Michele Raitano
The Future of Social Protection
What Works for Non-standard Workers?
Chapter 6. Italy: Para-subordinate workers and their social protection
Abstract
This chapter focuses on para-subordinate workers in Italy, i.e. individuals who are legally self-employed but who are often “economically dependent” on a single employer. The chapter first presents the welfare provisions that para-subordinate workers are entitled to receive, pointing out that the disadvantages facing para-subordinate workers relative to employees have been recently reduced. Afterwards, making use of the most suitable available longitudinal data, statistics about the evolution over time of the number and characteristics of para-subordinate workers are shown and the first phase of the working career of para-subordinate workers is observed, assessing vulnerability in terms of earnings level, employability and accrual of pension contributions over a ten-year period.
6.1. Introduction
The term “para-subordinate worker” is used in Italy to describe individuals who are self-employed in legal terms – they work at their own risk and are not formally the subordinate of an employer – but are often “economically dependent” on an employer since, in most cases, their activity is reliant upon one or a small number of clients.
In practical terms, the term covers categories of workers who are mandatorily enrolled in a special public fund called Gestione Separata, established by Law No. 335/1995, which is managed by the Italian Social Security Institute (INPS). They are obliged to pay pension contributions to this fund and – when they only work as a para-subordinate – additional contributions to finance sickness, unemployment and family benefits and allowances for maternity and parental leave. However, the group of workers enrolled in the Gestione Separata is heterogenous along many dimensions and welfare guarantees differ accordingly (see Section 6.2).
First, two main groups of workers are enrolled in the Gestione Separata: various types of “collaborators” – e.g. co-ordinated and continuous collaborators, occasional or project collaborators1 and so-called “additional workers” (also named “workers on vouchers”) – and professionals who perform their activity in an unlicensed sector not regulated by a specific professional register (e.g. web-designers, archaeologists).2 Those who work as a professional in a licensed sector regulated by a professional register – e.g. lawyers, architects, accountants, physicians – are not enrolled in the Gestione Separata and are not regarded as para-subordinate workers.3
Second, para-subordinate jobs can be associated with other types of employment or self-employed activities (in both the public and the private sector) or can be performed by retired persons. Therefore, a distinction can be made between exclusive para-subordinate workers (i.e. those only working as a para-subordinate) and non-exclusive para-subordinate workers (i.e. pensioners and those with another non-para-subordinate job).
Third, the category of collaborators is also heterogenous, since it includes generally low-paid workers hired under short-term arrangements (the types of collaborators mentioned above and workers on vouchers), relatively high-paid workers obliged by the law to enrol in the Gestione Separata – namely, company accountants, administrators and executives – and some high-skilled workers at the beginning of their careers (PhD students, postdoctoral fellows and physicians attending a postgraduate qualification course).
In the economic policy debate, especially before the outbreak of the economic crisis, many worries about the growing number of para-subordinate workers emerged (Berton et al. 2012), since, in most cases, these workers suffered low wages, frequent unemployment spells and were entitled to lower welfare guarantees than employees, with respect to both pension benefits and other payments (maternity, sickness, family and unemployment allowances). Furthermore, these worries were exacerbated by the feeling that many para-subordinate arrangements (for both collaborators and professionals) hid true employment relationships, which the employer organised as para-subordinate arrangements to save labour costs thanks to the lower social contributions levied on these arrangements.
A specific concern relates to the expected pensions of para-subordinate workers, because in the past they have paid a reduced contribution rate (and for professionals a reduced rate is still in force; see Section 6.2) and in the Italian notional defined-contribution public pension (NDC) scheme, introduced in 1995 and mandatory for those enrolled in the Gestione Separata, future pensions depend on the success of the working career, i.e. on lifetime earnings, on the frequency of unemployment spells (a risk worsened when notional contributions4 are not paid in the event of unemployment) and on the contribution rate (Palmer, 2006; Jessoula, 2012; Jessoula and Raitano, 2017). As a consequence, if para-subordinate workers are disadvantaged along these three dimensions, they risk receiving low pensions when retired. Furthermore, para-subordinate workers tend not to enrol in private pension funds due both to liquidity constraints (they very often receive low wages and are exposed to short-term unemployment risks) and administrative constraints: waiting and vesting rules requiring minimum enrolment periods hamper the participation of workers with insecure employment prospects. Moreover, occupational funds – defined by contractual agreements – usually do not include para-subordinate workers. And para-subordinate workers are not entitled to TFR (Trattamento di Fine Rapporto, a form of severance pay that can be used to finance contributions to private pension funds) (Jessoula and Raitano, 2017).
To deal with these issues the Italian government has followed a three-pronged strategy over the past decade: i) a process of gradual increases in contribution rates and welfare guarantees for para-subordinate workers to improve coverage for certain events (e.g. retirement, unemployment, maternity) and to reduce the advantage to employers – in terms of lower labour costs – of para-subordinate arrangements; ii) the introduction of stricter regulations aimed at detecting “false” para-subordinate arrangements for both collaborators and professionals5 (Law No. 92/2012); iii) the abolition of some types of contractual arrangements, namely, apart from some exceptions, project collaborations and continuous and co-ordinated collaborations (since 2016 and 2018 respectively; Decree No. 81/2015) and “vouchers” (replaced by a new type of arrangement for additional workers since July 2017; Decree No. 50/2017).
The debate about para-subordinate workers has been complicated also by the difficulty of finding comparable data to provide an extensive picture of the phenomenon. In fact, confusion about the extent of para-subordinate jobs emerged in the past due to the different procedures followed by various institutions to record the phenomenon (Raitano 2007). Indeed, while, on the one hand, survey data (e.g. the Labour Force Survey) identify as a para-subordinate those who self-report as “exclusively” working under such an arrangement at the time of the interview, administrative data (managed by the INPS) record all individuals who paid at least one contribution as a para-subordinate collaborator or professional over a period of a year (even if, during that year, these individuals also performed non-para-subordinate activities). As a consequence, the number of para-subordinate workers is higher when assessed through administrative data than through survey data. However, INPS data have the advantages of being based on an objective status rather than individual self-reporting, precisely recording the type of activity performed by a para-subordinate worker, and distinguishing between collaborators and professionals, on the one hand, and exclusive and non-exclusive para-subordinate workers, on the other hand. In this chapter we will therefore rely on administrative data from the INPS to describe the extent and characteristics of para-subordinate workers.
