html5-img
1 / 33

‘ Have earnings polarised in the UK?

‘ Have earnings polarised in the UK?. Craig Holmes Pembroke College, Oxford University and SKOPE ISER, University of Essex, October 21 st 2013. Outline. Background on the hourglass labour market Polarisation and earnings distributions – some theory

keahi
Download Presentation

‘ Have earnings polarised in the UK?

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. ‘Have earnings polarised in the UK? Craig Holmes Pembroke College, Oxford University and SKOPE ISER, University of Essex, October 21st 2013

  2. Outline • Background on the hourglass labour market • Polarisation and earnings distributions – some theory • Methodology – earnings distribution decomposition • Data • Decomposition results • Discussion and future work

  3. Background • Routinisation hypothesis (Autor, Levy and Murnane, 2003): • Refinement of SBTC - technology related to tasks, not skills • Routine tasks substitutable for computer capital • Growth in non-routine jobs, decline in routine jobs • Polarization hypothesis (Goos and Manning, 2007) • Routine occupations found in middle of income distribution • Non-routine occupations found at top and bottom of distribution • Ranking of jobs based on initial wages

  4. Background • Goos and Manning (2007) – 1979-1999:

  5. Background • Similar results observed in: • US (Autor, Katz and Kearney, 2006; Caranci and Jones, 2011) • Germany (Spitz-Oener, 2006; Oesch and Rodríguez Menés, 2011) • Spain and Switzerland (Oesch and Rodríguez Menés, 2011) • Across Europe (Goos, Manning and Salomons, 2009) • Other explanations have been put forward: • Offshoring • Growing wage inequality and demand for services

  6. Background • Wage inequality in the UK has risen since the 1980s

  7. Background • Earnings growth by percentile, UK

  8. Background • More high-wage and low-wage jobs  More inequality

  9. Background • Two main research questions: • To what extent has the shift towards non-routine employment decreased the number of middle wage jobs / increased wage inequality • Why, given that, has earnings distribution polarisation halted since mid 1990s?

  10. Polarisation and earnings distributions • The shift away from routine work should increase the number of high-wage and low-wage jobs, everything else being equal

  11. Polarisation and earnings distributions • However, wage structure of occupations unlikely to remain constant • Autor, Katz and Kearney (2006) – relative wage of routine occupations falls • “Wage polarisation” – a US phenomenon? • Wage differences between different non-routine occupations (Williams, 2012) • Other compositional changes – more educated workforce, lower union membership, greater female participation • Non-uniform increase in demand for non-routine tasks? • Change in returns to other characteristics

  12. Methodology • Standard quantileregressions compute quantiles of a distribution conditional on explanatory variables • However, we need to decompose unconditional distributions • Firpo, Fortin and Lemieux (2009) – two stage approach • Estimate a counterfactual distribution via reweighting initial distribution • Use re-centered influence functions to estimate distributional statistics (such as percentiles) as a linear expression of main explanatory variables

  13. Methodology • Data: • N observations, N0 from initial distribution, N1 from final distribution • Ti = 1 if from final distribution, i = 1,...,N. Pr(Ti) = p • Data can be reweighted • Reweighting: • where p(X) = Pr (T=1|X)

  14. Methodology • This counterfactual can be used to decompose wage and composition effects of a distributional statistic: • An recentered influence function of v(F) measures its sensitivity to each observation, where E(RIF) = v(F) • Assumea linear projection of RIF onto X: • where j = {0, C, 1}

  15. A quantile regression approach • Hence: • This is a more general case of the Blinder-Oaxaca decomposition, where v(F) is the mean.

  16. Data • Family Expenditure Survey, 1987-2001 • Around 10,000 observations each year • Usual gross pay and usual hours of work • Education – year left FT education  four levels • Union membership – subscription fees>0 • Quarterly Labour Force Survey, 1994-2007 • Around 150,000 observations each quarter (5 quarter membership) • Gross hourly pay directly reported • Educational qualifications • Union membership directly reported

  17. Data

  18. Composition and wage effects • FES, 1987-2001:

  19. Composition and wage effects • LFS 1994-2007:

  20. Composition and wage effects • Both periods find compositional changes decreasing the number of middle-wage jobs • Wage structure changes reverse this – partially in the 1987 and 2001, and completely between 1994 and 2007

  21. Individual composition effects

  22. Individual composition effects

  23. The wage structure - aggregate

  24. The wage structure - occupations

  25. The wage structure - education

  26. Discussion • Wage structure reduces increase in inequality, despite change in composition on the workforce • However, hard to interpret as educational or occupational opportunities pulling the middle up – despite the “room at the top” mindset of policymakers • An alternative interpretation – downward sloping wage structure is a ‘correction’ of compositional changes – not as many people in high wage jobs as we’d predict

  27. Discussion • Minimum wage has clearly having an effect at low-end • However, other wage structure effects also helping to reduce lower-tail inequality e.g. male-female wage gaps, relative pay of service jobs

  28. Discussion • Increasingly heterogeneous occupational groups

  29. Discussion • Unrelated to educational attainment? Graduates only:

  30. Discussion • This could reflect a supply problem if it reflects quality of graduates • Could also reflect suitability of university route into labour market vs. vocational education • Can not ignore changes on the demand side – in particular, are technology and skilled labour always complements? • Brown, Lauder and Ashton (2011): • “Knowledge work”  “Working knowledge” • “Digital Taylorism” – deskilling of high skill work • “War for Talent” – high premium paid for small pool of graduates at top universities

  31. Discussion • So far, defined high wage job as a fixed multiple of median pay • However, the size of the wage spectrum is relevant • Very high pay is increasing, even while high pay remains constant • Some higher paid workers move closer in relative terms to the middle as the top experience very rapid wage growth • What is an appropriate cut-off for these groups?

  32. Conclusion • Main points: • Earnings distributions polarised during 1980s and 1990s • This polarisation was less than compositional shifts would have predicted • Middle paying jobs stop declining in middle of 1990s, despite continuation of compositional shifts • Areas for future investigation: • Educational attainment and occupational demands do not seem to be offering opportunities to narrow earnings inequality • What happens to occupational mobility, particularly at low end? • What has happened since 2007?

  33. Contact Details Craig Holmes Pembroke College, Oxford, and ESRC Centre on Skills, Knowledge and Organisational Performance (SKOPE), Email: craig.holmes@pmb.ox.ac.uk

More Related