1. Presenter’s Name & Job Title | Date Capital accumulation, labour market institutions and unemployment in the medium run Engelbert Stockhammer* & Erik Klär** *Kingston University, **University of Trier

2. 2

3. 3 Motivation  “broad movements in unemployment across the OECD can be explained by shifts in labour market institutions”  Nickell et al 2005 (EJ), 1  “The theory can be summed up by saying given the psychology of the public, the level of output and employment as a whole depends on the amount of investment.”  Keynes 1937 (QJE), 221

4. 4 Overview  Mainstream view: Medium-run (‚natural‘) unemployment is primarily due to protective labour market institutions (potentially in interaction with adverse macroeconomic shocks)  Empirical validity called into question recently  OECD (2006) much more nuanced compared to its previous position (e.g. OECD Jobs Study 1994)  Howell et al (2007), Baccaro & Rei (2007)  Back to the question: What are causes of medium-run unemployment?  Contribution of the paper  estimates u(LMI, MS,  K) in 20 OECD countries  in a medium-run panel (5-year-averages)  using two LMI data sets and standard macro shocks  and capital accumulation (  K/K)

5. 5 Outline  Motivation, theoretical background  Empirical literature  Results  Bassanini-Duval (BD) dataset  BGHS dataset  Interactions among LMI  Illustrating contributions  Capital stock or investment?  Conclusion

6. Theoretical background  Basic theoretical vision  Actual unemployment is determined by demand, most importantly investment expenditures  At any point in time there is a well-defined NAIRU (= a short-run Phillips curve)  NAIRU is neither exogenous nor a strong attractor for actual unemployment  NAIRU gets dragged along with actual unemployment  Based on the work by Rowthorn, Arestis, Sawyer  Related papers:  Stockhammer 2008 Metroeconomica  Stockhammer 2011 Oxfd Rev Ec Policy

7. 7 Standard NAIRU model Demand (y) (goods markets) NAIRU (U N ) Actual Unemployment (u) ΔpΔp LMI

8. 8 Keynes‘ approach Demand (y) (goods markets) Actual Unemployment (u) ΔpΔpWS Capital accumulation (ΔK) Interest rates „animal spirits“

9. A modern Keynesian approach 9 Demand (y) (goods markets) NAIRU (U N ) Actual Unemployment (u) ΔpΔpWS LMI Capital accumulation (ΔK) Interest rates (i CB )

10. A NAIRU model  Wage setting  Price setting  Phillips Curve  Unemployment  NAIRU  Is the NAIRU endogenous or exogenous?  Medium-unemployment

11. NAIRU endogeneity  Standard arguments: deskilling, stigmatising...  Endogenous wage aspirations: workers (and the unemployed) regard wage of other workers as „normal“ and/or get used to current wage level  Difference to NK persistence: Not weak wage effects of long- term unemployed, but a shift of the reference wages („normal wage“) (Skott 2005, Stockhammer 2008)  Capital stock: imperfect substitution  Capital stock: increased K (for given Y) reduces price setting power of firms (Rowthorn 1979, 1995)  Profit claims / mark up depends on (long-term) interest rate (Hein 2008) 11

12. 12 Review of the empirical literature  Large mainstream literature on LMI and unemployment  (Nickell 1997, Blanchard & Wolfers 2000 (interaction of LMI and macroshocks), IMF 2003, Nickell et al 2006)  Two principal Keynesian criticisms:  LMI explanation does not actually work empirically (Howell et al 2007), Baccaro and Rei (2007)  Capital accumulation matters (Rowthorn 1995, 1999, Stockhammer 2004, Palacio-Vera et al 2006, Arestis et al 2007)  Demand (and hysteresis)  Limited substitutability (Rowthorn 1999)  Usually not capital accu and LMI in the same study  Exceptions: Stockhammer 2004, Arestis et al 2007  But: time-series = only limited LMI set, small range of countries

13. Overview selected empirical studies

14. 14 Classification of the Literature Econometric approachLMI controlled for Macro shocks controlled for TSDyn.PSta.PNonesomemanynonesomeK Nickell 1997 XXX Blanchard & Wolfers 2000 XXX Bassanini & Duval 2006 XXX Baccaro & Rei 2007 XXXX Stockhammer 2004 XXX Palacio-Vera et al 2006 XXXX Arestis et al 2007 XXX

