Jump to ContentJump to Main Navigation
The political economy of work security and flexibilityItaly in comparative perspective$

Fabio Berton and Matteo Richiardi

Print publication date: 2012

Print ISBN-13: 9781847429070

Published to Policy Press Scholarship Online: January 2013

DOI: 10.1332/policypress/9781847429070.001.0001

Show Summary Details

Flexibility and employment security: an analysis of work careers

Flexibility and employment security: an analysis of work careers

Chapter:
(p.61) Four Flexibility and employment security: an analysis of work careers
Source:
The political economy of work security and flexibility
Author(s):

Fabio Berton

Matteo Richiardi

Stefano Sacchi

Publisher:
Policy Press
DOI:10.1332/policypress/9781847429070.003.0004

Abstract and Keywords

Some employment contracts have an expiration date while others do not, but this does not imply that the former lead to less employment security than the latter. First, they do not necessarily last for a shorter period of time; second, fixed-term contracts may be used as probationary periods at the end of which a worker may be retained under a standard arrangement; third, many short-lived contracts may succeed each other with negligible interruptions and therefore produce a continuous employment status. An empirical analysis that brings all these considerations together suggests however that in the mid term workers in Italy with fixed-term contracts spend more time in unemployment than workers with open-ended arrangements. As the likelihood of getting a standard job does not compensate for this situation, we argue that in Italy non-standard employment relationships bring to more employment insecurity for the worker. Comparative analysis follows, highlighting similarities in Germany, and particularly in Spain and Japan.

Keywords:   Employment security, Employment duration, Unemployment duration, Job-to-job transitions, Port-of-entry hypothesis

Introduction

The first dimension that we will focus on is that of employment security, operationalised through employment continuity, that is, continuity in the condition of being employed, also with different jobs and different employers. Other conditions being equal, non-standard workers – those with fixed-term contracts in particular – run a higher risk of precariousness, vis-à-vis standard workers, if the higher contract discontinuity they are subject to, due to contract expiration, translates into higher employment discontinuity.

Contract discontinuity and employment discontinuity are indeed not the same thing. First of all, the fact that a contract has an expiration date does not necessarily mean that the contract will actually last for a shorter period of time. Moreover, the duration of a contract does not coincide with the duration of an employment relationship: fixed-term contracts might be used as probationary periods, at the end of which the selected workers are kept on by the firm. Lastly, employment relationships with a shorter duration are also compatible with essentially uninterrupted careers if the transitions from one job to the next – the so-called job-to-job transitions – are sufficiently frequent and the periods of non-employment sufficiently short. Hence, working with fixed-term contracts does not necessarily mean having more discontinuous careers.

Verifying if this is actually the case is an empirical issue, which is precisely what we will do in this chapter. We will prove that although open-ended contracts are far from being ‘permanent’, fixed-term contracts are even shorter. This greater contract discontinuity is generally not offset by more frequent transitions to a new job or by sufficiently shorter non-employment periods.1 As a consequence, workers whose careers are exclusively or almost exclusively made up of fixed-term contracts are exposed to high employment discontinuity, which leads to low employment security. It is therefore understandable that most fixed-term workers wish to have open-ended contracts. We will nonetheless argue that, in Italy, non-standard contracts hardly act as ports of entry into standard employment. The last section of this chapter will place the empirical evidence from the Italian situation within the context provided by Germany, Japan and Spain.

(p.62) A descriptive analysis of work careers

Careers analysed in this chapter have been pieced together using data from the Work Histories Italian Panel (WHIP: see Appendix A). Since employment discontinuity could be a hallmark of the initial phase of an individual's work career, we decided to focus on two specific groups of workers within the sample: those who entered the labour market at the beginning of the observation period and were aged between 16 and 35 (‘entrants’ in the following), and those who at that time were aged between 36 and 50 (‘experienced’); both for entrants and experienced workers we then separated the analysis for full-time and part-time workers.

