Ambition, Human Capital Acquisition and the Metropolitan Escalator

Abstract

Gordon I. R. Ambition, human capital acquisition and the metropolitan escalator, Regional Studies. This paper examines the relation between ambition, as a form of dynamic human capital, and the escalator role of high-order metropolitan regions, as originally identified by A. J. Fielding. It argues that occupational progression in such places particularly depends on concentrations both of people with more of this asset and of jobs offering preferential access to valued elements of tacit knowledge, interacting in thick, competitive labour markets. This is partially confirmed with analyses of British Household Panel Study (BHPS) data on long-term progression showing that only the more ambitious gain from residence in the extended London region, and that they only progress faster there.

Gordon I. R. 企图心、人力资本取得与大都会中的自动扶梯区域，区域研究。本文检视做为一种动态人力资本形式的企图心和最初由 A. J. Fielding 所指认的高阶大都会区域的电扶梯角色之间的关联性。本文主张，这些地方中的职业发展，特别取决于具有此类资质者和得以提供优先途径以获取隐性知识中的重要元素之工作的双重聚集，这些重要元素在厚实、竞争的劳动市场中互动。此一主张随着长期发展的英国家户盘区调查(BHPS)资料的分析而部分被确认,该 分析显示，只有较具有企图心者，能够从居住在伦敦的延伸区域中获益，且他们也仅能够在此地较为迅速的发展。

Gordon I. R. L'ambition, l'acquisition du capital humain et l'escalator metropolitain, Regional Studies. Ce présent article cherche à examiner le rapport entre l'ambition, en tant qu'une sorte de capital humain dynamique, et le rôle d'escalator des régions métropolitaines de rang supérieur, d'après A. J. Fielding. On affirme que l’évolution professionnelle dans de tels endroits dépend notamment des concentrations à la fois des gens dotés de davantage de cet atout et des emplois qui facilitent un accès préférentiel à des éléments prisés de la connaissance tacite, interagissant avec d'importants marchés du travail compétitifs. Cela est en partie confirmé à partir des analyses des données provenant du British Household Panel Survey (BHPS – enquête par panel auprès des ménages britanniques) sur l’évolution professionnelle à long terme qui montre que seulement les plus ambitieux profitent d'un lieu de résidence situé dans l'agglomération de Londres, et qu'eux seuls y progressent plus rapidement.

Gordon I. R. Ehrgeiz, Erwerb von Humankapital und die eskalierende Rolle der Metropole, Regional Studies. In diesem Beitrag wird die Beziehung zwischen dem Ehrgeiz als Form von dynamischem Humankapital und der zuerst von A. J. Fielding identifizierten eskalierenden Rolle von Metropolitanregionen höherer Ordnung untersucht. Es wird die These aufgestellt, dass das berufliche Vorwärtskommen an solchen Orten vor allem von der Höhe der Konzentration von stärker mit diesem Merkmal ausgestatteten Personen sowie von Arbeitsplätzen mit bevorrechtigtem Zugang zu wertvollen Elementen impliziten Wissens abhängt, wobei sich diese beiden Aspekte in dichten und konkurrenzbetonten Arbeitsmärkten gegenseitig beeinflussen. Diese These wird teilweise durch Analysen der Daten der British Household Panel Study (BHPS) über langfristiges Vorwärtskommen bestätigt, aus denen hervorgeht, dass nur ehrgeizigere Personen von einem Wohnort im Großraum London profitieren und nur dort beruflich schneller vorankommen.

Gordon I. R. Ambición, adquisición de capital humano y el papel escalador de áreas metropolitanas, Regional Studies. En este artículo se examina la relación entre ambición como forma de capital humano dinámico y el papel escalador de las regiones metropolitanas de primer orden, tal como lo identificó originalmente A. J. Fielding. Se sostiene que la progresión laboral en tales lugares depende específicamente de las concentraciones de personas con esta característica así como de los trabajos con acceso preferencial a elementos valiosos de conocimiento tácito, aspectos que influyen en mercados laborales densos y competitivos. Esto queda confirmado parcialmente con los análisis de los datos de la encuesta británica de presupuestos familiares (BHPS) sobre la progresión a largo plazo en los que se muestra que solo las personas más ambiciosas sacan provecho de una residencia en el área de Londres, y que solo aquí consiguen desarrollar una carrera profesional con más rapidez.

• Escalator region
• Migration
• Urban labour market
• London
• Social mobility
• Human capital

• 电扶梯区域
• 移民
• 城市劳动市场
• 伦敦
• 社会流动性
• 人力资本

• Région escalator
• Migration
• Marché du travail urbain
• Londres
• Mobilité sociale
• Capital humain

• Eskalierende Region
• Migration
• Städtischer Arbeitsmarkt
• London
• Soziale Mobilität
• Humankapital

• Región ascendente
• Migración
• Mercado laboral urbano
• Londres
• Movilidad social
• Capital humano

JEL classifications:

• J24
• J61
• J62
• R23

INTRODUCTION: SEEKING LADDERS TOPPED WITH GOLD

Famously, over the centuries great cities have attracted young people with talent and ambition who are set on making their careers in the metropolis in the face of uncertain odds. The iconic example is that of (the English) Dick Whittington, reputedly a poor orphan from the countryside who was attracted to London by tales of ‘streets paved with gold’, who after overcoming his initial disillusion (with the help of prophetic church bells urging him to ‘turn again’) and finding a rich master who encourages him to invest all his assets (his cat) in a trading venture, from which he gains a quite unexpected fortune, allowing him to progress to a social pinnacle as three times Lord Mayor. The story of the real Sir Richard's ascent to the mayoralty in the fifteenth century is rather more prosaic, starting from roots in the landed (if indebted) gentry and progressing via apprenticeship to the ‘honourable trade’ of mercery to become a royal financier (Fischer, 2005). In either version, however, the City is seen to present opportunities for kinds of success not available elsewhere, though ultimately attainable only by a few, through an (obscure) combination of personal character, social connections and pure chance. Away from the context of folk tales, Harris and Todaro’s (1970) formal representation of city labour markets in the global South accounts for their high unemployment levels as the equilibrium outcome of a gamble which rural migrants make on the chances of gaining access to a set of formal sector jobs with high wages, sustained (in practice) by forms of social closure mechanism not transparent to them.

