation lacks precision, especially in US higher education with a high
proportion of generic degrees (Roksa 2005, 225). Josipa Roksa and Tania Levey state:
Many educational credentials have no obvious matches in the labour market. This includes the majority of high
school graduates in general and academic tracks and a large portion of college graduates majoring in liberal arts
and sciences. Consequently, finding a job in one’s field of study is not only an individual dilemma, it is a process
296 S. MARGINSON

that reflects the relationship (or lack thereof) between the educational system and the labour market. (Roksa and
Levey 2010, 391)

Schneider and Stevenson (1999, 79–85) find that only 44% of students had ‘aligned’ educational
ambitions, meaning that they planned to complete the amount of education required by their
intended occupations (Arum and Roksa 2014a, 34). Many students keep their vocational options
open. Often, they enrol for more reasons than vocational planning, studying subjects they are
good at, or they enjoy, while hoping that the future will work out. Though this strategy embodies
uncertainty, because all graduates have a positional advantage in the labour market vis-à-vis non-
graduates, such confidence is not wholly misplaced. Robst (2007, 398) notes ‘the eventual match
between degree field and occupation is uncertain when selecting a major’. He finds that 55% of
respondents report a close relation between their work and field of study, 25% state that they are
‘somewhat related’, and 20% that they are not related (402), though Robst has difficulty defining
the work-relatedness of general degrees.
Even among specifically trained graduates, many enter occupations that are outside their fields of
training, often not always with income penalties (van de Werfhorst 2002, 301; Robst 2007, 403–404;
Melguizo and Wolniak 2012, 383). This lack of fit between formal training and work reflects the messy
way that labour markets operate. Many professional jobs are generic. They can be filled by graduates
from any field, and level of education and possibly institution attended are more significant than field
of study. Many graduates take jobs that provide the best pay and career prospects at the time of
selection. At this career point, some will depart from their qualifications and a proportion never
return. Some specialised positions are filled by persons trained in specialist fields other than that
of the position. For their part, employers select the ‘best’ person from the available pool. Specific
training and qualifications are only two of the factors in play. Studies of graduate selection indicate
that the attributes of potential employees that influence selection also include institution attended,
extra-curricular activities as students, subjective perceptions of ‘fit’ between graduate and workplace,
and personal ties (e.g. Bingley, Corak, and Westergård-Nielsen 2011; Tholen et al. 2013; Borgen 2015;
Rivera 2015).
There is more vocational specificity in education and predictable pathways to work in countries
such as Germany than in the US. In Germany, this is achieved not by market coordination in education
and work as Becker imagined, but by ‘tight linkages between occupational groups, education and
training practices, and certification boards’. German practice appears to conform ‘nicely to human
capital models’ but ‘these completely fail to capture the importance of the elaborate institutional fra-
mework that enables the German certification regime to operate as they predict’ (Hansen 2011, 32).
Nor does human capital theory explain how education enhances productivity (43) which remains a
black box. One constraint is its methodological individualism (Lukes 1973). It is impossible to accu-
rately attribute enhanced value to individuals working in a combined workplace, as are most employ-
ees (Piketty 2014, 330–331).

Other explanations of education and work

Human capital theory understands only some students/graduates, those who consider the lifetime
earnings attached to different choices and weigh them against the costs of study. Many students/
graduate fail at being a choice-making self-investing homo economicus. Jens Thomsen and col-
leagues report that at enrolment some students ignore forgone earnings during study (Thomsen
et al. 2013, 471). Others know graduate earnings only in their chosen occupation, not related
fields (Robst 2007, 399). Borgen (2015, 34) states that many students do not ‘self-select into colleges
based on expected gain’. Students have many interests in addition to credentials, future earnings
and careers, including network building (Armstrong and Hamilton 2013); the accumulation of
knowledge, generic skills and cultural capital; intellectual formation as an end in itself; cultural
activities; and social or political activism. They mix their goals, practices and modes of reflexivity.
 STUDIES IN HIGHER EDUCATION 297