On each group of workers we performed three sets of descriptive analyses. For the sake of readability, while tables and figures contain separate elaborations for full- and part-time workers as well as for entrants and the experienced, the text, unless dedicated comments are necessary, will focus on full-time entrants only.2

Our first set of analyses focuses on the duration of work contracts and leads to two key results: a large portion of open-ended contracts expires within a short period of time; yet, the duration of fixed-term contracts is shorter.3

Once it has been established that open-ended contracts usually imply greater contract continuity, it is necessary to understand if this leads to higher employment continuity. In order to do so, we need to compare the frequency with which open-ended workers and workers with fixed-term contracts become non-employed when their contracts expire, which is what the second set of analyses focuses on. First of all, it emerges that the frequency of transitions towards non-employment at the end of a contract is generally high for all types of contracts in Italy. The figures referring to open-ended workers and workers with fixed-term contracts are comparable – and occasionally higher for the second group – if experienced workers are considered. Conversely, in the case of entrants, some forms of fixed-term work are associated with a lower frequency of transitions towards non-employment in comparison to what happens to workers with open-ended contracts. These results are confirmed when medium-term transitions are examined.

The third set of descriptive analyses deals with the duration of non-employment. From this point of view, there are very small differences between the duration of non-employment periods after an open-ended contract and after a fixed-term one, which nonetheless tend to favour the latter.

Overall, experienced workers employed with fixed-term contracts are subject to higher turnover, which is not offset by a higher likelihood of immediately finding a new job or by a lower duration of the job-hunting period when not employed. The situation of entrants is, instead, less clear-cut due to less frequent transitions to non-employment. In order to ascertain whether this compensates for the lower duration of employment spells, we thus compare workers with open-ended and fixed-term contracts in terms of the number of months actually worked over a medium period of time.

(p.63) Duration of contracts

Only a third of entrants started their work career with an open-ended contract, while part-time work involves shares ranging from 6% to 33% depending on the type of contract (see Table 4.1).4

Table 4.1: Entrants by type of work arrangement

Type of work arrangement

Share of entrants (%)

Of which part-time (%)

Open-ended

33.4

22.8

Training

11.1

16.0

Apprenticeship

25.8

5.6

Temp agency work

1.8

22.9

Seasonal

1.3

33.3

Direct-hire fixed-term

8.4

32.9

Wage and salary independent contractors

6.3

Self-employment

11.9

Total

100.0

Note: Reference period: workers who entered the labour market in 1998 and 1999.

Source: Own elaboration on WHIP data.

Overall, among entrants, almost 43% of the contracts expired within the first 12 months, while 63% expired within the first two years. More specifically, 32% of open-ended contracts, 41% of apprenticeships, 85% of temp agency contracts, 82% of direct-hire fixed-term contracts and 67% of independent worker contracts expired within the first 12 months.5 Within the first two years, 47% of open-ended contracts and a share ranging from 62% to 96% of fixed-term contracts expired (see Figure 4.1a). (p.64)

Flexibility and employment security: an analysis of work careers

Figure 4.1a: Duration of contracts, entrants (full-time contracts)

Notes: Vertical axis shows the cumulative frequency of work contracts having a shorter duration than that displayed along the horizontal axis; higher curves thus refer to contracts with shorter duration.

Source: Own elaboration on WHIP data.

Flexibility and employment security: an analysis of work careers

Figure 4.1b: Duration of contracts, entrants (part-time contracts)

Notes: Vertical axis shows the cumulative frequency of work contracts having a shorter duration than that displayed along the horizontal axis; higher curves thus refer to contracts with shorter duration.

Source: Own elaboration on WHIP data.

(p.65)
Flexibility and employment security: an analysis of work careers

Figure 4.2a: Duration of contracts, experienced workers (full-time contracts)

Notes: Vertical axis shows the cumulative frequency of work contracts having a shorter duration than that displayed along the horizontal axis; higher curves thus refer to spells with shorter duration.

Source: Own elaboration on WHIP data.

Flexibility and employment security: an analysis of work careers

Figure 4.2b: Duration of contracts, entrants (full-time contracts)

Notes: Vertical axis shows the cumulative frequency of work contracts having a shorter duration than that displayed along the horizontal axis; higher curves thus refer to contracts with shorter duration.

Source: Own elaboration on WHIP data.

Hence, two facts clearly emerge: first, contrary to what is generally believed of the Italian labour market, open-ended contracts too are subject to a high turnover; second, the duration of fixed-term contracts is nonetheless markedly shorter. So, it can be concluded that – irrespective of age (see Figures 4.2a and 4.2b) – workers hired through fixed-term contracts are affected by higher contract discontinuity. In order to determine if this also implies higher employment discontinuity, the next step is to understand what happens when a contract expires.