The continuing power of legends such as the Whittington story to encourage young people wishing ‘to seek their fortune’ to head for the metropolis is not independent of their exposure to other real success stories coming from closer to home. Since there is less news from those who fail to ‘make it’, it must be very hard for anyone to appraise realistically the likely return on this gamble. But recent statistical analyses of socio-economic outcomes for those who have made such a move to some of the leading cities in the global North do suggest that they progress further and faster than their peers (Glaeser and Maré, 2001; de la Roca and Puga, 2012). City streets may not be ‘paved with gold’ just for the taking, but they do seem to provide access to a set of ladders potentially topped with gold – for those with the will and strength to climb them. In rather more dynamic terms, Fielding (1993, 1995) formulated the idea of an ‘escalator region’ to represent the enhanced chances of social mobility – across class boundaries – that he observed among residents in London's extended metro region (the South East) between 1971 and 1991. In particular, he found that in-migrants to the region gained strongly – though it should be noted that the evidence for their greater success (based on inter-Censal mobility between places and classes) would have included any one-off ‘elevator’ effects resulting from entry to a region with many high-status jobs, as well as any continuing ‘escalator’ effects arising from their presence in a set of more dynamic labour markets.¹

This paper focuses more specifically on what evidence there is for pure escalator effects on occupational progression, and on who actually benefits from these. In the terms used by Glaeser and Maré (2001), analysing effects of relocation on earnings, it focuses on (continuous) ‘learning’ effects rather than on (one-off) ‘productivity’ boosts. The assumption is that when people with similar characteristics are compared, any (pure) escalator effects ought to be just as relevant to long-term residents as to migrants. If, after controlling for initial ‘elevator’ effects, in-migrants are still found to gain proportionately more from their residence in an escalator region, this should be traceable to some catalytic attribute distinguishing those who choose to move to ‘escalator regions’ from those who were born and stayed there. Youth and talent are the obvious candidates for this role, but less readily observable factors might also be crucial, including the ambition and energy in pursuing career advancement to which Fielding (1993, p. 147) refers.

All these ‘bright lights’ stories clearly beg fundamental questions (for the prospective migrant) about how many places are actually available on a city's escalator, and what it takes to get access to these – whether great talent, great ambition, great connections or great luck. For the more detached observer (whether researcher or policy-maker) there is a related set of questions as to whether faster progress up career ladders in such cities reflects:

• simply their concentrations of:

  	○	progressive economic activities/educational institutions and powerful patrons; and hence of
  	○	talented and ambitious young people attracted from the countryside and provinces; or also

• specifically urban externalities in the form of:

  	○	informally diffused tacit knowledge – the trade ‘mysteries’ in ‘the air’ of Marshall’s (1890) specialized industrial districts, or the creative ‘urban buzz’ of Storper and Venables’s (2004) more diverse agglomerations; and/or
  	○	thick, flexible and dynamic labour markets offering workers a personal financial incentive to pursue such knowledge (Glaeser and Maré, 2001).

These issues have wider significance in two respects. The first is in terms of how far the existence of economically dominant cities actually contributes to the human assets of a society and/or to inequality in the social distribution of these (and thus of real incomes too). The second concerns the familiar puzzle (for neo-classical economists at least) as to what could motivate strong migrational flows toward particular regions, if their effect is to eliminate disparities in the real incomes received by people with comparable attributes in different locations (Evans, 1990). Three elements in a (strong) urban escalator theory which might resolve this are: the specificity of opportunities attracting migrants; the impacts of spatial location on the development of relevant attributes; and the dependence of positive impacts on acceptance by (some) workers in these places of costly efforts with uncertain rewards.

Within the context of a wider study, attempting to address all these issues in relation to potential escalator regions within the UK, the prime focus of this paper is on evidence relevant to the second and third of these elements – seeking answers to questions about whether jobs in some places contribute more to the development of human capital, and how selective these effects are, as between people with different backgrounds and attitudes.

In order to pursue these goals, the remainder of the paper is organized in three main sections. The first section presents a theoretical framework, linking progression up an occupational ladder to dynamic forms of human capital, uneven spatial opportunities for exploiting this, and their relation to signalling behaviour in labour markets with high degrees of uncertainty. The second section outlines an operationalization of this, including measures of advancement in job status, and levels of career ambition, for application to an analysis of mobility and status change over the 1991–2008 period, using micro-data from the British Household Panel Study. The third section reports on results from this analysis (and of shorter-term change data from the Labour Force Survey). A short conclusion summarizes the findings, together with outstanding questions, and links these to on-going work (with colleagues) using complementary data sources, including the Census-based Longitudinal Study and the Annual Survey of Hours and Earnings.

A THEORETICAL FRAMEWORK: DYNAMIC HUMAN CAPITAL, OPPORTUNITIES AND SPACE

The theoretical framework motivating this paper involves a strong version of the urban escalator thesis, focused on the potential role of urban externalities in creating real assets in the form of additions to human (and social) capital, with the active participation of workers who invest their energies acquiring and developing these capacities while ‘on the job’.

Human capital, occupational status and ambition

Though often approximated by time spent in formal education (or qualifications achieved), in training and in work ‘experience’, at a conceptual level human capital involves the full range of an individual's embodied assets, in the sense of those personal characteristics² that contribute to the value derived from (any of) their activities. These are clearly diverse, and the labour market is accordingly structured around many different occupational positions, each associated with distinctive combinations of human capital, and connected by multiple ladders (along with some ‘snakes’). In this context there is a two-way link between occupations and human capital: on the one hand, access to specific jobs requires credible evidence of a particular range of competences; on the other hand, performance of these roles is the main route to accumulating the relevant capabilities, and getting these recognized.

For both reasons, progress up occupational ladders (toward jobs with higher rewards) is indicative of success in developing types of human capital to which the productive economy currently assigns higher values. While at the micro-level progress can reflect a near zero-sum competitive game (with patronage/connections playing important roles), at more aggregative levels advancement both represents and reflects the accumulation of real economic assets. In relation to metropolitan escalators, this may be evidenced by the capacity of their beneficiaries to transfer acquired occupational advantage to other spatial locations (as Fielding, 1993, implies for London; and as de la Roca and Puga, 2012, show for leading Spanish cities).

In relation to individuals’ progression, there is an important distinction to be made between the capabilities contributing directly to current performance (static human capital) and those which contribute indirectly through the accumulation of further capabilities for future use – or maintenance and replenishment of depreciating ones (dynamic human capital). Beyond the contributions made by luck and prior social connections, occupational advancement may then be seen to depend on a combination of:

• exposure to the superior learning opportunities (that is, experience), which some places and institutions offer; and

• the desire/capacity to make the most of these (that is, dynamic human capital).

