Keywords

7.1 Introduction

The recenteconomic crisis has had a very strong impact on the labour market globally, generating a rise in unemployment and its duration as well as affectinglabour stability among the working population and the quality of their jobs (ILO2014). These effects have not been the same or present in all European countries or in all groups. Although older workers are already established in the labour market, have more experience, and have jobs that are more durable, young people who enter the working world are more exposed to the negative effects of the recession. The crisis and the weak recovery have exacerbated the difficulties in accessing employment for young people seeking their first job. This generation is being forced to accept jobs that are very low quality, part-time or temporary or even in the informal sector as the only path to work integration. Accessing employment and transitioning to decent work have become a very difficult process for these young people, with stable jobs being very scarce and difficult to land. The global youth unemployment rate is estimated to reach a maximum of 12.6% in 2013, with 73 million unemployed youth (ILO2013). On the one hand, the recruitment of young people has significantly decreased, and in many cases, they were the first to be terminated (Choudhry et al.2012). It also noted that the transitions of young people have worsened since the beginning of the crisis, increasing the likelihood of moving from employment to unemployment. Additionally, the probability of access to stable employment once they have finished their training has decreased, and what is observed is that young people are forced to begin careers in part-time and temporary jobs, suffering episodes of frequent unemployment along their career path, entering and exiting the labour market (Verick2011).

In the case of Spain, theunemployment rate for the entire population reached a maximum of 27% in the first quarter of 2013 (Active Population Survey), well above the European Union (EU) average of approximately 11% (Eurostat). This rate rose in the same period to 43% in the case of those younger than 30 years of age, though with differences within this group, given that young people between 16 and 19 years of age had an unemployment rate of 73%, the group between 20 and 24 years of age 52% and those between 25 and 29 years of age 33%. Additionally, although recent data show a certain decline in the unemployment rate to 20%, the group of young people under 30 years of age still suffers rates above 35%. Undoubtedly, as in other countries, in Spain, young people are the group most affected by the crisis, and they have suffered a significant worsening of their situation due to the difficulty in accessing employment, the low quality of their jobs and the instability in their situation (Dolado et al.2013; Cebrián and Moreno2015,2016).

The manner in which young people access the labour market is important because it can affect their subsequent career paths. Among the few young people who are employed, the temporary rate exceeds 50%. The fact that access to employment is temporary can be viewed as positive or negative. On the one hand, temporary employment can be a bridge to more stable employment and a desirable situation if the alternative is unemployment, which is known as “stepping stones” (Booth et al.2002; Gash2008). On the other hand, other theories and authors suggest that access to temporary employment can become a trap in which the worker is stuck without prospects for improvement; thus, the initial moments of the career path stigmatize the worker in later stages, which is known as “scarring effects” (The Economic Journal2001; OECD2014). The empirical evidence suggests that the majority of workers who access the labour market through a temporary contract end up achieving the worst results in terms of wages and job security. Even when the temporary work is voluntary, low rates of transition to a permanent contract generate inequalities that tend to persist (OECD2014). A study by the European Foundation for the improvement of living and working conditions (Eurofound) for2014 with data from several countries of the European Union concludes that, in Europe, workers with temporary employment have a greater rate of transitions to unemployment than older workers with an open-ended job. In addition, the crisis has increased the proportion oftemporary workers who become unemployed. Specifically, the ratio has increased from 11.6 between 2006 and 2007 to 18.5% between 2009 and 2010. Thus, the crisis has affected young temporary workers more due to the increase in the probability of losing their job than the reduced opportunities of finding a stable job. This study also highlights the high negative correlation between the rates of temporary work in a country and the proportion of transitions from temporary to open-ended employment, particularly in the Mediterranean countries such as Spain, with high rates of temporary work.

Some studies that focus on the Spanish case suggest that many young people are trapped in temporary work and that only some manage to take an open-ended job after some period in temporary jobs (Guell and Petrongolo2007; Toharia and Cebrián2007; Cebrián and Toharia2008; Garcia Perez and Muñoz Bullon2011; Garcia Perez et al.2014; Garcia-Perez and Vall Castelló2015).