(p.66) Transitions between contracts

Working with contracts of shorter duration does not imply that an individual is more frequently non-employed, provided that, at the end of a contract, the transition to a new job occurs more regularly. So, although workers hired with fixed-term contracts are subject to higher contract discontinuity in comparison to standard workers, they would not be affected by higher employment discontinuity if the job-to-job transition rates were able to offset the shorter duration of contracts.

In this regard, among full-time entrants, transitions after an open-ended contract lead to non-employment in 48% of the cases, similarly to what happens to apprentices (46%), whereas transition rates for direct-hire temps and temp agency workers are slightly lower (around four in ten). Conversely, independent contractors become non-employed in two thirds of the cases. The situation for experienced workers is more clear-cut, as transition rates from fixed-term employment to non-employment are either comparable or higher than from open-ended jobs (see Table 4.2).

Table 4.2: Employment outcomes after termination of former work arrangement

Former work arrangement

Employment outcome (row percentages)

Entrants

Experienced workers

Non-employment

Standard

non-standard

self-employment

non-employment

standard

non-standard

self-employment

Retirement

Open-ended

Full time

47.7

30.3

20.3

1.7

42.4

42.3

7.3

1.9

6.1

Part time

53.6

12.4

32.4

1.6

46.9

11.7

35.9

2.0

3.5

Apprenticeship

Full time

45.8

27.2

25.4

1.6

Part time

43.5

12.9

41.1

2.5

Temp agency work

Full time

40.2

20.9

38.7

0.2

42.0

30.2

27.6

0.0

0.2

Part time

50.9

12.1

37.0

0.0

46.9a

8.2a

44.9a

0.0a

0.0a

Direct-hire fixed-term

Full time

41.4

29.5

28.7

0.4

40.9

43.0

15.0

0.6

0.5

Part time

40.4

15.9

43.7

0.0

42.2

9.1

47.3

0.5

0.9

Independent contractors

66.2

9.8

21.6

2.4

77.8

6.1

13.8

2.0

0.3

Note: (a) Part-time temp agency entrants and both full-and part-time temp agency experienced workers account for less than 50 observations in the data. Reference period: separations occurred in 1998 and 1999.

Source: Own elaboration on WHIP data.

(p.67) Thus, while for experienced workers hired through fixed-term contracts there is no evidence of a (higher) job-to-job transition rate capable of offsetting the (shorter) duration of contracts, the overall effect on entrants is not as clear. The analysis of an individual's employment status immediately after the end of a work contract presents some drawbacks however: it provides little information about the evolution of the individual's work career, since future career outcomes may well depend on past employment statuses and not only on the current one; it implicitly selects, within a time frame, only the contracts that expire or are interrupted, excluding those that, although active, do not undergo any variation; and it is not informative of the duration of non-employment periods brought about by the expiration of a contract, which might vary greatly depending on the type of contract in question.

In order to overcome the limitations mentioned under the first item above, we now analyse the employment status four years after the end of a contract (see Table 4.3). As for entrants, non-employment is the outcome in 8% of the cases. Workers who are non-employed after an open-ended (7.6%), an apprenticeship (8.1%), a direct-hire fixed-term (7.9%) or an independent worker contract (8.5%) are close to the average, while in the case of temp agency workers, the rate of non-employment is almost half as small (4.7%). Among entrants, thus, only temp

Table 4.3: Employment outcome four years after termination of a work arrangement

Initial work arrangement

Employment outcome (row percentages)

Entrants

Experienced workers

Non-employment

Standard

non-standard

self-employment

non-employment

standard

non-standard

self-employment

Retirement

Open-ended

Full time

7.6

58.1

26.5

7.8

10.3

66.0

9.8

8.4

5.5

Part time

13.7

39.7

38.7

7.9

9.8

27.6

47.5

9.3

5.8

Apprenticeship

Full time

8.1

39.2

45.5

7.2

Part time

7.3a

31.7a

53.7a

7.3a

Temp agency work

Full time

4.7

65.9

25.4

4.0

16.2a

46.0a

27.0a

10.8a

0.0a

Part time

0.0a

70.0a

25.0a

5.0a

0.0a

50.0a

50.0a

0.0a

0.0a

Direct-hire fixed-term

Full time

7.9

53.3

33.1

5.7

11.8

55.1

23.7

6.9

2.5

Part time

6.9

55.7

36.6

0.8

10.3

34.0

38.1

15.5

2.1

Independent contractors

8.5

45.2

36.2

10.1

18.3

15.7

51.7

12.8

1.5

Note: (a) Part-time temp agency entrants and both full-and part-time temp agency experienced workers account for less than 50 observations in the data. Reference period: separations occurred in 1998 and 1999.