Important aspects of human capital always involve not just matters of knowledge (of what, how or who) but of attitude and character too, including: fitness of all kinds (physical, mental and emotional³), energy, patience, loyalty, attention span, capacity to concentrate and/or to endure tedium, etc. This applies especially to the dynamic component, for which an ability to ‘learn’ (whether innate or acquired through ‘learning to learn’ in academic settings) is necessary but not sufficient. What is also required is a positive desire to learn (in relevant contexts), whether because the process is felt to be stimulating/satisfying, or because it is recognized as functional for pursuit of some personal aspirations. Job-related ambition is thus potentially key to the capacity to convert superior learning opportunities into success in terms of occupational advancement.

This notion of ambition is close to McClelland’s (1961) concept of the Achievement Motive, identified as one of the key drivers of human behaviour, alongside the needs for Power and Affiliation – over and above subsistence and material desires. Those with strong drives to achieve were seen as motivated to choose and respond positively to realistic challenges, whereas (at the other extreme) those with a stronger drive to avoid failure would tend to go for either impossible targets, or ones involving no real challenge. Such drives are quite likely to be conditioned both by social background and formal education. But, together with a psychological capacity to take a long-term view, they have important implications for the acquisition of more advanced capabilities and effective exploitation of rich urban knowledge seams.⁴ For these drives to be directed specifically to acquiring capacities of recognized use in remunerative work – in competition with the claims of other domains of personal life – seems also to require perception of a job as of intrinsic importance/interest, and as offering potential for esteem and self-actualization (in Maslow's, 1943, terms).

Opportunities for acquiring tacit knowledge

The other key requirement for substantial occupational progression – beyond that offered after years of blame-free service – is access to suitable learning opportunities for acquiring relevant kinds of tacit knowledge. Tacitness is key because it goes beyond what is generally accessible (and hence of limited worth), because of its close relation to the practical mastery of a craft, and because of the need to share understandings with a group of others (Polanyi, 1958; Sennett, 2008).

Access to (potentially profitable kinds of) tacit knowledge is limited for several reasons. Know-how may be too new or specialized for codification to be economic yet. Or, it may be available only interactively, because it involves the ground of shared understandings required to make codified knowledge intelligible and practically applicable. And, in either case, the ‘guilds’ in possession of it may serve as gatekeepers to protect the rents derived from its use. Since not all such knowledge is equally valuable, moving on up the escalator would involve ‘entrepreneurial’ activity to find means of accessing, filtering and applying (what the current state of technology and tastes identify as) better versions of a still uncertain knowledge.

Generally, employers are not expected to allow their resources to be used to equip workers with transferable skills, except in circumstances such as formal training periods or internships where they can be compensated by paying lower/minimal wages. Becker (1962) developed this argument in relation to formal training, but the same principle should apply to less structured opportunities for learning with a transferable value to the worker. In practice, however, much human capital seems to be task-specific, acquired on the job but with transferable value to the worker,⁵ because broader human capital enhances the returns to specific skills (Oi, 1983). In fact, given a capacity and motive to learn, much task-specific capital will come from generalizing and recontextualizing the combined lessons of firm-specific training and non-routine work situations.

How far employers effectively facilitate the acquisition of such a potentially marketable asset will then depend on the nature of the activities in which a work unit is involved – being most likely: close to the cutting edge of market or technology change; where there is substantial customization of the product/service for quality-sensitive clients; with sophisticated forms of collaboration in providing a complex product; and/or the provision of strategic advice and support to highly placed decision-makers. Places with concentrations of such jobs ought then to provide the greatest opportunities for those with ambition and capacity to learn.

Escalator regions, migration and signalling

The nature of these activities means that many of them are likely to be concentrated in a limited set of high-order centres/city-regions offering superior face-to-face access to sophisticated external information sources, clients, suppliers and collaborators. For this reason alone, leading cities may exhibit stronger rates of occupational progression, at least for those groups with qualifications to access jobs of these kinds. If other work units required similar kinds of labour, but lacked the incentive to offer developmental opportunities, they might simply have to pay more in compensation, or else accept less qualified/ambitious recruits.

That might turn out to be virtually the whole story, with no other significant spillovers. Additionally, however, some such places may also generate substantial externalities through their labour markets. In particular, as Glaeser and Maré (2001) suggested, places possessing both a concentration of escalator portals of this kind, including a range of independent employers with overlapping demands in terms of skill sets, and a flexible, competitive labour market, may more strongly incentivize workers to pursue the available opportunities for on-the-job acquisition of human (and social) capital. In turn, such places are even more likely to attract labour migrants with high levels of dynamic human capital, and consequentially also to reinforce the concentration of work units dependent on a dynamic labour pool of this kind.

The standard neo-classical model of migration (from Sjaastad, 1962) represents it as an act of personal investment, with once-for-all relocation costs being incurred in the expectation of a flow of returns, in terms of higher (real) income over the years. Where the attractor is better employment prospects, the return on this investment should clearly be greatest in the young working ages, with lower flows among the middle aged, and a likelihood of reversal as retirement approaches if the most successful areas have highest living costs. If place-specific elevator processes primarily affect (ambitious) young people, the selectiveness of migration to such places would be intensified, and an earlier reversal of flows may be expected at the point when progression slows down – ‘stepping off the escalator’ (Fielding, 1993). In any case, migration may be motivated by the prospect of a once-for-all occupational uplift (in the elevator), either where lack of local opportunities holds people back in jobs below their actual capabilities (perhaps having down-shifted after a job break) or because an individual's promotion is contingent upon taking up a specific vacancy elsewhere.

In its strong versions (going beyond the simple availability of jobs offering a portal to learning) the notion of an urban/metropolitan ‘escalator’ is not really one of an effortless ride up the occupational scale. Rather, it involves intense competition and effortful learning on the part of ambitious, achievement-oriented workers seeking recognition from knowledgeable, sophisticated and selective employers, patrons or audiences. Some of the flavour of this was captured in the New York, New York context by the lyric asserting (in relation to one of that city's occupational specialities) that ‘If you can make it here, you can make it anywhere … it's up to you’⁶ – with the double implication that the city presents the most demanding challenge, but that success in meeting this yields capabilities of value anywhere.