15. 15 This paper Medium run unemployment (U MR ) Actual Unemployment (u) ΔpΔp LMI Capital accumulation (ΔK) Interest rates (i CB ) Macroec. shocks NAIRU (U N )

16. 16 Econometric Model & Data  u NAIRU = f(LMI, MS,  K)  panel least squares regression  DATA  Bassanini & Duval 1982-2003  most up-to-date OECD dataset  Basis of OECD Empl Outlook 2006  Baker, Glyn, Howell, Schmitt 1960-99  Updated and revised version of Nickell & Nunziata LMI-DB  u, K, CPI: EU AMECO  Capital stock or investment?  20 Countries (all Western OECD and JP, without LU and GR)  All data transformed into non-overlapping 5-yr- avgs (Blanchard & Wolfers 2000, Baccaro & Rei 2007)

17. 17 Regression results (BD 83-03 dataset)

18. 18 Regression results based on Baker et al dataset (1960-99)

19. 19 Findings I  LMI do play a role, but a minor one  Few are consistently statistically significant, many display perverse signs  Only Union Density (UD) statistically significant (and with the expected sign) across data sets and different specifications  In line with Baccaro and Rei (2007), Howell et al (2005)  Demand shocks do play an important role  Strong (and significant) effect of real interest rates  +1%-pt. real interest rate → +0.5%-pts. unemployment rate  Strong (and significant) effect of ACCU  +1% ACCU → -0.9%-pts. unemployment rate  Supply shocks play some (minor) role  TOT, LD have effects, particularly in the short (BD) sample

20. 20 Interactions among LMI  Interactions have been popular in the recent (mainstream) literature  e.g. Blanchard and Wolfers 2000, IMF 2003, Nickell et al 2005, Bassanini and Duval 2006 ... but have been reported rather selectively,  i.e. the ones that were found statistically significant  Theoretical background for interactions weak  Many possible interactions (we report 60)  Some (11/60) are statistically significant  About half of them with counterintuitive signs  Overall: unconvincing

21. 21 Interactions among LMI

22. 22 Contributions to unemployment in %pts relative to 1960-64 for ‚mean country‘ (BGHS dataset)

23. 23 Contributions across countries (in %-pts.) from 1990-94 to 2000-03 (BD 1982-2003)

24. 24 Findings II  Economic relevance over time  Early 70s: INT and ACCU lower unemployment, from 80s on strong positive effects: high real interest rates and diminished ACCU raise unemployment  70s and 80s: UD contributes to unemployment, in 90s the effect is negative  From mid 80s on, TW contributes to unemployment (however typically insignificant in our study, as opposed to B&D 2006)

25. 25 Measures of investment and capital  Is capital accumulation a demand-side or a supply-side variable?  Investment instead of ∆K/K: demand shocks  Perform very similar econometrically (log real gross fixed capital formation indeed even works somewhat better)  K (stock): supply side variable  Post Keynesian ‚capital shortage‘ argument: periods of insufficient ACCU leave capital stock ill-suited to accommodate high employment growth in the subsequent recoveries  There is some evidence that it has negative effect on unemployment, but not entirely robust (clearer in BD82-03 than in BGHS60-99)  We find evidence for both channels  weak evidence that demand-side effects of capital accumulation are stronger (more robust) than supply-side effects

26. 26 Capital accumulation: demand effects and supply effects (B&D dataset)

27. 27 Capital accumulation: demand effects and supply effects (BGHS data set)

28. 28 Summary / conclusion  The degree of capital accumulation has a strong impact on unemployment in the medium run  Real interest rate shocks play a significant role even when ACCU is simultaneously controlled for  The effects of LMI are relatively modest and unrobust  Only Union Density delivers consistent (positive) effect  Counteracted by COORD (or CBC)  Some role for TW; ‚perverse‘ results for EPL, UB  Investment (flow) seems to play a larger role (as a demand shock) than the capital stock (as a supply variable)  Policy conclusions:  focus on LMI as explanations for persistent unemployment is misguided – labour market reforms will not cure unemployment  Encourage investment and capital accumulation

29. 29 Thank you for your attention.

30. 30 Standard story: high u and high W/P WBC 1 (LMI) PS 1-u, e W/P WBC 2 (inflex LMI) A B e N = 1-NAIRU

31. 31 Stylized facts: high u and low W/P WBC PS 1-u, e W/P A B C