Source: Own elaboration on WHIP data.

(p.68) agency work displays rates of transition to non-employment lower than those of open-ended jobs.6

Nonetheless, these results might be affected by the second drawback mentioned earlier: all the contracts that within a time frame do not undergo any variation are automatically excluded from the sample. Since this problem increases with the duration of contracts – the longer a contract, the less likely it is to observe its interruption within a given period of time – the elaborations illustrated so far exclude a proportionally higher share of open-ended contracts in comparison to the share of fixed-term contracts; in loose terms, the sample analysed here comprises only open-ended contracts with a limited duration, thus potentially biasing the analysis.

In order to address this drawback, we focus on the contracts that are active at a given point in time instead of those that end within a time frame (see Table 4.4).7 Four years after, less than 8% of entrants are non-employed. When looking at the type of contract, 11% of those who enter with an open-ended contract do not have a job; similar figures apply to those who enter as apprentices or as direct-hire temps. The share for independent contractors is twice as high (22%). Concerning

Table 4.4: Employment outcome four years after observation, by initial work arrangement

Initial work arrangement

Employment outcome (row percentages)

Entrants

Experienced workers

Non-employment

Standard

non-standard

self-employment

non-employment

standard

non-standard

self-employment

Retirement

Open-ended

Full time

11.1

66.9

15.9

6.1

4.7

87.6

2.5

1.9

3.3

Part time

16.5

16.5

62.3

4.7

8.4

7.1

79.6

2.5

2.4

Apprenticeship

Full time

10.6

40.4

41.8

7.2

Part time

25.0a

16.7a

58.3a

0.0a

Temp agency workb

Full time

Part time

Direct-hire fixed-term

Full time

10.4a

54.2a

31.2a

4.2a

8.5

57.4

23.9

4.7

5.5

Part time

9.1a

72.7a

18.2a

0.0a

10.8a

10.8a

67.6a

10.8a

0.0a

Independent contractors

21.5

35.4

30.4

12.7

15.8

9.2

67.4

6.8

0.8

Note: (a) Part-time entrant apprentices, all direct-hire fixed-term entrants and part-time direct-hire fixed-term experienced workers account for less than 50 observations in the data.

(b) As temp agency work was introduced a few months before the observation date, observations for this type of contract are extremely rare. Reference period: initial work arrangement observed in May 1998.

Source: Own elaboration on WHIP data.

(p.69) experienced workers, the share of transitions towards non-employment is around 5% for workers holding an open-ended job four years before and always higher in the other cases.

To sum up, in the case of experienced workers, besides ensuring higher contract continuity, open-ended contracts also present higher job-to-job transition rates. On the other hand, in the case of entrants, some (fixed-term) contract types are characterised by lower rates of transition towards non-employment. Before determining if this is enough to offset the shorter duration of contracts, it is necessary to investigate how long the periods of non-employment last.

Duration of non-employment

Figures 4.3 and 4.4 describe the duration of non-employment periods. The various curves refer to the different types of contracts leading up to non-employment periods.

As for full-time entrants, non-employment lasts less than six months for 52% of workers previously hired with an open-ended contract, with the corresponding shares ranging from 48% to 65% for workers with fixed-term contracts. Within one year, the job-finding rate rises to 81% for open-ended workers and to 83% to 90% for the others.8 With the only exception of apprenticeships in the first few months of job search, we can thus assert that open-ended contracts are characterised by slightly longer times needed to re-enter the labour market.

Flexibility and employment security: an analysis of work careers

Figure 4.3a: Duration of non-employment, entrants (full-time contracts)

Notes: Vertical axis shows the cumulative frequency of non employment spells having a shorter duration than that displayed along the horizontal axis; higher curves thus refer to spells with shorter duration

Source: Own elaboration on WHIP data.

(p.70)
Flexibility and employment security: an analysis of work careers

Figure 4.3b: Duration of non-employment, entrants (part-time contracts)

Notes: Vertical axis shows the cumulative frequency of non employment spells having a shorter duration than that displayed along the horizontal axis; higher curves thus refer to spells with shorter duration

Source: Own elaboration on WHIP data.