However, if one other effect is admitted, then human capital theory can no longer function as a
closed system.
Because it is a closed system, human capital theory has never adequately addressed its cousin,
screening theory. There is evidence for the presence of the signalling function (just as there is evi-
dence some students estimate lifetime earnings in different jobs), though screening no more consti-
tutes a universal explanation than does human capital theory. For example, Arum and Roksa (2014b,
80–81) note that business graduates have strong early wage returns, despite relatively low measured
cognitive formation in that field. ‘Some majors serve as better signals of employability than others,
regardless of whether those degrees are underpinned by actual field-specific knowledge and
skills’. OECD data on earnings suggest that, in some countries, the returns to qualifications exceed
the returns to measured skills; in other countries, the ratio is reversed (OECD 2014a, 109). Both
human capital and signalling effects are at play (and without excluding other effects). Yet, often,
social scientists feel obliged to choose between them (e.g. Wolniak et al. 2008, 124–125; Baker,
2011, 8; Hu and Vargas 2015, 3).
Human capital theory also fails to encompass occupational and social status in education and
work, which is not fully captured by the earnings function. Arum and Roksa (2014b, 57) emphasise:
‘Rewards to occupations are related not just to income but also to occupational status and prestige. In
social settings, individuals are typically asked about what they do, not how much money they earn’.
Many studies identify status goals and effects, and variations in the respective roles of earnings and
status outcomes, by gender, by field of study and ‘college quality’, between countries and over time
(e.g. Zhao 2012; Triventi 2013, 55–57; Thomsen et al. 2013, 471; Arum and Roksa 2014b, 80–81; Hen-
nessy 2014, 47; Hu and Vargas 2015). Investigating outcomes for 13-year out graduates, Roksa finds
that for those with generic degrees working in the public and non-profit sectors, a managerial role is
often more attractive than higher earnings. ‘Graduates of female-dominated fields are disproportio-
nately employed in public and non-profit organizations which offer lower monetary rewards but
facilitate access to professional and managerial positions’ (Roksa 2005, 207). The passage of time
affects income and status in contrasting ways in different fields of study. ‘Occupationally specific
degrees are beneficial at the point of entry into the labour market but have the lowest growth in
occupational status over time’ (Roksa and Levey 2010, 389), though they do better in earnings
(399). Separately associated with both level of education and political standing, status opens the
way to jobs and income. Triventi (2013) in four European countries, and Hu and Vargas (2015) in
China, find that ‘college quality’ is associated with higher occupational status. Hu and Vargas
(2015) note that status is a signal of prestige to employers and correlates with the likelihood of
assuming a managerial position (19).

Conclusions
Human capital theory presents its core propositions about education, learning, productivity and earn-
ings as a necessary and sufficient truth about higher education and work, albeit (in some studies and
to varying degree) joining these core propositions at the periphery to caveats and moderating
factors. The theory’s claim about its universal application, in conjunction with the intellectual and
policy dominance it has long exercised, has disrupted the possibility of a social science of education
and work at a higher level of real-world complexity.
Higher education and work are different and separated social sites, though there are important
overlaps in practice. This is not a relationship of identity, regularity or a linear continuum. Nor is it
a dialectic, in which two contrasting parts form a unified system with a shared logic. Education
and work are heterogeneous in relation to each other. Their relations are never wholly resolved or
resolvable in practice; and if a final resolution is attempted in theory, then something can be lost
from view (for example, the generic or liberal component of intellectual formation in higher edu-
cation, which has no specific vocational aspect, slips from empirical sight, or appears negative).
Relations between higher education and work are also context-bound. They vary by country, field
298 S. MARGINSON