According to the information from the Muestra Continua de Vidas Laborales (MCVL), young people who have had some employment between 2005 and 2013 had on average 7.4 episodes of employment, of which 5.9 were temporary. Similarly, according to the records of contracts in the Public Employment Service (Servicio Público de Empleo (SEPE)), during 2014 and 2015, approximately one-third of the contracts were also registered for workers under 30 years of age, of which only 7% were open-ended. This situation has been constant for theSpanish job market in recent decades, even though stability has been a priority objective of the various Spanish governments and successive reforms enacted.

Along these lines is the implementation by the Ministry of Employment and Social Security of a set of measures contained in the “Strategy for entrepreneurship and youth employment 2013–2016”. Among these measures the implementation of theNational Youth Guarantee System (Hernandez Ten and Gentile2015) as a global and integrated system that allows young people to receive goodoffers of employment should be highlighted, continuing education, learning practices or apprenticeship training within 4 months from the time of ending their studies or becoming unemployed; it also rewards the open-ended employment of the young beneficiaries of this system. The number of contracts subsidized for this reason has been very small since its implementation and the effects and reach of the system have not been as wide as one might expect in a country with a very high youth unemployment rate. Other measures to boost recruitment and reward employment under 25 years of age have also been implemented (RDL 8/2014), such as the “first youth employment” contract (Law 11/2013), which is a modality of currently possible contracts whose cause is the absence of experience of the worker contracted (a subjective cause) and which encourages its transformation to open-ended, or open-ended contracts, known as “flat rate” contracts (RDL 3/2014, RDL 1/2015), which are rewarded with a flat rate or minimum extent (since 2015) of business contributions for common contingencies toSocial Security . Special mention should be given to some specific programmes that promote youth employment, such as open-ended contracts that support entrepreneurs, the training linkage contracts or contracts made under the youth guarantee system. However, in quantitative terms, it does not appear that these incentives have had much impact on youth recruitment (Cebrián and Moreno2016).

The objective of this study is to analyse how youth under 30 years of age in Spain access open-ended employment and how long they keep this type of employment. This is to detect whether entry into open-ended employment allows young Spaniards to start a path to work stability, given the importance that successive governments in Spain over the last decade have been placing on this issues.

The information that will be used is provided by theMCVL , in its panel version for 2005–2015. The MCVL is used, first, to study the probability that someone younger than 30 years of age will start an open-ended job and, second, to analyse the duration of this type of job, identifying the role played by motives that can explain why an open-ended working relationship ends and leads the individual to be placed outside or inside employment, with or without a trace of stability.

Previous studies (Cebrián et al.2011; Cebrián and Moreno2015,2016) have investigated the overallduration of episodes of open-ended employment, noting that approximately 60% of the episodes of open-ended employment last no longer than 2 years. Therefore, owing to the available data, which collected information on all episodes, the observation period for the analysis of access to open-ended employment through 31 December 2013 is established. This period ensures a 2-year observation window for all jobs begun in the period. The purpose of this window is to have a homogenousobservation period to track each and every episode of open-ended employment during the same time period. With this information, we find that out of the young people accessing their first job, 2 years later, only 8% maintain the same employment, but approximately 60% of those who began in temporary employment remain in temporary work, and less than 50% who began with an open-ended contract remain in a stable path. When there has been a change in the employment situation in more than 80% of the temporary jobs, the relationship ended for causes unrelated to the worker, and this percentage decreases to below 60% if the job was open-ended; voluntary job losses account for 13% and 27%, respectively.