This chapter is organised as follows. After having described the welfare provisions that para-subordinate workers are entitled to receive (Section 6.2), we present the main evidence on trends relating to individuals working in the various types of para-subordinate arrangements (Section 6.3) and on their characteristics (Section 6.4). Afterwards, we make use of an innovative sample dataset, called AD-SILC, built by merging INPS data with the Italian version of the EU-SILC (European Union Statistics on Income and Living Conditions), to compare distributions of earnings and weeks worked for para-subordinate workers and employees (Section 6.5), and to observe, in a dynamic perspective over a ten-year period, employment trajectories (Section 6.8) and the accrual of pension contributions (Section 6.8) by a cohort of younger workers in order to assess whether para-subordinate workers are disadvantaged in the early phase of their careers6. Section 6.9 summarises the main evidence and proposals raised to improve conditions for para-subordinate workers.
6.2. Welfare entitlements for para-subordinate workers
Table 6.1. Labour costs are still lower for para-subordinate workers
Contribution rates and welfare benefits by type of employment arrangement, 2018
Note: TFR is Trattamento di Fine Rapporto, a form of severance pay entitlement. Data about para-subordinate workers provided by the INPS in the Osservatorio sui lavoratori parasubordinati include collaborators and professionals but not workers on vouchers, whose information are collected in another archive, the Osservatorio sul lavoro accessorio. Therefore, the table lists workers on vouchers separately.
Table 6.2. Social protection coverage of para-subordinate workers
Note: TFR is Trattamento di Fine Rapporto, a form of severance pay entitlement.
* The unemployment benefit for para-subordinates, DS-COLL, is payable for half of the months worked between 1 January of the previous year and the end of the contract, up to six months. The standard unemployment benefit is payable for half of the months worked over the previous 48 months, up to 24 months.
In line with the classification used in the previous section, based on the categorisation of para-subordinate workers used by INPS in the administrative archives, we distinguish between collaborators (e.g. continuous, project and occasional collaborators plus other workers included as collaborators such as the administrators and accountants of firms and postgraduate and postdoctoral students) and professionals. We also divide both groups into “exclusive” para-subordinate workers – when only the para-subordinate activity is performed during the year by a worker who is not retired – and “non-exclusive” para-subordinates – when the worker is a pensioner or also performs non-para-subordinate activities, as an employee (in the public or in the private sector) or as a self-employed person. Note that, in line with normal procedures, we do not include “workers on vouchers” in the para-subordinate category, even if these workers pay contributions to the Gestione Separata. However, at the end of this section, the characteristics and welfare guarantees of “workers on vouchers” – a contractual arrangement that has been substantially revised in 2017 after an intense debate about their misuse by employers – will be also discussed.
Regarding pensions, para-subordinate workers are obliged to pay contributions to the public fund Gestione Separata which (like the whole Italian public pension system) is financed on a pay-as-you-go basis. Workers enrolled in the Gestione Separata will receive a pension calculated through the NDC formula (introduced in Italy by the Law No. 335/1995), which is based on principles of actuarial neutrality: pension benefit is based on the notional accrual of individual contributions over the whole career (a notional rate of return linked to nominal GDP growth is paid annually on the accumulated amount of contributions) and, at retirement, the accumulated amount is converted into an annuity based on average life expectancy at retirement age (Franco and Sartor, 2006).
The NDC formula can thus be considered as a reflection of a worker’s career: for a given set of macroeconomic conditions and demographic trends, and for a given retirement age, individuals’ pensions will depend on contributions paid during their whole working life, i.e. on earnings, the length of (effective or figurative) contribution periods and the contribution rate applied to earnings – in a NDC scheme, lower contribution rates mean lower future pensions. Furthermore, since it both lacks a minimum pension level and applies the same rate of return on contributions to every individual, the Italian NDC scheme has no explicit redistributive features. Consequently, the actuarial public pension pillar systematically transfers labour market inequalities at working age to pension differences at retirement (Jessoula and Raitano, 2017).
In this framework, a specific concern relates to para-subordinate workers, since they often experience low wages and frequent spells of unemployment (see Section 6.5) and until 2018 paid a lower contribution rate than that levied on employees. Contributions for employees stood at 32.7% in the period 1996-2006 and 33% since 2007 (23.81% is paid by the employer, 9.19% by the employee). Pension contribution rates for para-subordinate workers were much lower, especially in past years.
For exclusive collaborators, the contribution rate was just 10% in 1996 and 1997 and increased very gradually to 13.5% in 2002, then to 17.3% in 2004, 23.0% in 2007 and, then by around 1 percentage point per year until it equalled the 33% levied on employees in 2018 (Figure 6.1, Panel A). Pension contributions are not the only contributions paid by the client, however: since 1998 an additional rate (0.50% until 2007, 0.72% afterwards and 1.23% since July 2017) is paid to finance sickness, maternity and family benefits and, since July 2017, unemployment benefits.
Except in 1996 and 1997, contribution rates for non-exclusive collaborators have been lower than those paid by exclusive collaborators and are currently fixed at 24% (in 2003‑06 retired people paid a contribution rate higher than non-exclusive collaborators performing a “second job”, but the rates have been equal since then; Figure 6.1, Panel A). Additional contributions for other welfare guarantees are not levied on earnings of non-exclusive collaborators, who are not entitled to receive welfare benefits other than pensions (Table 6.1). For both exclusive and non-exclusive collaborators one-third of the contribution rate is paid by the worker, and two-thirds by the client.
Contribution rates for exclusive professionals were equal to those levied on exclusive collaborators up to 2013 when the process of gradual increases for professionals stopped. In fact, this rate has been reduced to 25% (plus 0.72% for financing the other types of benefits, apart from unemployment benefits that do not cover professionals) since 2017 (Figure 6.1, Panel B). Non-exclusive professionals have instead been subject to the same contribution rates as non-exclusive collaborators since 1996.
Figure 6.1. Contribution rates for para-subordinates mostly rose since the 1990s
Note: Panel A: Since 1998 the contribution rate is made up of the pension contribution rate plus an additional contribution paid by exclusive collaborators for financing sickness, maternity, unemployment and family benefits. This additional contribution rate was 0.5% in the period 1998-2007, 0.72% since 2008 to June 2017, while since July 2017 the additional rate has been increased to 1.23% to finance the increasing coverage of unemployment benefits for collaborators (DIS-COLL). The rest of the contribution rate is devoted to finance the pension system. Note that the contribution rate is paid for one-third by the worker and two-thirds by the client.