of study, type of institution, financing of education, occupation, industry, employment site and over
time. For example, both Goodman (2014) and Zhao (2012) note that status drivers are especially
important in China, arguably playing a larger role in comparison with income drivers than in the
US. The task of a realist study of education/work is to combine sensitivity to context with an
account of larger patterns, including aspects of social relations not directly observable (Sayer
2000). These patterns are both internal and external. Like all semi-bounded systems, the dyad of
higher education and work is connected to other systems or ‘fields’ (Bourdieu 1993; Fligstein and
McAdam 2015), including income determination and wealth creation, labour markets, state and poli-
tics, taxation, public spending and programmes, global flows.
Given that education/work relations entail complex and multiple phenomena – and no theoris-
ation can contain all phenomena, while retaining a bounded coherence – it is axiomatic that more
than one description of education/work relations can provide useful insights. Gerber and Cheung
(2008, 301) canvass four possible reasons for the higher earnings of graduates of elite institutions:
elite HEIs impart more valuable human capital, elite graduates signal their status to employers, stu-
dents in elite HEIs garner more valuable social capital, graduates from elite HEIs have enjoyed advan-
tages such as family affluence or ability that generate more favourable outcomes. However, in this
paper, they do not consider the possibility that all four factors are in play, with the mix varying
over time and between countries and between fields of study. In orthodox sociology, as in orthodox
economics, theoretical multiplicity is mostly a bridge too far. The drive for universal explanation, that
elusive talisman of social science, overrides real-world complexity. Hansen (2011) rightly argues that
all major theories of education/work relations, such as human capital, signalling and ‘credentialist’
certification’ are ‘to some degree wanting’ (31). The obverse is also true. Differing research-based
explanations of education and work contribute to knowledge. Some are more explanatory than
others. Confronting the complexity of education/work, the task of research is to determine which
explanation(s) is (are) primary, not to impose an exclusive straightjacket on the material.
It is not the purpose here to outline an alternative theorisation to human capital economics.
However, an alternative approach would be grounded in a meta-method that would use a semi-
open analytical system or model, admit multiple theories rather than one exclusive theory, and
draw on both quantitative and qualitative research and combine their insights. Statistical reasoning
would have a modest role. In slicing into parts of the empirical terrain, statistical studies can be sug-
gestive. For example, in research on top-end graduate incomes, the findings become interesting
where the linear patterns break down. The limits of statistical analysis show not when it is used
for specific inquiry but where it purports to provide a holistic picture, when it is substituted for an
historicised synthesis, and multi-variate modelling and calculation are used as a substitute for
more difficult processes of complex judgement. The use of multi-variate analysis should be limited
to instances when the variables are independent.

Disclosure statement
No potential conflict of interest was reported by the authors.

References
Armstrong, E., and L. Hamilton. 2013. Paying for the Party: How College Maintains Inequality. Cambridge, MA: Harvard
University Press.
Arum, R., and J. Roksa. 2014a. Academically Adrift: Limited Learning on College Campuses. Chicago: University of Chicago
Press.
Arum, R., and J. Roksa. 2014b. Aspiring Adults Adrift: Tentative Transitions of College Graduates. Chicago, IL: University of
Chicago Press.
Autor, D. H., L. F. Katz, and M. S. Kearney. 2008. “Trends in U.S. Wage Inequality: Revising the Revisionists.” The Review of
Economics and Statistics 90 (2): 300–23.
 STUDIES IN HIGHER EDUCATION 299

