Skip to main content

Labor market institutions and demographic employment patterns


We study collective bargaining’s effect on relative employment for youth, women, and older individuals. Our model of collective wage setting predicts that unionization reduces employment more for groups with relatively elastic labor supply: youth, older individuals, and women. We test this implication using data from 17 Organization for Economic Cooperation and Development (OECD) countries over the 1960–1996 period. We find that time-varying indicators of unionization decrease the employment–population ratio of young and older individuals relative to the prime-aged, and of prime-aged women relative to prime-aged men, and unionization raises the unemployment rate of prime-aged women and, possibly, young men compared to prime-aged men.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2


  1. 1.

    See Blanchflower and Freeman (2000b) on youth employment, Disney (1996) on older workers’ employment, and Ruhm (1998) or Blau and Kahn (2000) on women’s employment.

  2. 2.

    See, e.g., Edin and Topel (1997) or Davis and Henrekson (2005) for studies of Sweden and Kahn (1998) for Norwegian evidence with this conclusion.

  3. 3.

    See, for example, Card et al. (1999; Canada, France, and the USA), Krueger and Pischke (1998) or Blau and Kahn (2000; Germany and the USA) for studies that do not find that wage compression leads to relatively less employment for the low skilled; on the other hand, Freeman and Schettkat (2000; Germany and the USA) do find such effects.

  4. 4.

    We abstract from possible general equilibrium effects of union policy. For example, if unions reduce female employment, the demand for child care would presumably fall, possibly affecting the pattern of labor demand.

  5. 5.

    Below, we discuss the implications of relaxing this assumption. See Farber (1986) for further discussion of alternative models.

  6. 6.

    Note that when flat-rate welfare benefits serve as a floor for market wages, the labor supply of low-wage workers is more elastic than that of high-wage workers. Koeniger et al. (2007), in their study of overall wage inequality, made the related point that if unemployment insurance (UI) benefits are a higher fraction of wages for lower wage workers, then an optimizing union will compress wages.

  7. 7.

    In Bertola et al. (2002), we show that the same employment outcomes can be achieved through an employment tax whose proceeds are distributed to workers. In addition to affecting employment by influencing relative wages, unions may directly affect relative employment by agreeing to downsizing on the condition that older workers are separated first (Casey 1992); moreover, the most recent (and younger) employees are usually laid off on a last-in-first-out basis.

  8. 8.

    It may be interesting to note that non-separabilities may also play a role on the supply side, for example, in that the nonemployment opportunities of an individual may be larger when a spouse works: This channel can help rationalize polarization of institutional disemployment effects across demographic lines.

  9. 9.

    Traditional production–function estimation exercises, such as in Berger (1983) and some of the works on US wage inequality reviewed by Topel (1997), find evidence of substitutability between female and youth, or unskilled, male employment. Skilled prime-aged workers, however, are not close substitutes for youth, female, and older workers, whereas individuals within these groups are closely substitutable for each other (Disney 1996).

  10. 10.

    Note that it is possible that some workers, such as women, encounter labor market discrimination. Indeed, an extensive literature on the gender–pay gap suggests that both gender differences in productivity and discrimination play a role in causing the observed gender–wage differential. The possibility of discrimination can easily be accommodated in the model by adjusting “true” productivity by the discrimination coefficient. The same reasoning applies to other groups such as older or younger workers who may face discrimination. Because this issue is not central to our concerns in this study and leaves our basic reasoning unchanged, we do not explore it further theoretically but note that the adjusted productivity interpretation is most likely the relevant one for women. We do, however, explore the impact of controlling for antidiscrimination policies below.

  11. 11.

    Coverage can be different across groups (we discuss some of the empirical implications of this below), but to the extent that prime age males are more covered, the implications of such differences are as counterfactual as those of different bargaining power above.

  12. 12.

    Although wage insurance may explain wage compression within similar groups, it is harder to use an insurance argument to explain wage compression between ex ante identifiable groups such as men and women.

  13. 13.

    Christopher Ruhm kindly provided us with the data on weeks of paid parental leave that he used in Ruhm (1998) to find that they increase women’s relative employment while also reducing their relative wages at extended durations. Unfortunately, however, there was too little overlap between his data and ours in countries and periods covered to allow us to control for parental leave policies.

  14. 14.

    In Bertola et al. (2002), we estimated models with relative employment as the dependent variable, thus constraining the impact of the explanatory variables on the two comparison groups to be equal in absolute value. The results of those more restrictive models were very similar to those reported below.

  15. 15.