Panel B: Since 1998 the contribution rate is made up of the pension contribution rate plus an additional contribution paid by exclusive professionals for financing sickness, maternity, unemployment and family benefits. This additional contribution rate was 0.5% in the period 1998-2007 and 0.72% since 2008 to June 2017. The rest of the contribution rate is devoted to finance the pension system.
In spite of the steep increase in contribution rates, social contributions on para-subordinate arrangements are still lower than those levied on standard employees. Indeed, while para-subordinate arrangements are subject to a 34.23% contribution rate at the most in 2018, total contribution rates levied on private employees for pensions, maternity, sickness, family benefits and the Cassa Integrazione allowance (paid to an employee when the job relationship is temporarily suspended by the firm) are around 40%, depending on occupation, sector and size of firm (pension contribution rates are, however, constant across sectors, size of firm and occupations).7 Furthermore, private firms also pay an additional compulsory contribution of 6.91% to finance a worker’s TFR (a form of severance payment), which can also be used to finance contributions to private pension funds (Jessoula and Raitano, 2017).
Para-subordinate workers were not entitled to unemployment benefits until recently. An unsuccessful means-tested lump sum benefit was paid on an experimental basis in the 2009-11 period to those working as an exclusive collaborator with a single client and a low gross annual income (no higher than EUR 20 000). This benefit amounted to 30% of earnings with a EUR 4 000 ceiling). An unemployment benefit scheme for para-subordinate exclusive collaborators – named DIS-COLL – was introduced as an experiment by the Jobs Act reform (Decree No. 22/2015) and its provision was made permanent in 2017 (Law No. 81/2017), when an additional 0.51% contribution rate on exclusive para-subordinate earnings was established (Table 6.1). Professionals and non-exclusive collaborators are not entitled to DIS-COLL. A few categories of exclusive collaborators are also not entitled to DIS-COLL and therefore do not have to pay the additional contribution e.g. administrators of local authorities, occasional collaborators, postgraduate training physicians (after intensive discussions DIS-COLL has however been extended to PhD students and postdoctoral fellows).
DIS-COLL is paid for a period equivalent to half of the months worked between 1 January of the previous year and the end of the contract, up to a maximum of six months. The duration of payment is therefore lower than for the ordinary unemployment benefit for employees (NASPI), which is half of the months worked in the previous 48 months, but with a maximum duration of 24 months. The amount of the monthly DIS-COLL is, however, the same as the NASPI benefit: 75% of the average monthly income before the period of unemployment for incomes below EUR 1 195, with 25% of the difference between average monthly incomes and EUR 1 195 added in case of higher incomes, up to a maximum benefit of EUR 1 300. After the fourth month, the amount of the DIS-COLL payment is reduced by 3% per month. Unlike the NASPI, individuals receiving DIS-COLL are not entitled to notional pension contributions.
Exclusive collaborators and professionals (paying the additional 0.72% contribution rate; Table 6.1) are entitled to sickness, maternity and family benefits8.
Annual sickness allowance entitlement is at least 20 days and at most one-sixth of the total duration of the employment contract or 61 days, whichever is lower. The amount of sickness allowance is based on previous earnings and on the length of the contract.
In case of maternity, the contract is suspended for a five-month period and an allowance equal to 80% of the previous wage is paid if the worker paid contributions for at least three months in the previous year. Parental leave amounting to 30% of the previous wage is also available when the child is less than one year old.
Para-subordinate exclusive workers are also entitled to family allowance (ANF, Assegno per il Nucleo Familiare), a contribution-based scheme in which payments depend on household income and size (e.g. in 2017 the monthly benefit for a two-parent household with one child was EUR 137.50 if the household annual income was below EUR 14 383.37).
Finally, we should also focus on additional workers – usually termed “workers on vouchers”, since workers are paid by the client through a pre-printed form sold in shops – a further category that pays contributions to the Gestione Separata. This type of arrangement was introduced in 2003 (Decree No. 276/2003), to reduce the scale of the informal economy, to pay for some occasional jobs performed in a few cases (e.g. domestic care, private home lessons) by a limited number of categories (e.g. housewives, students, long-term unemployed) and with strict limits on the total amounts received every year and from a single client in the form of vouchers. Vouchers were very rarely used until 2008 when they were allowed in agriculture and a EUR 10 gross value for each voucher was established. This equates to a net wage of EUR 7.50 for the worker after taking account of a 13% pension contribution rate, while no other welfare guarantee is provided9.
Subsequently, several deregulations of the economic sectors and categories of workers entitled to use vouchers were introduced, especially in 2012 (Law No. 92/2012), when the usage of vouchers was extended to all sectors and all types of workers (including employees doing additional jobs) and in 2015 (Decree No. 81/2015) when the limits on voucher payments were increased (a worker can receive no more than EUR 7 000 in overall annual voucher payments and a maximum of EUR 2 020 from the same client).
The number of workers paid with vouchers increased strongly in recent years (see Section 6.3), possibly suggesting that the arrangement was being misused to replace regular employment or to pay some working hours performed by informal workers in case of administrative checks, and the left-wing trade union CGIL proposed a referendum on the abolition of vouchers. The government reformed the rules concerning additional jobs, limiting the type of firms allowed to use vouchers and increasing the pension contribution rate to 33% (16.5% for domestic activities), while no further welfare guarantees have been introduced (Decree No. 50/2017, phased in July 2017). The “new vouchers” can only be used by firms with up to five employees and by the public sector in some limited cases and a firm cannot pay vouchers for more than EUR 5 000 anually; a worker can annually receive as a voucher no more than EUR 5 000, with a limit of EUR 2 500 from a single client. Net hourly wage cannot be lower than EUR 9 and the daily net wage has to be equal to at least EUR 36.
6.3. Para-subordinate workers: the extent of the phenomenon
Using administrative data from the INPS which record workers paying at least one contribution a year to the Gestione Separata fund, a picture about the evolution over time of the number of para-subordinate workers is provided in this section, which also distinguishes between the different types of para-subordinate workers and highlights the evolution over time of additional “workers on vouchers”.10
The number of para-subordinate workers (collaborators and professionals, irrespective of whether they are exclusive or non-exclusive; Figure 6.2) hugely increased between 1996 and 2007: during that period the total number of contributors to the Gestione Separata increased 126% (129% for collaborators and 106% for professionals), while, in the same period, according to Eurostat figures, based on the Labour Force Survey, the mean number of employees and total workers over a year increased, respectively, by 19% and 14%. In 2007 around 1.9 million workers had a para-subordinate arrangement, and the number of collaborators was very large too (around 1.67 million workers versus around 220 000 professionals).