7.2 Youth Access to Employment

Between 2005 and 2015, in the SEPE, there were approximately 73 million contracts involving individuals younger than 30 years of age. This number of contracts represents approximately 10 million distinct individuals. During this period, contracts for young people accounted for 40% of the total registered contracts, though the proportion decreased from 50% in 2005 to 34% in 2015. A clear impact of the crisis was the expulsion of many young people from the labour market. If, during the precrisis years, young people could find a job even with little training, since 2008, the landscape has changed, and many of these many young who left the education system to work and who faced unemployment became inactive (Rocha2012). The participation rate for those under 30 years of age experienced a significant decrease over the period, 10 percentage points, linked to the arrival of the economic crisis. In 2005, the participation rate for this group was 67% and fell to 57% in 2015. The same is true for the occupancy rate, which decreased from 58% to 35% in the period (according to the data from the Spanish Labour Force Survey). This drop in the employment rate is in line with the evolution of the recorded youth contracts, and it should be noted that throughout the period, the unemployment rate grew, despite the declining active population. The reason for this result is that, although some young people left the labour market or delayed their entry, those who remained or accessed the market are largely unemployed.

In Spain, access to employment mainly occurs through temporary contracts, and this is true for all people, not only for young people. According to the recorded contracts during each of the years of the 2005–2015 period, open-ended contracts did not exceed 10% of all contracts registered each year for those younger than 30 years of age, with this figure decreasing to 6% in 2015. The latest Spanish labour market reforms have been aimed at promoting employment through various contractual modalities, as has already noted above.

With the goal of knowing which features lead to young people finding employment through an open-ended or temporary contract, a logit model can be estimated in which the dependent variable is having an open-ended or temporary contract. This estimate is performed by Cebrián and Moreno (2016) with the data for contracts recorded by the SEPE for the 2007–2014 period, controlling the probability ofaccess to open-ended employment by a number of variables concerning the characteristics of the individual and the job that somehow can affect the probability. The estimate of this probability has been repeated, expanding the sample to all contracts recorded for young people between 2005 and 2015. The following explanatory variables are used: sex, age, level of education, the autonomous community of the workplace, the branch of activity, occupation, the size of the company, the type of shift, whether the contract has some type of subsidy, the variation in employment in the quarter and the quarter in which employment begins. The results of the estimates show a decrease in open-ended contracts. Only between the fourth quarter of 2006 and the first quarter of 2009 is there a greater probability of recording an open-ended versus a temporary contract. Expanding the period of analysis allows one to observe in the results obtained the possible favourable impact of the reforms of 2006, which lasted until the economic stagnation of 2009. There is also a seasonal trend of more stable employment, given that the relative probability that the contract is open-ended is greater in the first quarter of each year, though throughout the period observed, the probability tends to decrease. The relative probabilities of the variables included in the analysis confirm what the mean values of the distributions show. Thus, men are more likely to have an open-ended contract than women and older workers. The level of education also increases the likelihood of having an open-ended contract, and the relative probability is significantly greater in the case of university studies. For branches of activity, the extractive industry and manufacturing, trade and financial and insurance activities have a greater probability of an open-ended contract. The group of occupations with the greatest probability of accessing open-ended contracts is administrative as well as occupations associated with higher levels of training, such as management and technical and professional support. Those that are least likely to access open-ended contracts are in the occupations of catering services, those dependent on trade or unskilled workers. With regard to the size of the establishment, companies with less than four workers have a greater probability of open-ended contracts; this probability decreases as the size of the establishment increases. Madrid and Catalonia are the two autonomous communities with the greatest probabilities of open-ended contracts, and Andalusia has the lowest relative probability. The rest of the explanatory variables show that the probability of open-ended contracts increases in full-time contracts, in those who have some type of subsidy, and in the quarters when the variation in employment is greater, reflecting the favourable effect of the economic cycle.

To be able to study the duration of these jobs and their stability, the probability of access to an open-ended contract is estimated using data from the MCVL. The added value of this analysis is the ability to identify those who access Social Security for the first time and to compare the probabilities obtained with all recorded episodes for young people.