    We implemented unit root tests for our panel using a method suggested by Maddala and Wu (1999). Because of our short panel, usually seven periods, we interpret these results very cautiously. To test for unit roots, we computed Dickey–Fuller statistics on the residuals for each country and their associated significance levels using the approximations in MacKinnon (1994). We then implemented the suggestion of Maddala and Wu (1999) to aggregate these individual country tests using an exact Fisher’s test, under which −2 times the sum of the logs of the significance levels has a chi-squared distribution with degrees of freedom equal to 2 times the number of countries. When we used this method to analyze the residuals from each of the basic regression models, we rejected the null hypothesis of no cointegration in each case (albeit not taking into account the fact that the residuals are themselves estimated variables due to the short panels). Thus, under these tests, we reject the hypothesis of spurious regression across our time-averaged observations.

  16. 16.

    Our results for the retirement variables are partially consistent with those of Blöndal and Scarpetta (1999) who examined the labor force participation rate of men aged 55–64 for 15 countries for the 1971–1995 period. Note that, of the explanatory variables in our analysis, the retirement variables are perhaps the most likely to suffer from reverse causation. We, nonetheless, present results including them to provide a sharper test of the impact of the collective bargaining variables, our primary focus. Results for these variables were similar when the retirement variables were excluded.

  17. 17.

    As explained in the Appendix, for countries for which the first period we observed coverage is, say, t 0, we assign the t 0 value to all prior periods. Our basic results for the unionization variables’ main effects were the same when we included a dummy variable for these observations and an interaction between collective bargaining coverage and this dummy variable. In this model, the only observations that contribute to the unionization main effects are ones with actual observations for collective bargaining coverage.

  18. 18.

    In addition, using the estimates in Table 5, we can conclude that unionization significantly lowers the employment of young and older women relative to prime age men, using either within-country or between-country standard deviations of the union variables, and whether or not we control for the unemployment rate.

  19. 19.

    In Bertola et al. (2002), we used a less flexible specification, constraining the impact of each institutional variable to be of equal magnitude and opposite sign. The results were qualitatively similar to those reported in this paper. OECD (2004), running that specification on a much shorter time span of annual data and computing collinearity-robust statistics of the type devised in that earlier version of our work, obtains much weaker evidence.

  20. 20.

    One might also speculate that cohort crowding effects would be larger with more rigid labor markets (Korenman and Neumark 2000). However, we were unable to obtain easily interpretable results testing such a prediction, possibly due to insufficient degrees of freedom, as the models in question had 65 coefficients and 101 observations.

  21. 21.

    Compositional effects are potentially important in interpreting aggregate movements in unemployment and wages (Blundell et al. 2003).

  22. 22.

    The EU, for example, issued directives on equal pay in 1957 and 1975 and on equal employment opportunity in 1976; the ILO issued conventions 100 (equal pay for work of equal value) in 1951 and 111 (equal employment opportunity) in 1958, and all of the countries in our sample except the USA (ratifying neither) and Japan (ratifying only convention 100) had ratified both ILO conventions by 1999. Most of these ratifications occurred during the 1952–1970 period, or before most of the period we examined. For details on these policies, see the ILO and European Union websites: and

  23. 23.

    Our main source for information on such policies was OECD (1988, pp. 167–168), which has data on equal pay and equal employment opportunity policies for all of our countries except The Netherlands and New Zealand. For these two countries, we obtained data from van der Sanden and Waaldijk (2001) and the New Zealand Human Rights Commission (

  24. 24.

    A similar argument applies to the results for older individuals, with retirement being the alternative to employment in their case, and motivates inclusion of retirement-system characteristics alongside unionization indicators in our employment and unemployment equations.

  25. 25.

    These results are consistent with Kahn’s (2000) findings for a cross section of 15 OECD countries. He found that collective bargaining coverage had a negative effect on youth relative employment and was positively associated with school attendance among young adults, but that enrollment did not fully account for the negative effect of union coverage on the relative employment of youth. Taking Kahn’s findings in conjunction with those reported above suggests that unions may increase the share of out-of-the-labor force youth who are neither at work nor at school. We should point out that different policies toward apprentice training may be an omitted factor in our regressions that could help explain the youth employment results. However, to the extent that national training systems do not change dramatically, we have accounted for these using the country dummy variables. For some discussion of cross-country differences in these policies, see Steedman (2001).

  26. 26.

    An additional reason for unions to raise the relative wages of youth is suggested by Lazear (1983) who argues that prime age union members wish to reduce competition from younger workers.