Afterwards, while the number of professionals increased almost constantly over the whole observation period to reach around 327 000 in 2016, the number of collaborators fell sharply to around 917 000 in 2016 (Figure 6.2). The number of collaborators is expected to decrease further in future years since, as mentioned, the Jobs Act reform of 2015 banned new project collaborations from 2016 and new co-ordinated and continuous collaborations from 2018. The only exclusive collaborations still allowed are occasional collaborations and collaborations in which the worker is clearly independent from the client (for instance, to be counted as an “independent collaboration” there can be no working hours schedule and the collaborator cannot be obliged to work in the client’s office).
Figure 6.2. The number of para-subordinate collaborators follows an inverted U-shape
Source: Own calculations based on INPS data “Osservatorio sui lavoratori parasubordinati”.
Distinguishing between para-subordinate workers on the basis of whether they are exclusive or non-exclusive, there has been a clear reduction in the number of exclusive collaborators, which has only been partially offset by an increase in the number of exclusive professionals (Figure 6.3). In the period 2011-16 the number of exclusive collaborators declined 49.6%, while the number of exclusive professionals increased 23.6%.
Therefore, it has to be emphasised that the decrease in the extent of exclusive collaborations came before the more binding regulatory constraints introduced by the 2012 reform and particularly by the 2015 reform. The decline started after 2007, when, as shown in the previous section, a process of gradual increases in the pension contribution rates for para-subordinate workers started.
As a consequence, in 2016 the number of exclusive collaborators is almost equal to that of non-exclusive collaborators (484 972 versus 432 535, respectively) while exclusive and non-exclusive professionals amount, respectively, to 241 338 and 85 600 individuals (Figure 6.3). This evidence is therefore consistent with the idea that the widespread use of para-subordinate collaborations of the previous decade was mainly due to the lower cost of these arrangements rather than to a real need to use collaborators rather than employees.
Figure 6.3. The number of exclusive para-subordinate collaborators fell in recent years
Source: Own calculations based on INPS data “Osservatorio sui lavoratori parasubordinati”.
Detailed data collected in the INPS (2017) report also show that the large majority of collaborators work for a single client (89.6%) while only 2.1% have at least three clients in a year. There were 174 035 collaborators who still had a project collaboration in 2016 (compared with 376 774 in 2015), while around 34 455 individuals worked in a co-ordinated and continuous collaboration in the public sector. The number of collaboration contracts based on the definition of “independence” of the collaborator according to the content of the Decree No. 81/2015 was 48 671 in 2016. Within the category of exclusive collaborators, there were 51 665 PhD and postdoctoral students, 34 376 postgraduate physicians in training and 504 315 administrators and accountants are also included.
This latter category is, as expected, paid more than other categories: on average, para-subordinate collaborators working as administrators and accountants had gross earnings of EUR 32 682 in 2016, compared with mean earnings for all collaborators of EUR 22 813.
Finally, it is interesting to report that, among the 432 553 non-exclusive collaborators in 2016, 19.3% were pensioners and 20.1% employees (in the public or the private sector), with the remainder accounted for by self-employed people (especially craftsmen and dealers). Interestingly, consistent with the evidence that pensioners who combine pensions and labour income were often high-skilled workers, mean annual earnings from non-exclusive collaborations performed by pensioners were in 2016 around EUR 30 000.
Finally, data on the trend of the number of vouchers sold to pay additional workers confirm the dramatic rise of the usage of this type of arrangement (Figure 6.4, Panel A). In 2016 over 130 million vouchers were sold to pay 1 765 810 individuals. From 2012 to 2016 the number of vouchers increased by 482%. This trend has reversed following the substantial reform of the voucher scheme, see Section 6.2).
Figure 6.4. The number of vouchers increased rapidly, and incomes on vouchers are low
Source: Calculations based on INPS data “Osservatorio sul lavoro accessorio”.
The distribution of the yearly amounts received in vouchers (Figure 6.4, Panel B) shows that very few individuals received more than EUR 3 000 annually (2.6% in 2016), while the large majority earn very low incomes: 75.1% of additional workers received less than EUR 1 000 in 2016. These figures clearly show that if work paid by vouchers is the only activity performed in a year by a worker, the amount received is totally inadequate to provide a decent wage. And if individuals are only working through vouchers, this points to them working more hours informally than they are paid through vouchers (Anastasia et al., 2016).
6.4. The main characteristics of para-subordinate workers
A disadvantage of INPS data is that, because they are based on administrative forms filled in by employers when contributions are paid, detailed workers’ characteristics are not recorded (e.g. education, marital status, number of children, parents’ background; INPS data, instead, record sex, age and area of work). In particular, information about the education of those working as a para-subordinate is crucial to better assess the characteristics of these workers.
To this end, in the rest of the chapter we will mainly rely on the micro-data collected in the AD-SILC dataset, a longitudinal dataset built by adding INPS data to the IT-SILC waves 2004-2012. Thus, the individual information recorded in IT-SILC has been enriched by the longitudinal information about the working history of each individual in the sample (e.g. recording for each working relationship, the type of arrangement, identified by the specific INPS fund where the workers pay contributions – which allows us to distinguish between para-subordinate collaborators and professionals – weeks worked and gross earnings).
Because individuals may have various contractual arrangements during a year, we recognise the working status at a point in time (31 December), thus identifying those exclusively working as a para-subordinate at that moment (if at 31 December an individual also has an employment or a self-employment arrangement or is retired he/she is therefore not classified as a para-subordinate worker).
Interestingly, data at our disposal show that the use of para-subordinate arrangements is much wider among those with a tertiary education qualification than among those with a lower level of education. Indeed, on the one hand, the distribution by education of private employees, para-subordinate collaborators and professionals in 2013 (Figure 6.5) shows that the share of those with a tertiary education is much higher in the two para-subordinate categories (34.0% and 46.8% of collaborators and professionals, respectively) than among private employees (11.2%), and conversely the share of low-skilled individuals (those with at most a lower secondary educational qualification) in para-subordinate arrangements is small. Similarly, 7% of tertiary graduates were para-subordinates at the end of 2013 (compared with a peak of over 10% in 2003), while the corresponding shares of those with at most upper secondary or lower secondary educational qualifications, respectively, 3.4% and 1.7% (not shown).