If all episodes of employment of those younger than 30 years of age between 2005 and 2013 are selected and, alternately, only those episodes constitute the first working relationship recorded by Social Security, then the probability is that these are open-ended contracts, given the variables included and the characteristics of the individual referenced; both cases are shown in Fig.7.1. In the chart, the y-axis is the value of the relative probability that an open-ended contract versus a temporary contract has been recorded. On the x-axis, the abscissas are different quarters. Therefore, if, for a given quarter, the relative probability that an open-ended contract has been recorded is greater than the probability that a temporary contract has been recorded, then the value on the y-axis will be greater than 1. If the probability that a temporary contract has been recorded is greater than the probability of an open-ended contract, then the value on the y-axis will be less than 1. Finally, if the relative probabilities of open-ended versus temporary contracts, in a given quarter, are not significantly different, then the value will be equal to 1 (marked by the line that appears in black on the graph). The variables that are included as explanatory are sex, age, nationality, autonomous community, tax group, size of the establishment, time elapsed since the first relationship recorded by Social Security, branch of activity, type of business, type of shift, starting quarter and variation in employment in the starting quarter.

Fig. 7.1
figure 1

Young people in Social Security records between 2005 and 2013: probability of having a permanent job, (Source: MCVL, 2005–2015)

Full size image

The results obtained are very similar to those noted above from the estimate using the data from recorded contracts. The probability of accessing an open-ended contract is greater in the years before the economic crisis, perhaps reflecting the effect of the 2006 reform; then, the probability tends to decrease and becomes negative. Similarly, there are seasonality patterns in open-ended hiring. Men are more likely to have an open-ended contract than women, Spaniards are more likely to have an open-ended contract than foreigners and older workers are more likely to have an open-ended contract. The contributing group variable is an approximation of the level of training of the individual, and the coefficients of this variable show a higher probability of having an open-ended contract if the group is not peons. Andalusia is the autonomous community with the lowest probability of open-ended contracts. Considering all contracts, the branches of activity with the highest relative probability are the extractive industry and the financial sector; however, when the probability of first access to employment will be open-ended is estimated, the branches of activity with the highest probability are construction, trade, hotels, transport and public administration. The probability is higher in smaller companies.

7.3 Study of the Probability of Ending an Open-Ended Contract by Cause That Motivates Termination

As noted above, previous research by Cebrián and Moreno (2015,2016) has studied the duration of open-ended employment when the worker is younger than 30 years of age. In such jobs, the effects of the characteristics of the worker and the job on theduration of employment are investigated. However, it is very easy to determine what is the cause that explains why employment ends and how the explanatory variables determine the probability that the termination of the working relationship occurs by the choice of the worker or by the decision of the company, either by contract termination or dismissal or by any other reason, such as leave or disability. Therefore, we next determine whether there are significant personal and labour differences that explain the termination of open-ended employment and the reason that prompt it during a 2-year observation period. The analysis is performed using the information provided by the MCVL, according to which there are various reasons for the termination of a contract. Using this information, it is possible to estimate competing risk models to assess why an episode that started between 1 January 2005 and 31 December 2013 remains in force 2 years after beginning or why it ended.

7.3.1 Specifying the Model According to the Causes of Terminating an Open-Ended Contract

The causes of termination have been grouped into four possible alternatives that identify the competing events. On the one hand, we consider the possibility that the employment remains active and, therefore, high in the Social Security system at the end of 2 years, in which case, the episode is right censored. On the other hand, in those cases in which there has been a decline in the Social Security system, the reasons recorded by Social Security include the possibility of a voluntary decision by the worker or a compulsory drop, caused by either termination (individual or collective) or the end of the contract, in addition to other cases that cannot be classified as voluntary or as mandatory (leave, demotion by Social Security, disability, etc.). In all of these cases, it is assumed that there is independence between the unobservable factors that affect each risk, making it possible to estimate competing risk models.

The events in competition are defined as follows. The reason for the situation of right censoring is that at the final moment of observation, i.e. 730 days after the start of the episode with an open-ended contract, it remains active and, therefore, is registered as such in the Security Social system. If it is not, it is considered that the episode has ended due to some cause. Because there can be three different reasons that explain the termination, models are used with multiple exits.

The duration of the episode is calculated as the difference between thetime of exit , identified by the date when the episode is lowered in the Social Security system, and the time when the contract started, i.e. the date it was high in the Social Security system over a 2-year observation window that begins the same day as the episode.

