The evolution of wage inequality in Italy

Abstract

This article analyses the evolution of inequality in yearly and daily wages between and within groups of blue and white collar, using the INPS-ISFOL database for the period 1985 to 1999 in Italy. Between-group inequality increased in the 1990s as clerical wages grew slowly, whereas blue collars' wages remained nearly constant. Within-group inequality increased only if measured by daily wages. The covariance structure analysis shows that inequality comes from persistent differentials among older workers and from high income volatility for younger cohorts. Within inequalities in office and manual workers are driven by the growth of permanency for the older cohorts (individual abilities, say experience, matter more) and by the growth of income volatility for the younger cohorts (luck in the labour market). Within each group, low paid workers during their career acquire earning gains for their abilities and reduce differentials with respect to high paid workers.

Acknowledgements

Database utilization has been made possible by Authors partecipation in the research partnership between the DSE (Dipartimento di Scienze Economiche, Università La Sapienza Rome) and ISFOL (Istituto per lo Sviluppo della Formazione Professionale dei Lavoratori). We thank participants to the XX AIEL conference held in Rome, 22, 23 September 2005 for their useful comments.

Notes

¹ The same technological change occurring in different institutions in the labour market has given rise to different outcomes in Europe and in the United States. Because of labour market rigidities, the common technological shock caused a pronounced effect on employment in Europe (Ljunqvist and Sargent, 1998), while in the United States this shock was mainly absorbed by prices, modifying the wage structure (Bertola and Ichino, 1995). Nevertheless, even in Europe wage distribution shifted towards rising inequality (Glyn, 2001; Ayala et al., 2002).

² See Leslie and Pu (1995) and DiNardo et al. (1996) for empirical support of this viewpoint.

³ To cope with this evidence, Aghion et al. (2002) relate ‘high generality’ of new technology to the amplification of the role of luck in the labour market. Furthermore, SBTC would have led to changes in firm organization and job tasks, from centralized, nonflexible, hierarchical and large firm structures to horizontal structures based on team work. Caroli and Van Reenen (2002) analysed the effect of the process of firm reorganization that was driven by massive injection of information technologies, on wage structure.

⁴ Income inequality leads to consumption inequality, ending in different economic growth patterns, see Blundell and Preston (1998).

⁵ Some facts should lead us to remark the impact of institutional change on inequality, for instance the widespread use of temporary contracts that made young workers less secure in their income than the older ones, lack of on-the-job training that reduces upward mobility for less-qualified workers and so on.

⁶ In 1984, in Turin, 40 000 white collar workers and managers of the FIAT demonstrated against the 35-day-strike decided by the Federazione Lavoratori Metalmeccanici (FLM)–Metal Workers Federation (the most important union in the metallurgical sector in Italy) in order to avoid 15 000 dismissals.

⁷ Gross wages are the sum of net wages, taxes and social contributions on workers; social contributions on firms are not included in gross wages.

⁸ ISFOL, is an institute for Training Workers.

⁹ For a detailed description of the dataset, see Centra and Rustichelli (2005).

¹⁰ In that way, we drop 7657 observations, that is a higher number than the expected one (0.2% of the whole sample) first because, we drop observations where the yearly or daily wage was in one of the two tails described above and second because all individuals with zero wages were obviously included in the lower percentile. In fact, we drop 441 individuals whose observations were in the two tails for the whole period.

¹¹ In fact, we consider these workers as two different individuals for the sub-period of their career. Individuals with the same qualification but with different jobs during the same year are also considered because our analysis leaves aside job characteristics; their yearly revenue is the sum of revenues in the different jobs.

¹² Calculated as revenue of the individual − average revenue of the individual + average global revenue, so that it can be negative.

¹³ For the United States DiNardo et al. (1996) analysed the relationship between the reduction in minimum wages and inequality.

¹⁴ To improve visibility, we do not plot observations with a yearly wage lower than 25 000 000 Italian lire. Doing so, we drop observation concerning younger cohorts in their youth.

¹⁵ We have 10 cohorts and 120 distinct elements in each variance–covariance matrix (T = 15, so T · (T + 1)/2 = 120). As such the analysis is based on 1200 measured covariances for each subset of data.

¹⁶ Cappellari (2002), used this formulation to compare the covariance structure of white collars between private and public sector.

¹⁷ MaCurdy (1982), analysed several stochastic specifications for the earning process in the USA, in order to better characterize the covariance structure. Similarly, Borella (2001), for the Italian earnings, on SHIW data.

¹⁸ Low-talented workers catch-up with high-talented ones by means of a higher wage growth rate during their career.

¹⁹ Cappellari (2002) found that white collars' wages in the public sector are even more persistent. Greater insider bargaining power along with the higher endowment in skill could be the reasons for this result.