Figure 6.5. Para-subordinate workers are well educated on average
Source: Own calculations based on AD-SILC data.
The use of para-subordinate arrangements among the well-educated may be due to compositional effects (e.g. para-subordinate arrangements are often used in the case of younger workers who are, on average, more educated) and to the fact that some specific activities performed only or mostly by tertiary education graduates (e.g. postgraduate fellows, administrators or accountants) are classified as para-subordinate jobs. However, multivariate econometric estimates confirm that, even after controlling for the type of job and characteristics of workers, the probability of working in a para-subordinate arrangement is much higher for tertiary graduates than for less educated individuals (Raitano, 2007). This might be driven by the specific characteristics of some highly skilled jobs, which mean that these jobs are more frequently carried out via para-subordinate work arrangements. Since survey data do not contain detailed job descriptions, this cannot be confirmed in this analysis.
As argued above, para-subordinate arrangements might mask – among both collaborators and professionals – actual employment relationships in which the employer uses a para-subordinate arrangement to save labour costs. The 2012 and 2015 reforms tried to avoid the misuse of these arrangements by adding requirements confirming that the para-subordinate worker is really independent.
To assess how many para-subordinate arrangements concealed what are effectively subordinate relationships between the employer and the worker before the recent reforms, a specific survey carried out by the ISFOL (Istituto per lo sviluppo della formazione professionale dei lavoratori – Institute for developing workers’ lifelong learning) on a sample of workers in 2010 is very useful. Indeed, this survey recorded how many para-subordinate workers reported in interviews that their job had some features that should not be typical of para-subordinate jobs, which suggests that there was actually a subordinate relationship between the employer and worker.
The figures in the ISFOL survey are worrying since they show that the large majority of para-subordinate workers performed activities similar to those of employees, which suggests that many para-subordinate arrangements masked subordinate relationships (Figure 6.6): indeed, the large majority of those interviewed reported having accepted a para-subordinate arrangement at the client’s request (70.5%), to work in the client’s office (71.7%), to follow prearranged working hours (67.0%) and because of a desire to convert their flexible para-subordinate arrangement into an open-ended employment contract (75%). Thus, rather than choosing this type of arrangement voluntarily, many workers seem to have been forced by employers to accept a very flexible arrangement, offering lower welfare guarantees and, very often, lower wages than employees.
Figure 6.6. Para-subordinates report working under conditions similar to those of employees
Source: Own calculations based on ISFOL-PLUS 2010 data.
6.5. Earnings distribution by contractual arrangements
As pointed out in previous sections, para-subordinate workers have less generous welfare entitlements than employees, or none at all. It is therefore interesting to analyse whether this disadvantage is compensated for by higher rewards – in terms of earnings or less frequent unemployment spells – or whether, on the contrary, lower welfare guarantees are associated with shorter working periods and lower earnings on average over a year.
To this aim, in AD-SILC we have identified the “main” working condition during a year as the contractual arrangement accounting for the highest number of working weeks in a year.
Focusing on total annual working weeks as the main working condition experienced during a year, a clear gap in terms of weeks worked emerges (Figure 6.7). While, on average, in the period 1996-2013 the mean number of weeks worked by private employees was fairly constant at around 44, total weeks worked by collaborators and professionals never reached this level. The gap in weeks worked between employees in the private sector and para-subordinate collaborators and professionals amounted to 7.5 and 7.0 weeks respectively in 2013.
Figure 6.7. Para-subordinates are less likely to work year-round than dependent employees
Note: All workers in the sample are assigned a “main” contractual agreement in any given year (the one with the highest number of weeks worked). Number of weeks worked are then calculated irrespective of contractual agreement. Retirees are excluded from the analysis. The low value in 1996 is related to the fact that the Gestione Separata fund was introduced in that year and, therefore, many workers started to work as a para-subordinate during the year. Annual weeks worked exclude paid annual leave.
Source: Own calculations based on AD-SILC data.
Para-subordinate workers also receive lower wages than employees (Figure 6.8, Panel A). Focusing on median earnings (which allows us to exclude the impact of possible outliers at the top of the earnings distribution) a large gap between private employees and collaborators emerges, while the gap between employees and professionals grew in 2012 and 2013, probably due to the effects of the economic crisis. In 2013, a median collaborator’s annual gross earnings were 33.8% lower than the median private employee wage, while professionals’ earnings were 31.8% lower.
Figure 6.8. Para-subordinate earnings are lower and more dispersed than those of employees
Note: This analysis only includes earnings from the main earnings source in any given year; that is, collaborators’ earnings refer to annual earnings from work as a para-subordinate collaborator, received by those who worked prevalently as a collaborator in that year. Earnings from other contractual arrangements, as well as the earnings of those who have a contractual arrangement as a para-subordinate collaborator as a secondary earnings source, are disregarded.
Source: Own calculations based on AD-SILC data.
However, median incomes do not provide a full picture of para-subordinate living standards since, as already noted, para-subordinate arrangements are very heterogenous and include, for instance, both well-paid activities (such as those performed by administrators and accountants) and low-paid and short-term occasional collaborations.
The evidence of a large heterogeneity emerges when calculating the Gini income of inequality of gross earnings received in a year by those mainly working under the various contractual arrangements (Figure 6.8, Panel B). Gini coefficient values were fairly constant over the period observed, but the level of earnings inequality between para-subordinate professionals and, particularly, para-subordinate collaborators is much higher than inequality between private employees.
The picture of large inequalities between para-subordinate workers is also confirmed when looking at the distribution by income brackets of annual earnings in 2013 (Figure 6.9). Indeed, 39.2% and 36.6% of those “mainly” working as collaborators and professionals, respectively, earned less than EUR 10 000 in 2013, while the corresponding share was 24.8% among private employees (part-time and fixed-term employees are in the bottom tail of the employed earnings distribution). Conversely, the share of high-paid workers – i.e. those earning more than EUR 75 000 in 2013– is higher among collaborators (5.3%) and professionals (2.3%) than among private employees (1.3%).