There is also a symbolic aspect, for both prospective workers and employers, to thronging around such escalators, particularly in the recognized centres. This stems from the special uncertainties involved in identifying: those who possess dynamic human capital that they will deploy on the employer's behalf; and those jobs which will provide the most favourable opportunities for human/social capital development. Recognition of a particular (geographic and trade) sub-labour market as fulfilling an escalator role can be expected to attract to it a share of ‘no-hopers’ (as achievement motivation theory suggests) for whom failure there would involve very little shame, but also a core of those more invested in becoming ‘king of the hill, head of the list, cream of the crop, at the top of the heap’.⁷ The first group should be simply weeded out in a competitive environment. For the rest, however, presentation as a job applicant in these sub-markets,⁸ particularly perhaps as an incomer, signals a degree of self-belief, seriousness of purpose and professional ambition which address key questions about the possession of dynamic human capital. Similarly, on the employer's side: to locate your business unit in a part of the city where the price of access to high-calibre workers is both higher property costs, and active local competition for the pool of talent the place attracts may be seen as a signal of intent to provide experiences that enable workers to develop proactively capacities with a marketable value. An intent to signal in this way would add force both to the self-selection of ambitious workers and progressive job opportunities, that together represent a weak (purely compositional) version of the escalator, and to the dynamic arising from their interaction within a thick, competitive labour market, representing the strong version of the escalator hypothesis.

Implications for the geography of occupational progression (and regression)

This sketched theoretical framework makes occupational advance primarily a function of on-the-job learning, stimulated (to varying degrees) by a combination of personal and contextual factors. On the one side: job-related ambition is seen to play the key role, while, on the other hand, it is access to supportive types of job opportunity plus more dynamic urban labour markets. If these forces are powerful enough, mutually reinforcing spatial selection on both sides of the labour market could lead a few places (maybe just one within a national economy) toward strong escalator roles that increasingly differentiate them from all others – and boost overall rates of advance. The question is how much of this idealized model is to be observed in practice, rather than just providing a modern version of the Whittington myth in a world where dominant cities simply attract ambitious people and high status jobs.

The labour market involves snakes as well as ladders, however, and some of these may also be concentrated in particular cities or regions. Accidents (both favourable and unfavourable) quite commonly place people in job niches imperfectly matched to their capabilities – with some chance of an early correction of the discrepancy, but (failing that) a gradual adaptation of capabilities (whether up, down or sideways) toward expectations. In the case of forced job losses, as one key type of accident, the likelihood of recovery will depend on the current tightness or slackness of the local labour market, with slack labour markets inducing a more extensive ‘bumping down’ process (à la Reder, 1955) as the downwardly mobile displace others. Some who mitigate their position through migration to a stronger labour market may nevertheless appear among those – recognized by Fielding (1993, p. 156) as moving from positions of weakness rather than strength – who do not visibly gain from moving to the escalator region. More generally, it would be expected that, even over the long run, all those with spells of interruption of employment were liable to show inferior occupational trajectories. A consequence would be a geography of generally weaker status progression and human capital development in regions exposed to substantial job losses – which is not simply the obverse of the expected upward movement in escalator regions, but needs to be controlled for when evaluating hypotheses about processes operating there.

The following sections of this paper partially address these hypotheses, with a focus on two empirical issues: about the evidence (after control for personal and job attributes) for the existence of place-specific pure ‘escalator’ effects in one or more metropolitan centres; and for the concentration of such effects on particular subgroups within the population, including those with stronger job-related ambition. Other issues are being addressed separately in papers with collaborators, using complementary data sources – including fuller treatment of the question whether any effects identified here as specific to places may just reflect more favourable opportunity structures.

METHODS/OPERATIONALIZING

In empirical terms, the analytic approach of this paper involves: charting the progress of people up (or down) an employment ladder linked to achieved/recognized levels of human capital; and then examining how variations in such progress are associated with personal attributes, areas of residence (and/or migration) and interactions between these. The questions to be addressed follow on from those of Fielding (1993), though being framed in more general labour market terms rather than as issues of class transition. Correspondingly, the basic methodology pursued here is one of conventional regression modelling with a continuous measure of progression, rather than tabular or statistical analyses of cross-classified categorical variables.

There are a number of important issues to be considered when applying this simple approach to analyses of the escalator processes discussed in the last section. In particular, these involve: data sources (and their limitations); variable definition (in relation both to the dependent variable and to the concept of ambition); and model specification (particularly the treatment of dynamics and disequilibrium positions).

Data sources

Several micro-data sources – including the (Census-based) Longitudinal Study (LS), the Annual Survey of Hours and Earnings (ASHE), the Labour Force Survey (LFS), and the British Household Panel Study (BHPS) – offer relevant data on the occupational progression of individuals in the United Kingdom. Each has distinctive strengths (and limitations) for some of the key questions about functioning of the escalator process, and both the LS and ASHE are being employed in forthcoming papers from this project.⁹ This paper, however – with its focus on hypotheses about the role of ambition as a form of dynamic human capital – concentrates on evidence from the BHPS. Uniquely (among UK data sources) this combines qualitative personal/family data, including the basis for a measure of work-related ambition, together with long-term work histories, from which shifts up and down the occupational ladder can be assessed, and these linked to areas of work/residence. The core analyses using this source are complemented by more descriptive analyses of occupational position and short-term change, using the very much larger sample provided by the LFS.

The BHPS micro-data files used here¹⁰ cover the years 1991–2008, with successive annual waves of this survey collecting data from a panel of respondents incorporating all base year members of an original sample of households. In each wave there are a repeated series of questions, including ones on employment status and experience over the year and on qualifications as well as on area of residence, etc., which are supplemented by others on specific topics that are one-off or repeated irregularly. Among these is a pair of questions on attitudes to jobs/work which are used here to construct a measure of individuals’ job-related ambition.

In order to make use of this (one-off) variable, the focus here is on change in job status over the whole period (for respondents in employment at both ends), rather than on year-to-year dynamics.¹¹ The full annual panel is used only in a supplementary role to generate measures of accumulated time spent in/out of employment, and in specific regions. Requirement of a full set of responses across eighteen waves in order to generate these measures, in a survey with some discontinuities as well as attrition in responses, together with the age and employment restrictions which were imposed, inevitably reduced the available set of observations rather substantially, leaving just 868 cases for the final analyses presented below.

Operationalizing key variables

Job status (JS)

In contrast to previous ‘escalator’ studies which have taken either transitions between discrete social classes (Fielding, 1989, 1995) or increments to actual earnings (Glaeser and Maré, 2001; de la Roca and Puga, 2012) as their outcome indicator, the dependent variable in the present analyses is designed to provide a continuous measure of occupational position in terms of the value of human capital associated with the job type.

Since human capital encompasses many different types of attribute, with relative values that can shift in response to changing sectoral mixes, technologies and educational trends¹² – any composite measure needs to be substantially grounded on evidence of (market-determined) pay differentials.¹³ The concern of this paper is not with tracking which skill types are becoming more or less valuable, but rather with assessing movement up and down a predefined occupational scale, using average market valuations for a period to place different ‘job niches’ (defined in terms of occupation and managerial/supervisory status) on that scale.