The explanatory variables related to the personal characteristics of the individual are the sex and age groups of individuals under 25 and over 24 years of age. The characteristics of the job include the autonomous community where the company is registered, the branch of activity, the size of the company, the tax group (as an approach to the level of occupation) and the type of open-ended contract, differentiating between a regular contract and an initially subsidized contract. Furthermore, a variable that indicates whether the individual had previous jobs registered with the system and the number of episodes that a worker had throughout the observation period has been included. Additionally included was the quarter of entry, to control for possible seasonal and cyclical factors and possible policy changes.

7.3.2 Results of the Estimated Model

The results of the estimated models show in the analysis of the duration of the episodes of open-ended employment what occurs in the final moment of observation, after 730 days since starting the contract.

This type of analysis provides evidence of how a young worker in open-ended employment progresses over time, identifying whether the loss of employment occurs with some frequency and what it depends on. It is important to determine whether open-ended contracts have a limited end in a relatively short timeframe, regardless of the cause that explains termination. This involves determining whether it makes sense to continue betting on these types of contracts without changing the system in which they are embedded and that may effectively be behind their instability.

These estimates include an estimate of the joint probability of exit, without differentiating the cause of termination, to compare the results with those obtained from the analysis of the duration of open-ended employment, considering the reason for the exit.

A risk rate greater than one means that it is more likely that the feature analysed has a positive effect on the exit rate, thereby reducing stability; conversely, a rate of risk of less than one indicates that the risk of ending the contract is lower, which favours stability.

The reference characteristics are being male, 16–24 years of age, working in a company based in a geographical area with an income level below the national average, having a regular contract, being an unskilled worker, working in the trade sector, working in a company with 1 to 49 employees when completing the MCVL and the beginning of the contract being in the first quarter of 2005.

First, it should be noted that there are no significant differences between men and women younger than 30 years of age, except regarding voluntary termination, which is 20% less likely among women. However, terminations for other reasons (e.g. leave) are 50% more likely among women.

However, a worker’s age group is important because younger workers are more likely to exit for all reasons except for others. But there is a quantitative significant difference between voluntary (30% more) and compulsory (6% more). Similar results are obtained if the company is based in an autonomous community with an income level above the national average. In fact, this result confirms other analysis outcomes: those regions which create more employment on the contrary have lower employment duration as Madrid or Cataluña. Then, the volatility of open-ended contracts, despite everything, is lower among older young people and regions with lower income levels.

It is equally remarkable that the contracts receiving some type of subsidy are overall 11% more unstable, but if the worker quits is almost 20%. If the contract is part-time, the probability of exit increases to 45% for all the episodes, though it is twice more likely to quit than to be fired. The residual cause has an opposite effect being more probable to end if the contract is a full-time ordinary.

Studying the results obtained for the branch of activity or economic sector in which the company develops its production reveals some interesting findings. On the one hand, it is notable that construction or services to businesses and households show a level of greater stability than trade, regardless of whether termination is voluntary or compulsory, and instability increases if there is any other reason for termination.

However, industry is significantly different from trade only if the motive is neither voluntary nor compulsory. In the transport sector , there is always more instability, and the same is true in the hotel sector, though in this case, compulsory causes are not significantly different from trade. In the financial sector, voluntary causes increase instability; however, a compulsory exit is less likely, which is similar to the collective services sector. Agriculture is generally more unstable, particularly for compulsory reasons.

In summary, following the details in Fig.7.2, we can say that dismissals are more likely in agriculture (19%) and transport (10.4%) and less likely in constructionFootnote 1 (30%) and financial, collective and other services (15%).Voluntary exits are more likely in hostel (195%) and in transport (90%), perhaps as a solution to escape from the precarious employment conditions.

Fig. 7.2
figure 2

Hazard rates: different reasons for leaving a contract by sector of activity (Source: MCVL, 2005–2015)

Full size image

The contribution group for Social Security, which relates the qualification necessary to perform the task of the job, is a very significant variable that shows a surprising result upon analysing exits because it is more likely for the contract to terminate as the occupational level increases. Thus, it is surprising to find that for the group composed of those at the upper end of the scale, the probability of a voluntary departure decreases.