Figure 6.9. Among para-subordinates, both high and low earners are overrepresented
Note: This analysis only includes earnings from the main earnings source in any given year; that is, collaborators’ earnings refer to annual earnings from work as a para-subordinate collaborator, received by those who worked prevalently as a collaborator in that year. Earnings from other contractual arrangements, as well as the earnings of those who have a contractual arrangement as a para-subordinate collaborator as a secondary earnings source, are disregarded.
Source: Own calculations based on AD-SILC data.
6.6. Para-subordinate workers’ working lives: a dynamic perspective
Risks related to para-subordinate arrangements cannot be assessed simply by looking at individuals’ labour market situations at a particular point in time, because individual employment trajectories need to be assessed over a longer time span. To this purpose, we have made use of the longitudinal AD-SILC dataset and observed the employment trajectories of individuals who entered the labour market in the 1996-2001 period over the ten years following the year of entry, using some indicators of poor careers related to unemployment spells and periods spent working in para-subordinate arrangements. In Section 6.8 we will also assess risks related to the starting phase of the working career by calculating total contributions accumulated in the NDC scheme over a ten-year period, which acts as both an indicator of the success of the early phase of the career and, in the case of limited accumulation of pension contributions, as a signal of a possible risk of old-age poverty in the future if the working career does not improve over the subsequent decades.
We first focus on new labour market entrants and show the share of individuals who started working in a para-subordinate arrangement (Figure 6.10, Panel A). This can only be calculated until 2011, since those working in the public sector are not recorded in AD-SILC after that year. The share of new entrants working as para-subordinate collaborators steadily increased over time until it reached a peak of 15.5% in 2008 and then declined to 11.1% in 2011 – maybe as an effect of the aforementioned rise in social security contributions for para-subordinate workers – while the share of new entrants who were professionals remained fairly constant in the 1-1.5% range.
Consistent with the evidence of Section 6.4, it is mostly tertiary education graduates who start working in a para-subordinate arrangement, while the share of workers with a lower level of education entering the labour market through a para-subordinate arrangement is much lower, since many para-subordinate jobs require high-skilled workers (Figure 6.10, Panel B). For instance, in 2009 the first working contracts of 29.3% of tertiary graduates were as para-subordinates, compared with peaks of 18.2% in 2008 for those with an upper secondary education and 6.0% in 2007 for those with at most a lower secondary qualification.
Figure 6.10. The number of labour market entrants working as para-subordinates peaked in 2009
Note: Labour market entrants are under 35-year-olds who, in a given year, worked 13 weeks or more for the first time.
Source: Own calculations based on AD-SILC data.
However, para-subordinate arrangements at entry might either represent a stepping stone towards a better paid and guaranteed arrangement or a sort of trap (a “dead end”) in less advantaged jobs (Muehlberger and Pasqua, 2006; Picchio, 2008; Berton et al., 2011).
The employment trajectories tracked in INPS data show that the persistency of para-subordinate arrangements is not too high: among new entrants in 2000, 19.2% still worked as para-subordinates in 2005, while 36.4% moved towards private dependent employment and 20.2% towards other self-employed activities. However, the share of those who were unemployed or inactive was very high (21.3%).
To better assess employment trajectories, we have used AD-SILC data – which, as mentioned, precisely record the number of weeks worked each year under the various types of contractual arrangements – to extract a sub-sample of around 13 000 individuals, and followed those entering activity in 1996-2001 over a ten-year period (i.e. 1997-2006 for entrants in 1996, 1998-2007 for entrants in 1997 etc.), distinguishing by types of arrangement at entry (Figure 6.11).11 On average, the share of workers spending periods working as a para-subordinate is low: only 2.3% of all workers and 0.9% of those who started on a non-para-subordinate contract spent at least five of the ten years as a para-subordinate worker. However, the number of years spent as a para-subordinate worker for those who started working as a collaborator or a professional is higher: indeed, 32.7% of these spent at least three of the ten years working as a para-subordinate.
Figure 6.11. Those who start their careers as para-subordinates often continue on this path
Note: Labour market entrants are under 35-year-olds who, in a given year, worked 13 weeks or more for the first time. The first 10 years start in the year after labour market entry.
Source: Own calculations based on AD-SILC data.
In general, what clearly emerges from the AD-SILC data is that the early phase of a career is characterised by frequent periods spent not working, irrespective of the contractual arrangement at entry, thus confirming a picture of widespread vulnerability among younger workers in Italy since the mid-1990s (Figure 6.12, Panel A). On average, new entrants in 1996-2001 spent around a quarter of the following ten-year period (i.e. 2.5 years) not working and not receiving unemployment benefit (the share of young workers receiving unemployment benefit in Italy is rather limited).
However, while unemployment risks are similar irrespective of the initial contract, those who start out as a para-subordinate worker do spend a much higher share of their weeks worked in para-subordinate jobs.
Figure 6.12. Those who start as para-subordinates do not work less than others
Note: Labour market entrants are under 35-year-olds who, in a given year, worked 13 weeks or more for the first time. The first 10 years start in the year after labour market entry.
Source: Own calculations based on AD-SILC data.
6.7. Actual accumulation of pension contributions by para-subordinate workers since 1996
The success of a working career over a certain period can be evaluated by measuring total pension contribution accrued over that period (according to the rules of the Italian NDC pension system, the annual rate of return on accrued contributions is equal to the five-year average of the GDP nominal growth rate). This is a useful indicator that summarises both the outcomes of various dimensions of a career over a period of several years (i.e. earnings, which are also influenced by the frequency of periods of part-time working, periods spent not working or receiving unemployment benefit, periods spent working in contractual arrangements that pay a reduced contribution rate, such as para-subordinate arrangements) and to highlight possible risks of inadequate future pensions due to a poor accrual of contributions in the early phase of the career in the NDC pension system where, it should be noted, future pensions strictly depend on contributions paid over the whole working life.
To assess the size of pension contribution accrued over the ten-year period following labour market entry for new entrants in 1996-2001 we used a relative approach, i.e. we calculated the relative accumulation with respect to the accumulation potentially achieved by a representative worker. We then chose as a benchmark an individual who worked continuously over the ten-year period as a full-time dependent employee in the private sector (thus paying the 33% contribution rate) and always earning the gross median wage (amounting, in real terms, to around EUR 20 000 per year). Consistent with the relative poverty approach, we then identified those who accumulated less than 60% of the potential accumulation of the ideal full-time, never unemployed median private sector employee over the ten-year period as poor in terms of pension contributions.