Baker, D. P. 2011. “Forward and Backward, Horizontal and Vertical: Transformation of Occupational Credentialing in the
Schooled Society.” Research in Social Stratification and Mobility 29: 5–29.
Becker, G. 1964. Human Capital: A Theoretical and Empirical Analysis with Special Reference to Education. Chicago, IL:
University of Chicago Press.
Becker, G., and K. Murphy. 2003. Social Economics: Market Behaviour in a Social Environment. Cambridge, MA: Harvard
University Press.
Bentele, K. G. 2013. ““Distinct Paths to Higher Inequality? A Qualitative Comparative Analysis of Rising Earnings Inequality
among U.S. States, 1980–2010.” Research in Social Stratification and Mobility 34: 30–57.
Berg, I. 1971. Education and Jobs: The Great Training Robbery. New York: Praeger.
Bhaskar, R. 2008. A Realist Theory of Science. 3rd ed. London: Verso.
Bingley, P., M. Corak, and N. Westergård-Nielsen. 2011. The Intergenerational Transmission of Employers in Canada and
Denmark. IZA Discussion Paper No. 5593. Bonn: Institute for the Study of Labor.
Blaug, M. 1976. “The Empirical Status of Human Capital Theory: A Slightly Jaundiced Survey.” Journal of Economic
Literature 14 (3): 827–55.
Boliver, V. 2011. “Expansion, Differentiation, and the Persistence of Social Class Inequalities in British Higher Education.”
Higher Education 61: 229–42.
Boliver, V. 2013. ““How Fair is Access to More Prestigious UK Universities?” British Journal of Sociology 64 (2): 344–64.
Borgen, N. T. 2015. “College Quality and the Positive Selection Hypothesis: The ‘Second Filter’ on Family Background in
High-Paid Jobs.” Research in Social Stratification and Mobility 39: 32–47.
Bourdieu, P. 1984. Distinction: A Social Critique of the Judgment of Taste. London: Routledge and Kegan Paul.
Bourdieu, P. 1988. Homo Academicus. Stanford, CA: Stanford University Press.
Bourdieu, P. 1993. The Field of Cultural Production. New York: Columbia University Press.
Bourdieu, P., and J.-P. Passeron. (1977)1990. Reproduction in Education, Society and Culture. London: Sage.
Bowles, S., and H. Gintis. 1976. Schooling in Capitalist America: Educational Reform and the Contradictions of Economic Life.
New York: Basic Books.
Britton, J., L. Dearden, N. Shephard, and A. Vignoles. 2016. How English Domiciled Graduate Earnings Vary with Gender,
Institution Attended, Subject and Socioeconomic Background. London: Nuffield Foundation.
Brown, P., H. Lauder, and D. Ashton. 2012. The Global Auction: The Broken Promises of Education, Jobs, and Incomes. Oxford:
Oxford University Press.
Carabelli, A. M., and M. A. Cedrini. 2014. “Chapter 18 of The General Theory ‘Further Analysed’: Economics as a Way of
Thinking.” Cambridge Journal of Economics 38: 23–47.
Chapman, B., T. Higgins, and J. Stiglitz, eds. 2014. Income Contingent Loans: Theory, Practice and Prospects. Basingstoke:
Palgrave Macmillan.
Collins, R. 1979. The Credential Society: An Historical Sociology of Education and Stratification. Oxford: Oxford University
Press.
Corak, M. 2012. Inequality From Generation to Generation: The United States in Comparison. Ottawa, ON: Graduate School
of Public and International Affairs, University of Ottawa.
Delaney, L., C. Harmon, and C. Redmond. 2011. ““Parental Education, Grade Attainment and Earnings Expectations
Among University Students.” Economics of Education Review 30 (6): 1136–52.
Denison, E. 1962. The Sources of Economic Growth in the United States and the Alternative Before Us. New York: Committee
for Economic Development.
Dow, S. 1990. “Beyond Dualism.” Cambridge Journal of Economics 14: 143–57.
Dow, S. 2012. “Different Approaches to the Financial Crisis.” Economic Thought 1: 80–93.
Fligstein, N., and D. McAdam. 2015. A Theory of Fields. Oxford: Oxford University Press.
Friedman, M. 1953. “The Methodology of Positive Economics.” In Essays in Positive Economics, edited by M. Friedman, 3–
43. Chicago: University of Chicago Press.
Gerber, T. P., and S. Y. Cheung. 2008. “Horizontal Stratification in Postsecondary Education: Forms, Explanations, and
Implications.” Annual Review of Sociology 34: 299–318.
Glomm, G., and B. Ravikumar. 1992. “Public Versus Private Investment in Human Capital: Endogenous Growth and Income
Inequality.” Journal of Political Economy 100 (4): 818–34.
Goodman, D. 2014. Class in Contemporary China. Cambridge: Polity.
Hallsten, M. 2014. “Inequality Across Three and Four Generations in Egalitarian Sweden: 1st and 2nd Cousin Correlations
in Socio-Economic Outcomes.” Research in Social Stratification and Mobility 35 (1): 19–33.
Hanley, C. 2011. “Investigating the Organizational Sources of High-Wage Earnings Growth and Rising Inequality.” Social
Science Research 40: 902–16.
Hansen, H. 2011. “Rethinking Certification Theory and the Educational Development of the United States and Germany.”
Research in Social Stratification and Mobility 29: 31–55.
Hennessy, P. 2014. Establishment and Meritocracy. London: Haus.
Hodgson, G. M. 2014. “What is Capital? Economists and Sociologists Have Changed Its Meaning: Should It Be Changed
Back?” Cambridge Journal of Economics 38: 1063–86.
300 S. MARGINSON