Flexibility and employment security: an analysis of work careers

Figure 4.4a: Duration of non-employment, experienced workers (full-time contracts)

Notes: Vertical axis shows the cumulative frequency of non employment spells having a shorter duration than that displayed along the horizontal axes; higher curves thus refer to spells with shorter duration

Source: Own elaboration on WHIP data.

(p.71)
Flexibility and employment security: an analysis of work careers

Figure 4.4b: Duration of non-employment, experienced workers (part-time contracts)

Notes: Vertical axis shows the cumulative frequency of non employment spells having a shorter duration than that displayed along the horizontal axis; higher curves thus refer to spells with shorter duration

Source: Own elaboration on WHIP data.

Employment security: results

The overall picture shows that fixed-term contracts have, indeed, a shorter duration in comparison to open-ended ones. However, besides displaying – at least for some contract types and above all in the case of entrants – higher rates of direct transition towards new jobs, workers hired through fixed-term contracts also experience, on average, slightly shorter periods of non-employment.

In order to determine what the overall effect is, that is, whether or not fixed-term workers experience, in the medium term, higher or lower employment discontinuity in comparison to workers with open-ended jobs, we now look at the share of career spent as non-employed by workers with different contracts (see Table 4.5). Among both entrants and experienced workers, fixed-term contracts are associated with longer portions of work careers spent in non-employment, vis-à-vis those working with open-ended contracts. So, even when fixed-term contracts are characterised by higher job-to-job transition rates and shorter non-employment periods in comparison to open-ended contracts, this is not enough to offset their shorter duration. An exception might be represented by entrants under a full-time temp agency contract, as the share of career they spend as non-employed is as small as for standard workers; however, when one looks at the work arrangement five years after entry instead of the entry contract as the reference employment status, the share of non-employment for temp agency workers is three times as large that for standard workers (results not shown in the table). (p.72)

Table 4.5: Share of career spent as non-employed in the medium term by first observed employment status

First observed employment status

Entrantsa

Experienced workers

Non-employment

28.2

Open-ended

Full time

9.4

3.6

Part time

10.7

5.5

Apprenticeship

Full time

11.2

Part time

10.2

Temp agency work

Full time

9.4

Part time

10.6b

Direct-hire fixed-term

Full time

13.1

9.1

Part time

11.8

8.2a

Independent contractors

15.1

15.3

Notes: (a) The first observed employment status for entrants is the entry contract.

(b) Less than 50 observations in the data. Reference period: from 1998 to 2003.

Source: Own elaboration on WHIP data

To sum up, we have seen so far that, as compared to standard workers, fixed-term contracts entail both shorter employment spells and shorter non-employment spells, for entrants as well as for experienced individuals. While for the latter this clearly results in less employment security as also short- and medium-run transitions among labour market statuses suggest that non-employment is more frequent for workers with fixed-term contracts, the picture is less clear-cut for some of the entrants. In particular, when they hold a full-time temp agency contract, transitions to non-employment, as compared to workers with standard contracts, are, indeed, slightly less frequent. All in all, in the medium run, this results in the number of months spent in the non-employment status by temp agency entrants being as small as for entrants with standard contracts; however, this holds true only if a temp agency contract is held at the very beginning of one's career.

We can thus argue that the chances of moving to an open-ended contract are extremely important in order to increase one's employment security, something we will deal with in the next section, which also takes into account the role played by workers' individual characteristics.

Port-of-entry and persistence effects

The topic of short- or fixed-term employment relationships as ports of entry towards more stable employment has been widely discussed in the literature. By limiting our focus here to some contributions regarding the Italian labour market, any work experience, even those with non-standard contracts and with a short duration, seems to have a positive effect on future employment outcomes (Ichino et al, 2008; Picchio, 2008). They do not entail a negative stigma, except in the case of older workers (Contini et al, 2000), provided that they are not repeated (p.73) over time or last too long, in which case they might even worsen the chances of stabilisation (Gagliarducci, 2005; Barbieri and Scherer, 2009). These matters are further investigated by Berton et al (2011), who study the transitions between different employment statuses by distinguishing the various specific contract types.

The analysis yields very clear results:

  • Individual characteristics – such as gender, level of education and preferences – account for a large portion of the phenomena described in the previous sections of this chapter, but not all of them.

  • Transitions to open-ended contracts are always more likely for the employed – regardless of their type of contract – than for the non-employed: fixed-term work thus enhances, on average, the probability to work with an open-ended contract in the future (port of entry).