  1. Agell J, Lommerud KE (1992) Union egalitarianism as income insurance. Economica 59(235):295–310

    Article  Google Scholar 

  2. Autor DH, Donohue III JJ, Schwab SJ (2006) The costs of wrongful discharge laws. Rev Econ Stat 88(2):211–231

    Article  Google Scholar 

  3. Berger MC (1983) Changes in labor force composition and male earnings: a production approach. J Hum Resour 18(2):177–196

    Article  Google Scholar 

  4. Bertola G (1999) Microeconomic perspectives on aggregate labor markets. In: Ashenfelter O, Card D (eds) Handbook of labor economics, vol 3c. North-Holland, Amsterdam, pp 2985–3028

    Chapter  Google Scholar 

  5. Bertola G, Blau FD, Kahn LM (2002) Labor market institutions and demographic employment patterns. Working paper 9043, National Bureau of Economic Research, Cambridge, MA; also, discussion paper 3448. Centre for Economic Policy Research, London

  6. Blanchard OJ, Wolfers J (2000) The role of shocks and institutions in the rise of European unemployment: the aggregate evidence. Econ J 110(462):C1–C33

    Article  Google Scholar 

  7. Blanchflower DG, Freeman RB (2000a) The declining economic status of young workers in OECD countries. In: Blanchflower DG, Freeman RB (eds) Youth employment and joblessness in advanced countries. University of Chicago Press, Chicago, pp 19–55

    Google Scholar 

  8. Blanchflower DG, Freeman RB (eds) (2000b) Youth employment and joblessness in advanced countries. University of Chicago Press, Chicago

    Google Scholar 

  9. Blau FD, Kahn LM (1996) International differences in male wage inequality: institutions versus market forces. J Polit Econ 104(4):791–837

    Article  Google Scholar 

  10. Blau FD, Kahn LM (2000) Gender and youth employment outcomes: the U.S. and West Germany, 1984–91. In: Blanchflower DG, Freeman RB (eds) Youth employment and joblessness in advanced countries. University of Chicago Press, Chicago, pp 107–167

    Google Scholar 

  11. Blau FD, Kahn LM (2002) At home and abroad: U.S. labor-market performance in international perspective. Russell Sage, New York

    Google Scholar 

  12. Blau FD, Kahn LM (2003) Understanding international differences in the gender pay gap. J Labor Econ 21(1):106–144

    Article  Google Scholar 

  13. Blöndal S, Scarpetta S (1999) The retirement decision in OECD countries. OECD Economics Department working paper no. 202. OECD, Paris

  14. Blundell R, MaCurdy T (1999) Labor supply: a review of alternative approaches. In: Ashenfelter O, Card D (eds) Handbook of labor economics, vol 3c. North-Holland, Amsterdam, pp 1559–1695

    Chapter  Google Scholar 

  15. Blundell R, Reed H, Stocker TM (2003) Interpreting aggregate wage growth: the role of labor market participation. Am Econ Rev 93(4):1114–1131

    Article  Google Scholar 

  16. Calmfors L, Driffill J (1988) Centralization of wage bargaining. Econ Policy 3:14–61

    Article  Google Scholar 

  17. Card D, Kramorz F, Lemieux T (1999) Changes in the relative structure of wages and employment: a comparison of the United States, Canada, and France. Can J Econ 32(4):843–77

    Article  Google Scholar 

  18. Casey B (1992) Redundancy and early retirement: the interaction of public and private policy in Britain, Germany and the United States. Br J Ind Relat 30(3):425–443

    Google Scholar 

  19. Davis SJ, Henrekson M (2005) Wage-setting institutions as industrial policy. Labour Econ 12(3):345–377

    Article  Google Scholar 

  20. Disney R (1996) Can we afford to grow older? A perspective on the economics of aging. MIT Press, Cambridge, MA

    Google Scholar 

  21. Edin P-A, Richardson K (2002) Swimming with the tide: solidarity wage policy and the gender earnings gap. Scand J Econ 104 (1):49–67

    Article  Google Scholar 

  22. Edin P-A, Topel R (1997) Wage policy and restructuring: the Swedish labor market since 1960. In: Freeman RB, Topel R, Swedenborg B (eds) The welfare state in transition: reforming the Swedish model. University of Chicago Press, Chicago, pp 155–201

    Google Scholar 

  23. Erickson CL, Ichino AC (1995) Wage differentials in Italy. In: Freeman RB, Katz LF (eds) Differences and changes in wage structures. University of Chicago Press, Chicago, pp 265–305