Hoxby, C., and C. Avery. 2013. The Missing ‘One-Offs’: The Hidden Supply of High-Achieving, Low-Income Students. Spring:
Brookings Papers on Economic Activity.
Hu, A., and N. Vargas. 2015. “Horizontal Stratification of Higher Education in Urban China.” Higher Education. doi:10.1007/
s10734-014-9833-y.
Hussain, I., S. McNally, and S. Telhaj. 2009. University Quality and Graduate Wages in the UK. IZA Discussion Paper No. 4043.
Bonn: Institute for the Study of Labor.
Keeley, B. 2007. Human Capital. Paris: OECD.
Kerr, C. 2001. The Gold and the Blue: A Personal Memoir of the University of California, 1949–1967. Volume 1: Academic
Triumphs. Berkeley: University of California Press.
Keynes, J. M. (1921) 1973. A Treatise on Probability. Reprinted in E. Johnson and D. Moggridge, eds. The Collected Writings
of John Maynard Keynes, Vol. 8. London: Macmillan, for the Royal Economic Society.
Keynes, J. M. (1933) 1972. Essays in Biography. Reprinted in E. Johnson and D. Moggridge, eds. The Collected Writings of
John Maynard Keynes, Vol. 10. London: Macmillan, for the Royal Economic Society.
Keynes, J. M. (1936) 1973. The General Theory of Employment, Interest and Money. Reprinted in E. Johnson and
D. Moggridge, eds. The Collected Writings of John Maynard Keynes, Vol. 7. London: Macmillan, for the Royal
Economic Society.
Lawson, T. 2003. Reorienting Economics. London: Routledge.
Lawson, T. 2012. “Mathematical Modelling and Ideology in the Economics Academy: Competing Explanations of the
Failings of the Modern Discipline?” Economic Thought 1: 3–22.
Lemieux, T. 2006. “Postsecondary Education and Increasing Wage Inequality.” AEA papers and proceedings, May.
Li, H., L. Meng, X. Shi, and B. Wu. 2012. “Does Attending Elite Colleges Pay in China?” Journal of Comparative Economics 40:
78–88.
Lukes, S. 1973. Individualism. Oxford: Basil Blackwell.
Marginson, S. 2016a. The Dream is Over: The Crisis of Clark Kerr’s California Idea of Higher Education. Berkeley: University of
California Press and the Center for Studies in Higher Education, University of California.
Marginson, S. 2016b. “High Participation Systems of Higher Education.” The Journal of Higher Education 87 (2): 243–71.
Marshall, A. 1898. “Distribution and Exchange.” Economic Journal 8 (29): 37–59.
Marshall, A. (1920) 1961. Principles of Economics. Revised Edition. London: Macmillan.
Melguizo, T., and G. C. Wolniak. 2012. “The Earnings Benefits of Majoring in STEM Fields among High Achieving Minority
Students.” Research in Higher Education 53: 383–405.
Mincer, J. 1958. “Investment in Human Capital and Personal Income Distribution.” Journal of Political Economy 66 (4):
281–302.
Mouw, T., and A. L. Kalleberg. 2010. “Occupations and the Structure of Wage Inequality in the United States, 1980s to
2000s.” American Sociological Review 75 (3): 402–431.
Organisation for Economic Cooperation and Development, OECD. 2014a. Education at a Glance 2014. Paris: OECD.
Organisation for Economic Cooperation and Development, OECD. 2014b. United States: Tackling High Inequalities,
Creating Opportunities for All. Paris: OECD.
Piketty, T. 2014. Capital in the Twenty-First Century. Cambridge, MA: Belknap Harvard University Press.
Psacharopoulos, G. 1994. “Returns to Investment in Education: A Global Update.” World Development 22 (9): 1325–43.
Rivera, L. 2015. Pedigree: How Elite Students Get Elite Jobs. Princeton, NJ: Princeton University Press.
Robst, J. 2007. “Education and Job Match: The Relatedness of College Major and Work.” Economics of Education Review 26:
397–407.
Roksa, J. 2005. “Double Disadvantage or Blessing in Disguise? Understanding the Relationship Between College Major
and Employment Sector.” Sociology of Education 78 (3): 207–232.
Roksa, J., and T. Levey. 2010. “What Can You Do With That Degree? College Major and Occupational Status of College
Graduates Over Time.” Social Forces 89 (2): 389–415.
Romer, P. M. 1990. “Endogenous Technological Change.” Journal of Political Economy 98: S71–S102.
Saez, E. 2013. “Striking it Richer. The Evolution of Top Incomes in the United States.” Berkeley: University of California,
Berkeley, Department of Economics. http://eml.berkeley.edu//~saez/saez-UStopincomes-2012.pdf.
Sayer, A. 2000. Realism and Social Science. London: Sage.
Schneider, B., and D. Stevenson. 1999. The Ambitious Generation: America’s Teenagers Motivated but Directionless. New
Haven, CT: Yale University Press.
Schultz, T. W. 1959. “Investment in Man: An Economist’s View.” Social Service Review 33 (2): 109–117.
Schultz, T. W. 1960. “Capital Formation by Education.” Journal of Political Economy 68 (6): 571–83.
Schultz, T. W. 1961. “Investment in Human Capital.” American Economic Review 51 (1): 1–17.
Soares, J. 2007. The Power of Privilege: Yale and America’s Elite Colleges. Stanford: Stanford University Press.
Social Mobility Commission, United Kingdom. 2016. Socio-economic Diversity in Life Sciences and Investment Banking.
London: Social Mobility Commission.
Spring, J. 2015. Economization of Education: Human Capital, Global Corporations, Skills-based Schooling. New York:
Routledge.
Stiglitz, J. 2013. The Price of Inequality. Penguin: London.
 STUDIES IN HIGHER EDUCATION 301