  • An order among non-standard contracts with respect to the probability of taking an open-ended job nonetheless clearly emerges, with direct-hire temps and temp agency workers outperforming apprentices and independent contractors at the bottom; the best port of entry into open-ended employment, however, are open-ended contracts themselves.

  • Individual characteristics being equal, regardless of the contract currently regulating the employment relationship, the most likely event in the future is that a worker will again be hired with the same type of contract (persistence).

All in all, this means that the chances of fixed-term workers getting an open-ended job are lower than those of getting another fixed-term contract. As Berton et al (2011) also document, a substantial share of such persistence occurs within the firm, without any transformation into open-ended contracts even beyond any reasonable amount of time needed for screening. This behaviour may indicate that the strategy adopted by the firms aims at taking advantage of non-standard contracts mainly in order to reduce costs. This is to be appraised in the light of the evidence showing poor human capital accumulation in fixed-term jobs (European Commission, 2011), which is therefore likely to be both a cause and a consequence of persistence in a vicious circle that leads fixed-term contracts to have negative effects on employment security also in the long run (Barbieri and Scherer, 2009).

Career dynamics in comparative perspective

A main concern emerging from the analysis of employment continuity in Italy is the cost-reduction-oriented strategy of firms using fixed-term workers, with a leakage into poor human capital accumulation that is likely to negatively affect non-standard workers' career dynamics. In the following, we will see that Italy and Spain display similar patterns in this perspective. The comparison with Germany and Japan – whose labour markets are instead characterised by employers investing in long-term employment relationships – thus turns out to be of particular (p.74) relevance; as we will see, despite this similarity, Germany and Japan perform in a deeply different way, with the former allowing many fixed-term workers to climb the ladder of employment security and the latter moving instead towards a model of neat segmentation.

Germany

Fixed-term work in Germany mainly concerns individuals with university education, those who are low-qualified (McGinnity et al, 2005) and – in about 70% of cases – those who fail the transition from training to work (Hagen, 2003).

According to McGinnity et al (2005), they are all subject to substantial turnover, but only at the beginning of their career: individuals who enter the labour market with a fixed-term contract are indeed six times as likely to be unemployed one year after entry as those who enter with an open-ended one; the risk is then twice as high three years after entry and is the same five years later. No difference is found, however, with respect to unemployment duration. When fixed-term workers' career perspectives are instead compared to those of the unemployed, Hagen (2003) finds a higher probability of being employed during the following three years, a lower probability of being outside the labour force and no effect on the risk of unemployment. At least until the medium run, fixed-term workers thus suffer from a negative employment security gap in comparison to standard workers. Nonetheless, holding a fixed-term job is better than remaining unemployed in this perspective: even if in the short run this does not reduce the risk of unemployment, it enhances participation in the labour market, which in turn positively contributes to one's employment perspectives, thus acting as an indirect stabilisation device.

In this perspective, Mertens and McGinnity (2004) directly assess the existence of a port-of-entry effect. They find that about 40% of fixed-term workers make a transition to an open-ended position every year; this frequency increases with age and – somehow surprisingly – decreases for holders of tertiary education degrees.9 Hagen (2003) sheds further light on such a process: three years after a fixed-term job, 60% of German workers hold an open-ended contract, while the same figure for people observed in unemployment is 43.5%; the marginal impact of a fixed-term experience on the probability of finding an open-ended position in the following three years is positive and – contrary to what happens in Italy – larger than on the probability to persist in fixed-term positions. We thus argue that for workers who do not enter or fail the vocational training path to stable employment, fixed-term jobs, despite the initial high turnover, represent an alternative pathway to gain employment continuity in the medium run.

Temp agency workers represent an exception to this positive picture nonetheless. Germany stands among the largest markets in Europe for agency work (2% of the total employed workforce according to Ciett [2011]). The progressive liberalisation of this market has had a negative effect on temp agency workers' tenure in the last 15 years (Antoni and Jahn, 2009); this resulted in the strongly (p.75) pro-cyclical behaviour of temp agency work (Jahn and Bentzen, 2010), which is all the more serious as its port-of-entry effect into open-ended employment is poor in Germany (Kvasnicka, 2005).