In practical terms, occupation is defined in terms of the 371 units of the SOC90 occupational classification, while four levels of managerial/supervisory responsibility were also recognized (managers in establishments with twenty-five or more employees, managers in smaller establishments, supervisors and other workers¹⁴). Out of the 1484 logically possible combinations, 1197 such niches were actually represented (with earnings data) in the pooled 1993–2000 rounds of the UK LFS used to generate the JS scale values.¹⁵ These scale values were derived from logged hourly earnings¹⁶ – available for employees only but applied also to the self-employed. After (regression-based) adjustment of individual earnings data for three contextual factors (year, workplace region¹⁷ and establishment size¹⁸), values for each niche were computed on two bases: as simple averages; and as predicted values from regressions of adjusted log earnings on occupation and status dummies. In order to mitigate problems of sampling error in the less common niches, these two estimates were combined, with weights reflecting the computed standard errors of each.¹⁹

JS scale values were then applied to BHPS cases, using its records of occupation and managerial/supervisory responsibilities. Survey design, data collection and coding procedures should make these measures comparable between the two surveys (LFS and BHPS) and across waves of each.²⁰

Ambition

For an indicator of individuals’ potential motivation to seek and acquire on-the-job human capital, the study made use of a couple of questions (each appearing in both the 1991 and 1999 waves of the BHPS) about the two most important aspects of, firstly, work in general, and, secondly, of particular jobs. For a composite measure of ‘job-related ambition’, responses were combined that highlighted promotion, career, initiative and job content (scored positively) as against those citing people's company or money for essentials (scored negatively) – and discounting other references to pay.²¹

In its raw form this measure showed a clear, but ‘U’-shaped, relation with age, which partially explained a relatively weak correlation in raw scores between the two years. This age factor was controlled for and the scoring converted to a 0–1 scale. In order to derive a measure applicable to the 1991–2008 period, the age-standardized measures from the 1991 and 1999 surveys were then simply averaged.²²

Model specification and estimation

In principle, the analyses of individual progression relate JS change measures to a series of independent variables, involving location, prior education, etc. However, a consistent feature of all the analyses with different sources and time intervals varying between one and seventeen years is that change seems to be substantially affected by JS level at the start of the period. When this variable is included in the analysis, it consistently attracts (negative) coefficients, significantly below 0 and above –1.

Similar findings are familiar in wage equations (for example, Blanchflower and Oswald, 2005), but in the context of JS changes there appear to be two distinct kinds of explanation requiring different kinds of response. On the one hand, there is a clear possibility of significant measurement error, affecting both JS levels and change, in which case inclusion of the lagged dependent variable could be a source of bias in the analyses. This might arise either because of the open-ended way in which occupations are initially reported (however systematic the coding of these) or because of insufficient recognition of occupational heterogeneity in one or both tails of the distribution, causing an artificial truncation of the scale. On the other hand, it could be well be that some real processes are in play involving a ‘reversion’ in JS scores. This might be either toward a ‘mean’ consistent with an individual's real capabilities, from an initial niche out of line with these; or toward zero, reflecting a natural tendency for stocks of human capital to depreciate, in the absence of positive inputs from learning experiences that renew or upgrade these. In this case it is omission of lagged values from the right-hand side which would be liable to bias other estimates.

A series of experiments was undertaken to assess the potential influence of these different explanations. JS scores were averaged over several waves of the BHPS to see how far this weakened the relationship between change and lagged JS values, as would be expected with random measurement errors (or atypical short-term jobs). The length of the observation period was also reduced to see whether this also served to attenuate the relationship, as would be expected from a systematic tendency to reversion. And the lagged JS score was replaced with a set of quantiles to see how truncation effects might be responsible.

In fact, though analyses with larger datasets (over shorter intervals) did suggest that truncation played a role, there was no real evidence of this with the BHPS. Support was, however, found for both of the other hypotheses, with their contradictory implications as to what should be done. The compromise solution eventually adopted involved averaging the first five waves of the survey to define the baseline JS level, down-weighting this by 20–30% when calculating JS change, and omitting JS levels from the right-hand side of the regressions.

ANALYSIS AND RESULTS

Short-term occupational progression

Preliminary regression analyses were undertaken with the LFS, which has a very much larger sample of workers, enabling finer geographic distinctions to be made, though it only has short-term (one year) information on changes in position, and it lacks attitudinal data. These served to confirm that, as Fielding had indicated, if there was an escalator city/region in the UK, it was within the Greater South East (GSE) – that is, London and its two adjacent regions. But they suggested that the relevant job opportunities were heavily concentrated in its core. Simple analyses of JS levels²³ for respondents to the 2004–2008 rounds of the LFS showed that, for given demographic and educational characteristics, JS levels among Central London (that is, central business district – CBD) workers were some 13% above the national average as compared with 3–4% in the rest of London, 2% in outer parts of the GSE, and 1% in the Birmingham and Manchester conurbations – though below average in the other five conurbations.

A more restricted analysis of JS changes between second quarters of 2008 and 2009 (the results are presented in Table 1) identified a similarly large differential in progress (of 18%) for Central London workers, though in this case it was specifically associated with moving residence into the GSE during the year. The basic results here are presented in column 1 of Table 1, with the other columns confirming their robustness to omission of the lagged dependent variable (which dominates the basic regression) and exclusion of the majority whose job did not change.²⁴ Overall, it seems that young people and those with higher education were more likely to progress. In relation to geographies, there is evidence of a broad north–south contrast, with a modest but significant difference of 0.7% in occupational advancement between residents of southern England and those from the north and west. This is evidence of an escalator effect – involving continuing learning and progression – since it essentially relates to people who stayed within one of these two super-regions. It might possibly be stronger among Central London workers, though the evidence on that is not clear here. As a continuing advantage to the average resident it is by no means trivial – but not of anything like the scale of gains made by interregional migrants with Central London jobs. These have then to be interpreted as predominantly reflecting a large one-off elevator effect, from moving to a place with superior opportunities to deploy already existing capacities. Other interregional migrants may make some gains of this kind, but they would be of a very much more modest scale.

