Employment and unemployment in Italy: measurement issues and behavioural analyses


Research Project supported by the Italian Ministry of University and Scientific Research, 1999

[Home] [Versione italiana] [Printer-friendly version]

Projects | 1997 | 1999 | 2001 | 2003 | 2005 



The programme deals with issues of measurement and modelling of participation at work in Italy, by using suitable micro databases. From a general perspective, the research programme aims at bringing together: (i) substantive issues about the analysis of the Italian labour market; (ii) methodological issues on the measurement of labour phenomena, both on the side of variables measuring and on the side of model specification and estimation; (iii) issues of utilisation and/or development of micro-data bases, both from surveys and administrative sources.
The programme focuses on some five broad topics, largely related one to another: (a) analysis of job mobility, wage mobility and flexibility, based on a longitudinal employer-employee linked database from Social Security files; (b) information on labour force participation from Labour Exchange archives (NETLABOR); (c) measurement and modelling of employment and unemployment from surveys; (d) combining/linking data from complementary sources; (e) analysis of some weak segments of the labour market. The primary interest is on the Italian labour market, but some comparative analyses extended to EU countries and the USA are also planned.
Moving from the employer-employee panel 1986-'94 from the Social Security files, we first plan to update it up to 1998. We then plan to study; (i) non-sampling errors in the gross earnings measure from Social Security files; (ii) issues of earnings mobility and flexibility; (iii) the dynamics of labour market segmentation, possibly using simulation systems.
Research units: Torino
Moving from the case study performed within the 1998 programme on some Labour Exchanges in the Veneto region (see further section 2.2), we plan the carry out the research programme along the following lines. (i) Analysis of the incidence among demographic groups and the evolution over time of part-time and short-term jobs. (ii) Estimation of unemployment, especially for small areas, by suitably exploiting information on registered 'unemployed' in the archive. (iii) Use of NETLABOR to evaluate the impact of selected labour market policies.
Research units: Padova
(C1) The main data source here is the Labour Force Survey, from which we obtained several (anonimysed) sub-samples over the years and across the country within a partnership with Istat. The first research task is to improve on current criteria for counting the unemployed. We shall look at the issue from different points of view: (i) we will go on with, and expand, exploratory analyses of labour market attachment; (ii) we will improve on the formulation of a "Double-Hurdle" model of labour supply with unemployment and fallible indicators of labour force state; (iii) we will focus on youth unemployment and on the potential role of the family in explaining the phenomenon; (iv) finally, we will try to extend the analysis of the relevance of the classification of non-employed persons (by distinguishing persons out of the labour force and unemployed with various degrees of labour market attachment) in a dynamic setting, along the lines suggested by Flinn and Heckman (1983) and Jones and Riddell (1999).
(C2) A second target in this area is related to the pilot survey on participation at work based on a 'time-use' approach, to be run by ISTAT in fall 1999 (see section 2.2). We plan (i) to carry out an in-depth analysis of its results, both on the substantive issue of measuring participation at work, chiefly "hidden participation", and on the methodological aspects of the survey operations, (ii) and possibly to go on with the planning and design of large scale national survey on the topic.
(C3) As for the analysis of the dynamics of employment and unemployment, we will use panel sub-samples from the Labour Force Survey (obtained thanks to a linkage procedure developed within the 1998 programme) and from the Bank of Italy Survey on Family Budgets. We plan to focus on the following topics: (i) estimation of gross flows matrices consistent with the constraints on the marginal distributions provided by cross-sectional estimates (see point (D1) below), and their further refinements by allowing for measurement errors in reported labour force state; (ii) analysis of unemployment duration: improvements in the measurement of duration itself, by exploiting the panel structure of the data and the consistency checks on reported duration and labour force state thus allowed; reduced-form modelling of unemployment duration with specific attention to the phenomenon of "long-term unemployed".
(C4) As for school-to-work youth transitions, longitudinal data from a survey on four cohorts of students attaining the compulsory level will be analysed, accounting for attrition.
Research units: (C1) Padova; (C2) Padova and Siena; (C3) Padova; (C4) Siena.
(D1) The issue of combining data from different sources arises both at the aggregate and at the micro level. At the aggregate level it demands for calibration methods. We will use such an approach for three types of problems: (i) estimation of transition matrices from the Labour Force Survey, with gross flows constrained to be consistent with marginal distributions estimated from the two cross-sectional surveys; (ii) reconciliation of labour mobility estimates coming from different sources, i.e., from the Labour Force Survey and from Social Security data-base; (iii) combined use of sampling information on unemployment from the Labour Force Survey with information on registered unemployed to the Labour Exchanges provided by NETLABOR.
(D2) At the micro level, the issue is about the linkage of individual records from complementary sources. It is likely that, at the beginning of 2000, a first version of the integrated Labour Force Survey-Social Security files micro database (planned to be produced within the 1998 programme: see section 2.2) will be available. Potentially, the integrated database is quite interesting, because the two sources are essentially complementary. We plan to start working on this integrated data-base with reasonably moderate targets: quality assessment; comparisons of the two sources with respect to their capacity to identify the employed. In the final phase of the research programme, we plan to use the database for specifying and estimating models of labour supply and job mobility.
Research units: (D1) Padova and Torino, (D2) Padova.
(E1) Relationships between participation at work and spells of poverty will be studied using data from the Survey on Household Budgets (by the Bank of Italy) and from the En Community Household Panel (, taking into account the non negligible attrition in both surveys.
(E2) The participation at work behaviour by females, youths and immigrants will be studied using Social Security files over the period 1974-1996 with reference to the Veneto region.
Research units: (E1) Siena and Padova; (E2) Venezia.

Info: anna.giraldo@unipd.it