Japan

Recruitment policies in the Japanese labour market are traditionally based on the search for high-school leavers and university graduates, every year in the spring. This was and still is part and parcel of the ideal and practice of lifetime employment. As seen in Chapter 3, lifetime employment has always represented a relatively minor share within the broader category of standard work. Even standard jobs outside lifetime employment, however, enjoy high levels of job and social protection. In their quest for numerical flexibility and cost reduction, in the past decades Japanese firms have progressively substituted non-standard for those standard workers outside lifetime employment, thus maintaining the core while eroding the belt.

Non-standard workers are subject to a high degree of turnover: both job creation and job destruction are larger than for standard employees (Genda, 1998), non-standard workers' probability to become unemployed is five times as high as for standard ones (Esteban-Pretel et al, 2011) and the share of non-standard employment is found to be highly pro-cyclical (Houseman and Osawa, 1995). On top of that, the chances of entering standard employment are extremely low for non-standard workers: in the short run, the probability of getting a standard job is lower for non-standard workers than for the unemployed (Hiroki, 2001) and it takes 20 years of participation in the labour market for the probability of holding a standard job to be the same for everybody irrespective of their starting position (Esteban-Pretel et al, 2011). This is indeed an extremely large portion of one's career (about a half) and gives rise to deep segmentation.

According to Kondo (2007), the reason behind this evidence is that Japanese firms' recruitment strategies for core positions basically have not changed in the last decades: most of them still look for potential hires only among high-school leavers or university graduates and not within the employed or experienced unemployed workforce. If external careers for non-standard workers are thus extremely difficult, internal ones turn out to be unlikely as well: indeed, only 25% of Japanese firms consider promotion from non-standard to standard positions as a viable recruitment strategy. Past work experience thus deeply determines one's career perspectives and its impact accumulates over time; Japan appears therefore to be one of the most segmented labour markets among developed economies (Keizer, 2007).

Spain

Spain is often referred to as a clear-cut example of labour market segmentation, in which reforms at the margin created a secondary segment of poorly protected (p.76) fixed-term workers beside a primary one where workers with open-ended contracts enjoyed a high(er) degree of job security (European Commission, 2011). This picture resembles the Italian one: both job and worker turnover are larger among fixed-term positions in Spain (Garcia-Serrano, 1998). A (partially) counterbalancing effect is observed with respect to unemployment duration. On the one hand, the job-finding rate for fixed-term work is, indeed, 10 times as high as for open-ended jobs, entailing a shorter waiting time for workers accepting a fixed-term contract (Bover and Gomez, 2004). On the other hand, however, workers moving back and forth from unemployment to fixed-term jobs are found to systematically anticipate the long-term unemployed in searching for a new job; for the latter, the liberalisation of fixed-term employment has come at the price of even longer unemployment (Güell, 2003).

The problem with fixed-term jobs in Spain is once again related to their port-of-entry effect, since only a small fraction of fixed-term positions is eventually converted into open-ended employment relationships and this happens either in the presence of strictly binding legal constraints or when fixed-term workers present a credible threat of quitting their current job (Güell and Petrongolo, 2007). On top of the negative impact on employment security, this cost-reduction-oriented behaviour of Spanish firms appears to be detrimental for human capital accumulation (Dolado et al, 2002; Bentolila et al, 2008) and, in turn, for labour productivity as well (Dolado and Stucchi, 2008).

Conclusions

When compared to open-ended contracts, fixed-term contracts have negative effects on a worker's employment continuity that go beyond their shorter duration in all the four countries taken into account. This occurs to different extents however. While in Germany, with the exception of temp agency work, fixed-term employment does not appear to be excessively detrimental to one's medium-run employment perspectives, poor human capital accumulation and cost-reduction-oriented recruitment strategies lead workers with fixed-term contracts to move back and forth from employment to unemployment for a relevant share of their work career in both Italy and Spain. In the former, though, temp agency work seems to emerge as an exception (in this case positive), but only when occurring early in a worker's career. Segmentation is even deeper in Japan, where the way workers enter the labour market almost completely determines their career outcomes; the fact that Japanese firms heavily rely on high-school and university graduates for the recruitment of the incoming workforce, and on school performance to screen it, makes it hard for a worker employed under fixed-term agreements to enter both internal and external career paths.

In our approach, though, a lower level of employment security might be offset by higher wages during periods of employment or by adequate levels of social protection when non-employed, thus potentially compensating for the gaps that have emerged so far. The next chapter will deal with the first of these issues: wages.