Table 1. Regressions of single-year job status (JS) change on individual and spatial attributes

	Full sample		JS changers only
Constant	0.122*** (20.3)	–0.006*** (3.0)	1.061*** (29.6)	–0.046 (2.8)**
JS at the start of the year	–0.057*** (29.6)	–	–0.477*** (34.4)	–
Age less than fifty years	0.004* (2.2)	0.004* (2.5)	0.002 (0.9)	0.037* (2.5)
Higher education	0.029*** (15.2)	0.005** (2.7)	0.196*** (16.1)	0.036** (2.6)
Resident in Southern England	0.007*** (4.1)	0.005** (3.1)	0.002 (0.9)	0.038** (3.0)
Interregional mover plus working in Central London	0.186*** (4.3)	0.183*** (4.2)	0.789** (3.0)	1.085*** (3.7)
Other Central London worker	0.006 (1.3)	–0.009 (2.0)	0.054 (1.9)	–0.063* (2.0)
Other interregional mover	0.017 (1.7)	0.017 (1.7)	–0.012 (0.3)	–0.032 (0.7)
Adjusted R^2	0.026	0.001	0.212	0.007
N	34 230	34 230	4564	4564

Note: The dependent variable is JS change since the reported occupation and status twelve months previously. Interregional movement is over the previous year, between standard regions, excluding moves between parts of the Greater South East (GSE). Southern England comprises London, East Anglia, East Midlands, South East and South West. Figures in parentheses are t-statistics; asterisks indicate significance: *5%, **1% and ***0.1%.

Source: Analysis of the Quarterly Labour Force Survey (QLFS), 2nd quarter 2009.

This preliminary evidence on short-term changes echoes Fielding's earlier findings and argument in pointing to strong links between upward social mobility and migration into the GSE. But this effect now seems to be specific to Central London job opportunities – and in that context to be predominantly a matter of once-for-all (elevator) gains from migration rather than continuing movement up the escalator.

From these short-term observations, however, one clearly cannot tell:

• what kind of qualities were deployed by the migrants in pursuing and then securing higher status jobs in Central London;

• how these might feed into superior long-term progress as compared with others within that labour market; nor

• what part of the non-migrant population within the escalator region shares the capacity to take advantage of superior opportunities for progression.

To follow these up the paper turns to the longer-term change (and attitudinal) data available from the BHPS. The analyses here first relate job changes to a combination of personal attributes (including job-related ambition, and qualifications) and location/relocation of residence, and then explore the geography of ambition and qualification.

Long-term occupational progression

Two initial sets of exploratory regressions were undertaken in order, firstly, to identify the personal attributes and regional distinctions that actually mattered and, secondly, to find the most appropriate form for the dependent variable. Among the independent variables, these analyses made it clear that age, degree-level qualifications and job-related ambition were all salient, whereas gender, race and religion were not – and that the really significant regional distinction was between the GSE and the rest of the UK. With the aim of distinguishing between potential escalator and elevator effects, regional indicators were defined both in terms of the proportion of the period spent in a region (representing exposure to any escalator processes), and also in terms of interregional shifts between the start and end of the period (to pick up any elevator effects). For the dependent variable, an adjusted JS change measure was defined which set the baseline in terms of the average of recorded JS scores for the first few years (in order to mitigate measurement errors) and it discounted these by a factor (intended to allow for reversion toward the mean or bottom). These decisions were guided by effects on the root mean square error (RMSE) for a preferred model specification, leading to choice of a five-year average for the baseline (as the point of diminishing returns to adding further years) and a 28% discount factor (as the value minimizing the RMSE).

Regression results with this adjusted job change measure, for what is effectively a fifteen-year period (1993–2008), are presented in Table 2. A basic model (in column 1) shows that occupational progression was strongest for young people and those with degrees (with additional one-off gains for those acquiring a qualification at this level during the period), while those who had significant spells out of employment suffered big setbacks, especially if they were older. There is also evidence here that those with higher levels of job-related ambition did significantly better. As far as geography is concerned, the results suggest that continuing residence in the GSE led to rather stronger progress – not quite significant statistically at the 5% level, but equivalent to about half the value of a degree in terms of long-term advancement. None of the migrational variables appears to have any real effect, however: for interregional migrants as a group, the suggestion is that they did a bit worse than stayers, with no indication that those moving toward London or the GSE did any better, and the more positive sign for those moving away from London is far from being statistically significant.

Table 2. Regressions of long-term job status (JS) change on personal and spatial variables

	(1)	(2)	(3)	(3a)
Age range (years), 1991	22–46	22–46	22–46	22–28
Constant	1.809*** (11.4)	1.748*** (9.8)	1.839*** (11.8)	0.593 (0.7)
Age	–0.364*** (8.0)	–0.338*** (6.7)	–0.361*** (8.0)	0.032 (0.1)
Ambition
Scale	0.146* (2.5)	0.063 (0.9)
Upper 60%			0.013 (0.6)	–0.044 (1.0)
Degree
At the start of the period	0.074** (3.0)	0.069* (2.2)	0.081*** (3.4)	0.089 (1.9)
Gained during the period	0.151*** (3.9)	0.153*** (3.8)	0.156*** (4.1)	0.225*** (3.2)
Proportion of time out of work
Aged less than thirty-two years in 1991	–0.631*** (4.1)	–0.639*** (4.1)	–0.618*** (4.0)	–0.536** (2.8)
Aged thirty-two or more years in 1991	–0.910*** (3.9)	–0.927*** (4.0)	–0.971*** (4.1)	–
Proportion of time spent in GSE	0.035 (1.8)	–0.081 (1.6)
Interacted with: less than thirty-two years in 1991	–	0.037 (1.0)
Graduate	–	0.009 (0.2)
Ambition: scale	–	0.271* (2.1)
Bottom 40%	–	–	–0.017 (0.6)	–0.056 (1.0)
Upper 60%	–	–	0.067** (2.7)	0.186*** (3.3)
Interregional migrant	–0.026 (0.8)
Toward London/GSE	0.002 (0.0)
From London beyond the GSE	0.042 (0.5)
Standard error (SE)	0.063	0.062	0.062	0.070
Adjusted R^2	0.135	0.141	0.141	0.096
N	868	868	868	238

Notes: Dependent variable is calculated as: JS score for 2008 less 0.72*mean of JS scores for 1991–1995 – for respondents in employment and with recorded occupation at each date.

GSE, Greater South East.

Source: Micro-data were from the British Household Panel Study (via the Economic and Social Data Service).

A second model discards these migration variables and concentrates on seeing whether there was evidence of a stronger GSE escalator effect among subgroups who might be particularly suited to exploiting its opportunities. Three hypotheses – relating to the young, to graduates and to those with higher levels of ambition – were tested by interacting the relevant variables with the measure of time spent in the GSE. Results (in column 2 in Table 2) suggest no particular advantage for graduates, an averagely positive (though not significant) effect for those under thirty-two years of age at the start of the period, but a clearly stronger effect of GSE residence for those with higher levels of ambition. For those lacking any of these attributes, the suggestion (from the coefficient on the simple measure of time spent in this region) is that they did less well in the GSE than elsewhere. This is a bit misleading, however, since no one in the sample actually received a zero score for job-related ambition – but it does seem to be the case that the average middle-aged person with low levels of ambition could expect no benefit from a GSE escalator, beyond the modest gains they might enjoy elsewhere.