    Google Scholar 

  24. Farber HS (1986) The analysis of union behavior. In: Ashenfelter O, Layard R (eds) Handbook of labor economics, vol. 2. North-Holland, Amsterdam, pp 1039–1089

    Chapter  Google Scholar 

  25. Farber HS, Krueger AB (1993) Union membership in the United States: the decline continues. In: Kaufman BE, Kleiner MM (eds) Employee representation: alternatives and future directions. Industrial Relations Research Association, Madison, WI, pp 105–134

    Google Scholar 

  26. Freeman RB, Schettkat R (2000) The role of wage and skill differences in U.S.–German employment differences. Working paper 7474. National Bureau of Economic Research, Cambridge, MA

  27. Jimeno JF, Rodríguez-Palenzuela D (2002) Youth unemployment in the OECD: demographic shifts, labour market institutions, and macroeconomic shocks. Documento de trabajo 2002-15. FEDEA, Madrid

  28. Kahn LM (1998) Against the wind: bargaining recentralisation and wage inequality in Norway, 1987–1991. Econ J 108(448):603–645

    Article  Google Scholar 

  29. Kahn LM (2000) Wage inequality, collective bargaining and relative employment 1985–94: evidence from fifteen OECD countries. Rev Econ Stat 82(4):564–579

    Article  Google Scholar 

  30. Koeniger W, Leonardi M, Nunziata L (2007) Labor market institutions and wage inequality. Ind Labor Relat Rev (in press)

  31. Korenman S, Neumark D (2000) Cohort crowding and youth labor markets: a cross-national analysis. In: Blanchflower DG, Freeman RB (eds) Youth employment and joblessness in advanced countries. University of Chicago Press, Chicago, pp 57–105

    Google Scholar 

  32. Krueger AB, Pischke J-S (1998) Observations and conjectures on the U.S. employment miracle. In: German-American Academic Council (ed) Third Public German-American Academic Council Symposium: Labor Markets in the USA and Germany. German–American Academic Council, Bonn, pp 99–126

    Google Scholar 

  33. Lazear EP (1983) A competitive theory of monopoly unionism. Am Econ Rev 73(4):631–643

    Google Scholar 

  34. Lewis HG (1986a) Union relative wage effects. In: Ashenfelter OC, Layard R (eds) Handbook of labor economics, vol 2. North Holland, Amsterdam, pp 1141–1181

    Google Scholar 

  35. Lewis HG (1986b) Union relative wage effects: a survey. University of Chicago Press, Chicago

    Google Scholar 

  36. MacKinnon JG (1994) Approximate asymptotic functions for unit-root and cointegration tests. J Bus Econ Stat 12(2):167–176

    Article  Google Scholar 

  37. Maddala GS, Wu S (1999) A comparative study of unit root tests with panel data and a new simple test. Oxf Bull Econ Stat 61(special issue):631–652

    Article  Google Scholar 

  38. McDonald IM, Solow RM (1981) Wage bargaining and employment. Am Econ Rev 71(5):896–908

    Google Scholar 

  39. Mincer J (1976) Unemployment effects of minimum wages. J Polit Econ 84(4, part 2):S87–S104

    Google Scholar 

  40. Montenegro CE, Pagés C (2004) Who benefits from labor market regulations?: Chile 1960–1998. In: Heckman JJ, Pagés C (eds) Law and employment: lessons from Latin America and the Caribbean. University of Chicago Press, Chicago, pp 401–434

    Google Scholar 

  41. Neumark D, Wascher W (2004) Minimum wages, labor market institutions, and youth employment: a cross-national analysis. Ind Labor Relat Rev 57(2):223–248

    Article  Google Scholar 

  42. Nickell S, Layard R (1999) Labor market institutions and economic performance. In: Ashenfelter O, Card D (eds) Handbook of labor economics, vol. 3C. North-Holland, Amsterdam, pp 3029–3084

    Chapter  Google Scholar 

  43. Nickell S, Nunziata L, Ochel W, Quintini G (2003) The beveridge curve, unemployment and wages in the OECD. In: Aghion P, Frydman R, Stiglitz J, Woodford M (eds) Knowledge, information and expectations in modern macroeconomics: in honor of Edmund S. Phelps. Princeton University Press, Princeton, pp 394–431

    Google Scholar 

  44. OECD (various issues). Labour force statistics. OECD, Paris

  45. OECD (1988) OECD employment outlook, September 1988. OECD, Paris

  46. OECD (2004) Wage setting institutions and outcomes. In: OECD employment outlook. OECD, Paris, pp 127–182

  47. Polachek SW, Xiang J (2006) The gender pay gap: a cross-country analysis. Paper presented at the society of labor economists meetings, Boston, MA