Taylor, C. 2007. A Secular Age. Cambridge, MA: Harvard University Press.

Teichler, U. 2009. Higher Education and the World of Work: Conceptual Frameworks, Comparative Perspectives, Empirical
Findings. Rotterdam: Sense.
Tholen, G., P. Brown, S. Power, and A. Allouch. 2013. “The Role of Networks and Connections in Educational Elites” Labour
Market Entrance.” Research in Social Stratification and Mobility 34: 142–54.
Thomsen, J. P., M. D. Munk, M. Eiberg-Madsen, and G. I. Hansen. 2013. “The Educational Strategies of Danish University
Students From Professional and Working-Class Backgrounds.” Comparative Education Review 57 (3): 457–80.
Triventi, M. 2013. “The Role of Higher Education Stratification in the Reproduction of Social Inequality in the Labor
Market.” Research in Social Stratification and Mobility 32: 45–63.
Trow, M. 1973. Problems in the Transition From Elite to Mass Higher Education. Berkeley, CA: Carnegie Commission on
Higher Education.
United Nations Educational, Scientific and Cultural Organisation, UNESCO. 1968. Readings in the Economics of Education.
Paris: UNESCO.
van de Werfhorst, H. 2002. “Fields of Study, Acquired Skills and the Wage Benefit from a Matching Job.” Acta Sociologica
45: 286–303.
Wolniak, G. C., T. A. Seifert, E. J. Reed, and E. T. Pascarella. 2008. “College Majors and Social Mobility.” Research in Social
Stratification and Mobility 26: 123–39.
Zhao, W. 2012. “Economic Inequality, Status Perceptions, and Subjective Well-being in China’s Transitional Economy.”
Research in Social Stratification and Mobility 30: 433–50.