This finding suggests the possibility that such benefits might actually be quite heavily concentrated among a small minority of highly ambitious people enjoying very much faster progression within this opportunity-rich region. Experiments replacing the scalar measure of ambition with a set of quantile dummies indicate, however, that those with very high levels of ambition did not gain much more than those with median levels – the real contrast being between the latter and those at the bottom end of the distribution. This can be captured roughly by a division between the bottom 40% and the upper 60% of observations on the ambition scale. A third model thus replaces all the other GSE residence-related variables with a pair which interact (the proportion of) time spent in the GSE with dummies for these two groups of less/more ambitious people, applying this first to the full sample (column 3 in Table 2) and then to a subgroup of the youngest (column 4). The results now show clearly significant gains (averaging 6.3%) from continuing residence in the GSE for the more ambitious group, and none for the less ambitious. When this analysis is broken down by age, these gains seem to be heavily concentrated among the youngest (those aged between twenty-two and twenty-eight years in 1991), with average increment to their JS level of 16.6% over the period. Though not shown here (since the GSE effects are not statistically significant), the corresponding results for those in the older (29–47-year-old) group suggest gains to the more ambitious group from staying in the region which are only about one-fifth of those enjoyed by the younger cohort. Given the large overlap in ages between members of the two groups during the observation period, this contrast implies that escalator effects may only really apply while people are in their twenties and early thirties.

Social and spatial influences on job-related ambition

These results point to a very strong interaction between job-related ambition and residence in the escalator region (that is, the GSE in this case), both in the sense that only the more ambitious seem to benefit from the escalator, and also in that access to this escalator seems necessary for this group to convert their ambition into faster progress in the job market. But there is also a question as to how far ambition itself may be unevenly developed among residents of different regions. This cannot be explored in much depth with the BHPS since the relevant questions were only asked in two waves of the survey, and the derived measure clearly contains quite a lot of noise. Nevertheless, patterns and associations in the data provide some basic evidence about the geography of ambition and its role in reproducing social positions.

On a cross-sectional basis, individuals’ scores on the index of job-related ambition are most clearly (and positively) associated with their levels of academic achievement, and with their father's social class²⁵ (both directly and also indirectly, via educational outcomes which are linked to parental class). There is also a statistically strong, though substantively much weaker, association with gender – women recording significantly higher levels of job-related ambition.²⁶ Marital status also seems to be a factor, with the never married displaying significantly higher levels of ambition, and with entry into couplehood as one of the clear influences (together with acquiring additional qualifications) on changes in measured job-related ambition between the 1991 and 1999 waves. This looks like a question of trade-offs between priorities – and, since establishing a new household commonly involves extra financial demands, it tends to support the idea that the constructed ‘ambition’ measure relates to occupational achievement, rather than simply pursuit of higher earnings.

Spatial variables all seem to have weaker links to ambition, though ambition levels do tend to be rather higher among residents of the GSE. On what now seems the most salient measure – based on the split between those in the bottom 40% on the ambition scores, or above that – just 32% of GSE residents in 1991 were in the former group as compared with 42% in the rest of the UK. Controlling for qualifications and parental class (both favouring the GSE) only reduces this gap by about one-quarter. It seems that this regional disparity could have come about in three main ways: it could reflect differences in the context of primary socialization and attitude formation; or it could involve a subsequent adaptation to the differential availability of relevant opportunities in the region of residence; or it could reflect selective migration of the more ambitious toward regions in which they could realize their drive for advancement.

Taking the 1991/1999-based ambition classification as a constant reference point and comparing residence inside or outside the GSE for the first and last survey waves (1991 and 2008), it is evident both that a much smaller proportion of the low ambition group actually shifts between these two broad regions (2% as compared with 7%), and that the higher ambition group is entirely responsible for the net shift into the GSE. But even over seventeen years, this shift is quite modest, representing just 4% of the GSE sample, and lowering the share in the low ambition group by only 2% (as compared with the 10% gap to be explained). Looking at the relation between place of birth and region of residence in 1991, however, it is striking that the bias toward the upper ambition group within the GSE then seems almost entirely attributable to people born outside the GSE, whether elsewhere in the UK or overseas. Earlier movement to the GSE seems to have brought it (by 1991) a group of people who whether naturally (through primary socialization) or subsequent adaptation have proved more than usually ambitious – rather than this being a natural advantage of (and for) those born and brought up there.

CONCLUSIONS

Starting from a conception of occupational progression focused on the development of human (and social) capital, this paper has emphasized the crucial contributions of both dynamic human capital (including ambition as well as learning skills) and access to opportunities for acquiring tacit knowledge, together potentially with those urban contexts (in leading agglomerations) which have concentrations of both.

A distinction has been drawn between three processes that contribute to an uneven geography of human capital stocks: bumping down of displaced workers in slack labour markets; an elevator operating at the point of migration for those whose talents have been underemployed in such areas (or whose promotion within a specialist role depends on taking opportunities where they arise); and an escalator offering continuing prospects of upgrading for those workers in the core region who are able and willing to pursue them. Within the last of these a further distinction has been made between two versions of an escalator region hypothesis: a ‘weak’ one simply reflecting compositional effects stemming from a highly selective mix of resident and job types in successful high order centres; and a ‘strong’ one, involving positive interactions between each of these in the context of particularly thick and competitive sub-labour markets.

Empirical investigation with two British surveys confirmed Fielding's observation for an earlier period that some version of the London region stood out in its association with strong progress up the occupational ladder. From observation of short-term changes, the dominant effect appeared to be an elevator one associated with graduates coming to work in Central London jobs. Evidence on long-term changes, however, clearly showed the operation of escalator processes for residents of a much wider GSE region, though essentially confined to young people with at least reasonable levels of job-related ambition. And, consistent with a strong version of the escalator thesis, such ambition appeared only to be rewarded within the context of this region. This is only a partial test of the thesis, since other sources with larger samples are required to examine how much of the escalator effect is simply dependent on a concentration in/around London of job opportunities offering privileged access to cutting-edge stocks of tacit knowledge.