Our calculations clearly show that the ten-year period after labour market entry has not been successful for a high share of Italian workers in previous years. Indeed, just under half of workers that entered the labour market in the period 1996-2001 accumulated less than 60% of the hypothetical median employee (Figure 6.13): for those who started work as a para-subordinate, the share of those with poor contributions accumulation – which can be also regarded, as mentioned, as an indicator of a risk of a poor future pension – is 46.7%, but it is also extremely high among those who started work as an employee in the public or private sector (40.9%). As expected, the risks of an unsuccessful career are higher for females than for males and for workers with a lower level of education than for tertiary education graduates, but these risks are also fairly high among males and high-skilled workers (Figure 6.13).
Figure 6.13. Many have low pension contributions, but para-subordinates are more vulnerable
Note: Labour market entrants are under 35-year-olds who, in a given year, worked 13 weeks or more for the first time. The first 10 years start in the year after labour market entry. Workers “at-risk of low future pension” have, during the first 10 years of their career, accumulated less than 60% of the amount a person who has worked consistently on a full-time basis over 10 years at the median wage would have contributed.
Source: Own calculations based on AD-SILC data.
A crucial factor in the success of the starting phase of the career is the continuity of the employment period. As expected, indeed, risks of low accumulation (and potential pension inadequacy in the future) reduces when the number of years effectively worked increases, even if a worrying 20.6% were rated “poor” in terms of accumulated contributions over the 10-year period among those who worked at least eight of the ten years, due to low earnings and periods spent working in arrangements with reduced contribution rates (left panel in Figure 6.14).
Conversely, consistent with the heterogeneity of the types of para-subordinate workers, risks of poor accumulation do not increase linearly with the number of years spent working as a para-subordinate. Indeed, the share of workers with poor accumulation, while always very high, is lower among those who worked at least five years as a para-subordinate – both “strong” para-subordinates such as accountants and administrators and very disadvantaged workers belong to this group – than within the group of those who worked a lower number of years as a para-subordinate – poor workers with very fragmented careers may, indeed, belong to this group (right panel in Figure 6.14).
Figure 6.14. Low pension accruals are mainly driven by disrupted employment histories
Note: Labour market entrants are under 35-year-olds who, in a given year, worked 13 weeks or more for the first time. The first 10 years start in the year after labour market entry. Workers “at-risk of low future pension” have, during the first 10 years of their career, accumulated less than 60% of the amount a person who has worked consistently on a full-time basis over 10 years at the median wage would have contributed.
Source: Own calculations based on AD-SILC data.
6.8. Concluding remarks
Using administrative data from the INPS, in this chapter we have surveyed the evolution over time of the number of individuals working in para-subordinate arrangements in Italy, in light of the evolution of the rules governing hiring through the various types of arrangements and the welfare guarantees applied to them, which clearly affect the relative labour cost of para-subordinate workers and employees.
The disadvantages in terms of welfare guarantees for para-subordinate workers have been reduced by the recent reforms, even if some gaps still exist, particularly concerning unemployment benefit. However, those who worked as a para-subordinate in previous years were clearly disadvantaged by the reduced contribution rates applied to such contracts (below 20% until 2007) which will have a major impact on future pensions from the Italian public pension scheme where, being based on an NDC formula, pensions depend on contributions paid during the whole career.
Recent reforms have also tightened the requirements for clients signing a para-subordinate arrangement and have ended the very disadvantageous – in terms of wages and welfare guarantees – “voucher arrangement” which has been replaced by a more favourable and less deregulated arrangement since the middle of 2017.
There has probably also been a reduction in the use of para-subordinate collaborations in recent years due to fiscal incentives (i.e. a strong reduction in social contributions paid by firms) introduced by the 2015 Jobs Act reform for those who hired new workers through open-ended employment contracts in 2015 and 2016.
However, as pointed out by the dramatic increase in the number of “workers on vouchers” before the 2017 reform – also due to limits on hiring through para-subordinate collaborations – Italian firms continue to search for opportunities to hire less costly and more flexible workers and are liable to use these contracts even when they mask what are effectively subordinate working relationships.
Also to avoid the adverse effects related to the existence of para-subordinate arrangements – even if, as noted, their use was made more difficult by the recent reforms – the government introduced in the Stability Law for 2018 a reduction in social contributions paid by firms for the first two or three years for young workers hired through an open-ended employment arrangement.
Relying on a temporary reduction in social contributions to improve workers’ prospects – a measure that is costly for the public finances since it reduces contributions to the pay-as-you-go public pension system12 – assumes a belief in a path dependency of the employment trajectory, according to which the contractual arrangement at entry influences the following career path. However, in past years, employment trajectories have often varied widely irrespective of the initial contractual arrangement (Fabrizi and Raitano, 2012). While young people who started their careers as para-subordinates in the late 1990s were more likely to continue to work as para-subordinates, they are not significantly less likely to work overall than those who started out in other contractual arrangements. Therefore, reducing the cost of the employment arrangement at labour market entry might not be sufficient to improve employment trajectories that, as summarised in Sections 6.7 and 6.8, have proved to be unsuccessful for a very high share of workers in past years.
A particular concern relates to the future pensions of those with unsuccessful career patterns over a long period (in terms of earnings, unemployment spells, atypical and less guaranteed contractual arrangements). Indeed, in an NDC scheme (as the Italian public pension system is for those who started work since 1996) less successful careers will be exactly mirrored by lower pensions and there are no redistributive tools, apart from personal income tax progressivity and the existence of means-tested welfare benefits for the elderly poor. Even if Italian labour market conditions suddenly improve from 2018 onwards, as shown in Section 6.7, a high share of individuals have so far only accumulated low contributions in the NDC pension scheme – partly due to periods spent working as para-subordinates – and will therefore risk receiving a low pension at retirement if their career pattern does not significantly improve.