(p.77) Notes

(1) Non-employed individuals are those who simply do not have a job, whereas unemployed individuals are those who do not have a job but are looking for one. Despite the precautions we have adopted to narrow down our sample to include only the unemployed, we deem it safer to refer to the broader category of non-employment.

(2) Similarly, the sample selection criteria – which exclude workers who only occasionally participate in the labour market – strongly reduce the gap between men and women; therefore, from the gender perspective, this chapter will present only aggregate elaborations.

(3) Open-ended employment in Italy leads instead to lifelong work relationships in the public sector, which is not observable in the social security data used here and amounts to about three million workers. Two aspects are worth noting: first, due to binding budget constraints, open-ended employment in the Italian public sector displays a negative trend over several years now, thus being of minor relevance in the flows despite its enduring importance with respect to the stock of employed workers; and, second, in this chapter's perspective, including public open-ended positions in the analysis would lead to larger employment security divides between open-ended and fixed-term workers (see eg Barbieri, 2009).

(4) The amount of working hours cannot be observed in the case of self-employed workers and independent contractors. From now on we will exclude from the analysis in this chapter trainees, as these are now only possible in the public sector (after the 2003 Biagi law), and seasonal workers due to their peculiar cyclical employment patterns.

(5) We refer here to the duration of temp agency work contracts – that is, to the length of the relationship between the worker and the agency – and not to the duration of the assignments that temp agency workers perform in the user firm.

(6) The analysis performed on experienced workers yields partially different – and once again neater – results since the frequency of transitions towards non-employment is higher on average (11%) and always smaller for open-ended workers than for those with fixed-term contracts.

(7) In more technical terms, we will use the stock sampling method rather than the flow sampling method. It should be noted, however, that this strategy is not completely free from drawbacks, because it tends to underestimate the number of contracts with a short duration.

(8) A difference seems to emerge between full-time and part-time temp agency workers in relation to the time it takes them to re-enter the labour market, which is very short for the first group and very long for the second group. However, these figures might derive from the markedly limited number of observations concerning part-time temp agency workers; hence, their statistical reliability is questionable.

(9) Pfeifer (2008) proposes an explanation in terms of risk aversion, which is found to be lower for more educated workers. (p.78)

Notes:

(1) Non-employed individuals are those who simply do not have a job, whereas unemployed individuals are those who do not have a job but are looking for one. Despite the precautions we have adopted to narrow down our sample to include only the unemployed, we deem it safer to refer to the broader category of non-employment.

(2) Similarly, the sample selection criteria – which exclude workers who only occasionally participate in the labour market – strongly reduce the gap between men and women; therefore, from the gender perspective, this chapter will present only aggregate elaborations.

(3) Open-ended employment in Italy leads instead to lifelong work relationships in the public sector, which is not observable in the social security data used here and amounts to about three million workers. Two aspects are worth noting: first, due to binding budget constraints, open-ended employment in the Italian public sector displays a negative trend over several years now, thus being of minor relevance in the flows despite its enduring importance with respect to the stock of employed workers; and, second, in this chapter's perspective, including public open-ended positions in the analysis would lead to larger employment security divides between open-ended and fixed-term workers (see eg Barbieri, 2009).

(4) The amount of working hours cannot be observed in the case of self-employed workers and independent contractors. From now on we will exclude from the analysis in this chapter trainees, as these are now only possible in the public sector (after the 2003 Biagi law), and seasonal workers due to their peculiar cyclical employment patterns.

(5) We refer here to the duration of temp agency work contracts – that is, to the length of the relationship between the worker and the agency – and not to the duration of the assignments that temp agency workers perform in the user firm.

(6) The analysis performed on experienced workers yields partially different – and once again neater – results since the frequency of transitions towards non-employment is higher on average (11%) and always smaller for open-ended workers than for those with fixed-term contracts.

(7) In more technical terms, we will use the stock sampling method rather than the flow sampling method. It should be noted, however, that this strategy is not completely free from drawbacks, because it tends to underestimate the number of contracts with a short duration.

(8) A difference seems to emerge between full-time and part-time temp agency workers in relation to the time it takes them to re-enter the labour market, which is very short for the first group and very long for the second group. However, these figures might derive from the markedly limited number of observations concerning part-time temp agency workers; hence, their statistical reliability is questionable.

(9) Pfeifer (2008) proposes an explanation in terms of risk aversion, which is found to be lower for more educated workers. (p.78)