Within this account it is not specifically migrants who benefit from the escalator process, but they are found to include a disproportionate share of the relatively ambitious people who are its main beneficiaries. Job-related ambition is found to be a mediating variable which contributes to inter-generational continuities in relative JS, with the GSE having the biggest share of those likely to produce ambitious offspring. But their strong representation within this region has much more to do with the selective attraction of ambitious migrants, both from abroad and from the rest of the UK.

A key aspect of the version of the escalator thesis developed here is that it makes a substantial part of the rewards derived from operating within the leading city-region contingent on (effective) efforts by the workers concerned to develop their capacities. Together with the specificity of job opportunities to which migrants are attracted, this can help to explain how, even with strong spatially equilibrating forces operating in the labour market, large inflows of migrants continue to be attracted to the core region. These serve to reinforce measured differences in economic performance between north and south – but also serve to equalize life chances for at least the more ambitious half of the populations who start life on one side or other of this ‘divide’.

Acknowledgements

Earlier versions of this paper were presented at the European Regional Science Association (ERSA) Congress in Jonkoping, Sweden, and at seminars held at Reading University and the London School of Economics (LSE). The author is grateful to the participants at these sessions for helpful comments on this work; and to Nick Buck for comments made on a much earlier attempt at the same topic. None of these is responsible for the paper's shortcomings. Data from the UK Labour Force Survey (LFS) and from the British Household Panel Survey (BHPS) were supplied by the UK Data Archive. Neither the original collectors of these data nor the archive bear any responsibility for the analysis or interpretations presented here. Research for this paper was supported by the UK Economic and Social Research Council (ESRC) via its Spatial Economics Research Centre.

Notes

1. These labels for the two processes reflect the implied time profiles of change in occupational status: with a simple vertical uplift in the escalator case; and movement on a continuing upward slope in the escalator case. This (more natural) usage differs from that of Findlay et al. (2009) who applied the ‘escalator’ term to promotions that do not require a residential move, and ‘escalator’ to ones that do.

2. Formally such items of human capital which are wholly embodied are to be distinguished from elements of social capital which involve the active participation of others. In the (micro-level) context of the present paper, that distinction is neither crucial nor very clear, since key skills often involve managing social interactions, and both involve context-dependent elements.

3. Including the ‘erotic capital’ to which Hakim (2010) has recently drawn attention.

4. Though aspects of McClelland's own empirical work were challenged, and this social psychological approach to mobility studies later lost out to more structural perspectives, the classic socio-metric modelling incorporating such factors did find ‘that the influence of education on occupation is mediated by motivational as well as cognitive and institutional factors’ (Duncan and Featherstone, 1972, p. 121).

5. Accounting, on Gathmann and Schönberg’s (2010) estimates, for upwards of 25% of wage growth overall, and at least 40% of (the more substantial) growth among university graduates.

6. Or in the original Fred Ebb lyric of 1977, for a singer ‘leaving today’ from his ‘little town’: ‘If I can make it there, I'll make it anywhere. … It's up to you, New York, New York.’

7. Same source as the previous note.

8. Which might only be a relatively small part of the general labour market, even in a high-order centre.

9. With Mike Coombes and Tony Champion.

10. This survey was undertaken by the Institute for Social and Economic Research and distributed via the UK Data Archive, both at the University of Essex, Colchester.

11. True longitudinal analyses are being undertaken separately using the ASHE dataset to investigate other issues about the escalator region hypothesis.

12. Among other factors, including migration policies and the social attractiveness of different kinds of work.

13. This approach offers an objective basis for measurement in the sense that it reflects the evaluation of worker capacities emerging from the product and labour markets in which they are involved, rather than academic or administrative preconceptions. From a social rather than an individual perspective several qualifications need to be borne in mind: market evaluations are contingent on the current distribution of wealth and educational capital in the society; even within generally competitive markets the actual pattern of rewards will be affected by market failures due to the incidence of monopoly power, misinformation, etc.; and (as a referee suggested) some occupational skills may be valued for their contribution to such market imperfections.

14. This categorization is based on answers to questions about such responsibilities, applied independently of employee/self-employed status.

15. To limit the duplication of cases (among people surveyed over five successive quarters), responses from only one of the four quarters were used for each year, yielding a total of some 67 000 usable observations for the calculation of the JS scores.

16. Because logged values were used for all calculations, ‘averages’ all relate to geometric means.

17. In fact, just three regions – London, Eastern/South Eastern, and the rest of the UK – differentiated on the basis of housing costs, as an independent influence on earnings.

18. Distinguishing those in establishments with 1–10, 11–24, 25–49 and more than 50 workers.

19. Standard errors for the simple averages were computed on the basis of simple random sampling and the number of cases observed in the niche, while those for the regression estimates were simply derived from the standard error of the estimate, which (in relation to true population values) was assumed independent of the number of cases in a cell. Weights were then computed to minimize expected error in the synthetic estimates. In the largest cells, the simple averages were given weights over 0.9, while for the smallest they were below 0.1; for cells with fifty cases in the sample the two estimates received equal weights.

20. Where the same edition of the Standard Occupational Classification was applied. This includes the 1991–2000 rounds of the LFS, but all waves of the BHPS.

21. Along with those to security, hours and relations with managers. For the form and scoring of the two questions, see Appendix A.

22. As will be noted in the analysis section, explorations of differences in values between the two surveys suggested some real shifts associated with changes in marital status and qualification levels – though not with job status.

23. These were computed using a parallel methodology to that described in the third section for the BHPS dataset, though benchmarked with later earnings data and occupational classification.

24. The analyses reported here rely on retrospective reporting of baseline jobs, etc. twelve months previously, taken together with reports of current positions from a single wave of the LFS, rather than using its longitudinal element to compare responses made in two separate waves. The former were judged to produce more reliable JS change measures because of the potential for inconsistent recording and coding of occupations in the latter case (which suggested that 35% of respondents changed niches over the year, as compared with just 15% from the retrospective data).

25. Based on their occupation when the respondent was aged fourteen.

26. The fact that this is not translated into higher rates of JS advancement seems to reflect the much greater incidence of periods of non-employment among women.

APPENDIX A: BRITISH HOUSEHOLD PANEL STUDY (BHPS) QUESTIONS USED TO CONSTRUCT THE INDICATOR OF JOB-RELATED AMBITION

The most (and second-most) important reason for working:

Follow my career; [scored positively].

Essential foods, etc.; and People's company [both scored negatively].

Working is normal; Enjoy working; Money for extras; Earn money for self; and Other reason [all unscored].

The most important (and second-most) aspect of a job:

Promotion prospects; Using initiative; and Actual work [all scored positively].

Total pay; Good relations with manager; Job security; Hours worked; and Something else [all unscored].