  48. Rosen S (1970) Unionism and the occupation wage structure in the United States. Int Econ Rev 11(2):269–286

    Article  Google Scholar 

  49. Rowthorn RE (1992) Centralisation, employment and wage dispersion. Econ J 102(412):506–523

    Article  Google Scholar 

  50. Ruhm CJ (1998) The economic consequences of parental leave mandates: lessons from Europe. Q J Econ 113(1):285–317

    Article  Google Scholar 

  51. Steedman H (2001) Benchmarking apprenticeship: UK and continental Europe compared. Centre for Economic Performance D.P.513, London

  52. Topel RH (1997) Factor proportions and relative wages: the supply-side determinants of wage inequality. J Econ Perspect 11(2):55–74

    Google Scholar 

  53. van der Sanden J, Waaldijk B (2001) Background data report: Netherlands. In: European Union, European Union research project HPSE-CT-2001-00082, employment and women’s studies: the impact of women’s studies training on women’s employment in Europe. University of Hull, Hull

Download references


This paper benefits from helpful comments by Richard Disney, Christian Dustmann, Richard Freeman, Harry Holzer, Steve Machin, Justin Wolfers, three anonymous referees, and seminar participants at Cornell University, University of Texas, University of Turin, Juan March Institute (Madrid), and conference participants at the AEA/IRRA meetings, Regensburg, Bergen, and Bonn. We are also grateful to Justin Wolfers and David Neumark for providing data, and to Julian Messina, Abhijay Prakash, and Thomas Steinberger for the excellent research assistance. Portions of this paper were written while the second and third authors were Visiting Fellows in the Economics Department at Princeton University, supported by the Industrial Relations Section. They are grateful for this support. The corresponding author is Lawrence M. Kahn.

Author information



Corresponding author

Correspondence to Lawrence M. Kahn.

Additional information

Responsible editor: Christian Dustmann



This paper’s data set builds on that constructed by Blanchard and Wolfers (2000), documented at The data set contains macroeconomic and institutional data on 26 OECD countries for 8 5-year periods covering the time span 1960–1999. We have added data on labor force by age groups, population by age groups, and unemployment rates by age groups for male and female workers separately, as well as time-varying data on a number of institutions.

Demographic disaggregration of employment and unemployment:

The labor force and population data are taken directly from the International Labour Organization (ILO) database “Economically Active Population 1950–2010.” The data on unemployment rates by age group were constructed from data found in the OECD (various issues) publication Labour Force Statistics. These are country-source data, and we did not attempt to harmonize their definition. To compute the average unemployment rate for each 5-year period, we calculate the arithmetic mean of the yearly unemployment rates within the period. To obtain similar data on as many countries as possible, we also aggregate the data to broad age groups by computing the labor force weighted average of the time-averaged unemployment rate of the relevant age groups. The labor force weights themselves are constructed as linearly interpolated weights from the labor force data used above.

Institutional indicators:

Time-varying employment protection legislation indicators are from the Blanchard and Wolfers (2000) dataset.

Union density, collective bargaining coverage and coordination, and labor tax rates are from the data appendix to Nickell et al. (2003), kindly attached by the authors to the discussion paper version of their study at The collective bargaining coverage was available for some countries from 1960 to 1999 and for other countries from 1980–1994. We used interpolation and assigned the authors’ earliest figure to all dates before its date.

The UI year 1 and 5 replacement rates were taken from a OECD database and were available for the entire 1960–1996 period.

The data on retirement system characteristics were interpolated from those available in Blöndal and Scarpetta (1999): male and female retirement ages in 1961, 1975, and 1995 from Table III.1; 10-year pension accrual rate in 1967 and 1995 from Table III.4; pension replacement rates for 1961, 1975, and 1995 from Table III.3; disability and unemployment special scheme replacement rates for 1961, 1975, and 1995 from Table IV.3.

Individual union membership data were taken from the International Social Survey Programme (ISSP) for 1985–1998.

The data on enrollment in education were taken from the World Bank’s 1995 CD edition of the World Tables.

Table A1 Correlation matrix for the institutional variables

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Bertola, G., Blau, F.D. & Kahn, L.M. Labor market institutions and demographic employment patterns. J Popul Econ 20, 833–867 (2007).

Download citation


  • Unions
  • Demographic employment differentials

JEL Classification

  • J21
  • J23
  • J51