Some measures to improve expected NDC pensions of workers with fragile careers have been proposed recently by some scholars and have been discussed by the government and the main trade unions (for a survey, see Raitano, 2017). Some proposals are based on the introduction of a flat “citizenship” pension to be associated with an NDC scheme with a reduced contribution rate (to cut labour costs), but these proposals would be very expensive for the public budget due to the immediate reduction in pension contributions in the pay-as-you-go system. Some relate to the possibility of extending the amount of social assistance benefits paid in future years to poor pensioners, while other proposals suggest introducing a “minimum pension” in the NDC scheme with the amount increasing with the length of the individual career, in line with the principle of “making contributions pay”. These last two types of measures would have the advantage of increasing public spending only after around 2040, when the first cohort of individuals wholly enrolled in the NDC scheme start to retire. However, after an intense debate following the agreement between the government and the trade unions signed in September 2016 and the promise of measures to improve the pension prospects of younger workers, discussions about how to provide future guarantees for poor pensioners in the NDC scheme have stopped and no measure on this is going to be included in the Stability Law for 2018.
References
Anastasia B., Bombelli S., Maschio S. (2016). Il Lavoro accessorio dal 2008 al 2015. Profili dei lavoratori e dei committenti. WorkINPS papers, n.2.
Berton F., Devicienti F., Pacelli L. (2011). Are temporary jobs a port of entry into permanent employment?: Evidence from matched employer-employee. International Journal of Manpower, vol. 32(8), pp. 879-899.
Berton F., Richiardi M., Sacchi S. (2012). The Political Economy of Work Security and Flexibility: Italy in Comparative Perspective, Policy Press.
Fabrizi E., Raitano M. (2012). Rigido, flessibile o liquido? L’immagine del mercato del lavoro italiano dal dataset AD-SILC, Economia e Lavoro, n.3/2012.
Franco D., Sartor N. (2006). NDCs in Italy: unsatisfactory present, uncertain future. In Holzmann R., Palmer E. (eds.), Pension reform: issues and prospects for ndc schemes, Washington: World Bank Publishing.
INPS (2017). Il lavoro parasubordinato. Aggiornamento all’anno 2016. Rome.
Jessoula M. (2012). A risky combination in Italy: selective flexibility and defined contribution pensions. In Hinrichs K., Jessoula M. (eds.), Labour market flexibility and pension reforms, Palgrave Macmillan.
Jessoula M., Raitano M. (2017). Italian pensions from “vices” to challenges: assessing actuarial multi-pillarization twenty years later. In Natali D. (eds.), The New Pension Mix in Europe. Peter Lang Publishing.
Muehlberger U., Pasqua S. (2006). The “Continuous Collaborators” in Italy. Hybrids between Employment and Self-employment?. CHILD Working Papers no. 10/2006.
Palmer E. (2006). What Is NDC?. In Holzmann R., Palmer E. (eds.), Pension reform: issues and prospects for NDC schemes, Washington: World Bank Publishing.
Picchio, M. (2008). Temporary jobs and transition to stable jobs in Italy, Labour, Vol. 22 No. 1, pp. 147-74.
Raitano M. (2007). I lavoratori parasubordinati: dimensioni del fenomeno, caratteristiche e rischi. Economia e Lavoro, n. 3/2007, pp. 135-160.
Raitano M. (2017). Poveri da giovani, poveri da anziani? Prospettive previdenziali e vantaggi della pensione di garanzia. OCIS Social Cohesion Papers No. 3.
Raitano M., Fana M., Patriarca F. (2017). ll costo e l’efficacia degli sgravi contributivi connessi al Jobs Act. In Pizzuti F.R. (eds), Rapporto sullo Stato Sociale 2017, Sapienza University Press, Roma
Notes
← 1. To be considered as continuous the contract duration shall be at least 30 days per year. If this requirement is not fulfilled the collaboration is considered “occasional”. Project collaborations are based on a specific project to be carried out by the worker (i.e. the worker cannot simply carry out repetitive tasks).
← 2. Note that data about para-subordinate workers provided by INPS in the Osservatorio sui lavoratori parasubordinati include collaborators and professionals but not workers on vouchers, whose information are collected in another archive, the Osservatorio sul lavoro accessorio. As a general rule, if not differently specified, we thus refer to collaborators and professionals when we refer to para-subordinate workers.
← 3. Collaborators’ contributions are paid by the client to the Gestione Separata, while professionals pay contributions on their own. Note also that in Italy liberal professionals with a professional association pay pension contributions to specific social security funds privately managed by their professional association.
← 4. Within the regular unemployment insurance system (e.g. NASPI), the government pays notional (or imputed) pension contributions for the unemployed, see Section 6.2.
← 5. Law 92/2012 introduced an “assumption of a subordinate job” – for both collaborators and professionals – when at least two of the three following conditions are fulfilled: i) the arrangement for the same client lasts at least eight months over a two-year period; ii) the workers receives at least 80% of his/her earnings from the same client in a two-year period; iii) the worker has a regular workstation in the client’s office. When two of these conditions are fulfilled the worker is considered to be an employee and the firm is fined.
← 6. I thank Fondazione Giacomo Brodolini for allowing me access to the AD-SILC dataset.
← 7. For instance, the total contribution rate is 41.57% for blue-collar workers and 39.35% for white-collar employees in industrial firms with at least 50 employees; 40.07% and 37.85% respectively for firms with up to 15 employees; and 39.37% for both groups in commercial services companies of over 200 employees.
← 8. All Italian workers are also insured against workplace and occupational injuries through a specific contribution to INAIL (Istituto nazionale Assicurazione Infortuni sul Lavoro). All workers participating in the social insurance system, are also entitled, in the case of disability, to the "ordinary disability allowance" (if the percentage of disability is at least 66%) or to the "disability pension" (if the percentage of disability is 100%).
← 9. Additional workers are guaranteed against workplace injuries as they pay a 3.5% contribution rate to INAIL (7% before the 2017 reform). The gross value of vouchers also includes EUR 0.50 to cover administrative costs.
← 10. As already mentioned, because these data record the number of individuals who paid, at least once, a contribution to the Gestione Separata, the figures are different from those provided by survey samples (e.g. the Labour Force Survey) which record the number of individuals who self-report as performing a para-subordinate activity at the time of the interview.
← 11. The analysis does not distinguish between non-para-subordinate contractual contracts, such as temporary vs. permanent contracts because, (i) no significant differences between different non-para-subordinate contractual arrangements persist in the medium term (after five years; Fabrizi and Raitano, 2012) and (ii), the INPS only provides information on permanent vs. temporary contracts since 1998, reducing the sample size of our exercise.
← 12. Raitano et al. (2017) estimate that the contribution reliefs introduced by the Jobs Act reform in 2015 cost around EUR 15-20 billion (according to the various scenarios) in terms of missed revenues for the pension system in the 2015‑18 period.