With regard to the size of the centre, there is, as expected, a very positive relationship in the case in which the company had closed at the time of the data extraction. Although the data do not represent the company size when the contract was signed, employees have a greater probability of exiting both voluntarily, before the approaching fact, and compulsorily. In addition, in larger companies, exits are more likely, particularly for voluntary or other causes, and compulsory exits are less likely (15%). These results add evidence in favour of what Cebrián and Moreno (2016) pointed out about the lower (higher) duration of contracts in small (large) firms.

The starting quarter is not very significant in general, except towards the end of the recessionary cycle. In Fig.7.3 we can observe that since 2007 this is an average of the decline or the increase of the probability of voluntary and compulsory exits in all quarters, except during 2013 when these trends change.

Fig. 7.3
figure 3

Hazard rates: different reasons for leaving a contract by starting quarter of the contract (Source: MCVL, 2005–2015)

Full size image

7.4 Conclusions

Thepromotion of youth employment is a current goal in the programmes of all governments. Therefore, over recent decades, measures have been enacted, ranging from long-term projects, such as reforms to the education system to improve the employability of young people, to measures intended to have an immediate impact on employment, such as incentives for open-ended hiring. The paradigms of stability are employment and, in particular, employment with an open-ended contract.

The goal of this study is twofold: on the one hand, to describe the paths to the employment for young people younger than 30 years of age, emphasizing those that facilitate entry into more stable employment, and, on the other hand, to analyse theduration of this employment and the reasons for its stability, identifying the reasons that may lead to termination. This analysis is based on 2005–2015 information from the MCVL.

The analyses performed lead to some important conclusions. First, it should be noted that despite the budgetary efforts to subsidize and encourage open-ended recruitment equalling the costs of the different types of hiring, the weight of open-ended contracts in the set of youth recruitment is very small: the episodes of employment collected in the MCVL for the period analysed do not even reach 10% of the total affiliations. However, the analysis of probability that young people will enter open-ended employment shows an increase during the years before the Great Recession of 2009, particularly in the case of the first access to employment. This finding may indicate that incentives with regard to first-time access work to some degree, but subsequent career paths do not guarantee stability. It could also be indicating the effectiveness of the measures introduced in promoting stability and limiting temporary work by the reform of 2006 to combat the phenomenon known as the “culture of the temporary” that prevails among Spanish entrepreneurs (Toharia2005). However, the arrival of the Great Recession itself combated temporary work, curbing hiring in all forms but particularly open-ended.

One possible explanation of why companies are not attracted to the modalities of incentivized hiring can be that they all carry the restriction that the company must maintain the net size of the template for at least 3 years or the duration of the subsidy. This rigidity can pose a greater cost for the company than having to bear the labour costs of entry and exit related to unsubsidized open-ended contracts.

Second, when studying the stability of open-ended jobs by analysing the duration and the causes of termination, it is clear that young people suffer a high degree of insecurity, given that 2 years after starting a contract, only approximately 40% remain active, a result obtained in prior analyses. In addition, a high percentage of young people, regardless of the type of contract, terminate their labour relationship due to a decision by the company. Estimates that differentiate the effect of the cause of exit from a job on its duration show that young women have a lower probability of abandoning employment voluntarily, though they are more likely to exit due to other causes, such as leave. Age has a positive effect on stability, as well as regular contracts, lower-income autonomous communities, large firms and higher occupational level. Trade, hostel and transport are sectors with a high proportion of young workers. The probability of ending a job because of dismissal is alike in three of them, although in hostel and transport sectors the probability of leaving the job is higher than in trade. This could be an evidence about young workers escaping from precariousness.

In summary, if open-ended hiring and its promotion are not being successful and temporary contracts remain high, how much longerlabour market policies should continue to focus on the type of contract. Additionally, it would be desirable to divert part of the efforts and resources towards policies that encourage the creation of larger companies in sectors with more stable hiring, such as the industrial sector. Similarly, active policies should adapt the training of younger workers to the needs of the productive system, which requires improving channels of communication and information between the different actors in the labour market. Moreover, it would be interesting to know what is needed by Spanish companies and under which conditions they are willing to hire young people, given that they are not increasing stable employment under the incentives introduced by legislative changes.