Abstract
Introduction
People with disabilities usually have lower labour market participation and employment rates. In addition, ordinary employment is not a feasible option for some persons with disabilities. Sheltered employment is among the measures taken to promote employment opportunities for individuals with disabilities. The International Labour Organization (ILO) has recommended enacting this policy of sheltered employment for “disabled persons for whom access to open employment is not practicable” 1 . Sheltered employment refers to employment in firms created specifically for the employment of people with disabilities. These firms usually receive strong support and funding from the state. Differences in the design of such programs across countries (and even within countries) can be significant. Sheltered employment could segregate workers with disabilities and offer low wages and precarious employment relationships. Other forms of sheltered employment are similar to productive work in ordinary firms and in several countries during recent decades, this kind of employment has lost its predominant role as ‘protected’ employment for people with disabilities, while other measures (such as supported employment) have been implemented. The nations with the highest percentages of people working in sheltered employment (over 50 percent) are Belgium, Italy and Spain [14].
In Spain, the Act for Social Integration for People with Disabilities (Ley de Integración Social del Minusválido, hereafter LISMI) was enacted in 1982. This initiative defined the Sheltered Employment Centres (SEC) system as a measure to promote employment of people with disabilities. The centres operate as a sort of ‘protected’ labour market for people with disabilities; the purpose of the centres is to provide employment opportunities for people with disabilities who have difficulty entering into the ‘ordinary’ labour market.
SECs have been a key component of all strategies developed since the 1990s to promote the employment of people with disabilities in Spain. A long-lasting debate continues in Spain about whether SECs promote the integration of people with disabilities into social and economic life. Some organizations advocating for people with disabilities argue that the main role of these centres is to provide job or long-term career opportunities for people with disabilities without considering whether the workers should transition to the non-protected labour market. In contrast, other experts argue that SECs should be an intermediate step towards a job in an ‘ordinary’ firm. From this intermediate step perspective, workers hired by a SEC should eventually move on to a job in a non-protected firm. The debate remains open because there are very few data to clearly support either of these views. In an analysis of Sheltered Employment Centres in Madrid, it is found that satisfaction of workers is very high [9]. The authors suggest that personal and social adaptation services should be used as a means to promote the transition to normalized employment.
Given this perspective of the SECs as a ‘protected’ labour market for workers with disabilities, the aim of this paper is to analyse the stability of employment in SECs during 2005–2011 as compared to employment in ordinary firms. The regulation of SECs in Spain and the role they play in the lives of workers with disabilities is explained. The data used in the analysis are described and the results are presented.
The role of Sheltered Employment Centres in the promotion of employment of people with disabilities
Although SECs are widely-used and meant to increase the employability of people with disabilities, few analyses have provided empirical evidence on this issue. At the international level, some authors have reviewed the similarities and differences of sheltered employment policies across countries[3, 15]; however, few provide accurate and detailed data about program uptake and impact on the employment of people with disabilities. Analysis of policies towards people with disabilities in the member countries of EU-15, Australia, Canada and the United States, suggest that traditionally sheltered employment legislation was made independent of other employment policies, although this has changed in recent years [15]. Countries such as France, Spain and Portugal have regulations that place more importance on this sheltered employment, while other countries such as the Netherlands and the United Kingdom have moved towards greater promotion of supported employment 2 . In the Netherlands, innovative practices in sheltered work have been developed and sheltered work has changed to increase the efforts into transition to the regular labour market [16]. It is not a question of being in favour or against sheltered work but rather to determine which practices promote help for people with disabilities to full inclusion in the labour market.
There are common trends in the evolution of SECs in European countries [4]. The staffs of these centres are likely to be young, with a greater presence of workers with physical and sensorial disabilities. There is not a large difference in terms of those who are psychologically impaired. More than half of workers perform unskilled tasks but skilled professions are becoming an important part of these programs. Very specialised jobs are largely unrepresented. Supervisory positions are often filled by workers who are not disabled. Less than 3 percent of workers passing through sheltered employment get a job in ordinary firms, in most countries, the social support and training provided are designed to increase the skills and personal safety of workers in SECs and not to prepare these workers for jobs in ordinary firms.
Focusing on Spain, a qualitative study of SECs concludes that these centres provide training, employment and development of working habits (among other things) for people with disabilities, but they rarely help workers with disabilities move towards the ordinary labour market [1]. Other studies, using the Spanish Survey on Disabilities, Impairments and Health Status (Encuesta de Discapacidades, Deficiencias y Estado de Salud, EDDES-1999) launched in 1999, find that among programs promoting employment of people with disabilities, employment in SECs was the most-commonly used option among men and the second most-commonly used option among women [17]. During the period 2006–2011 in Spain, the number of contracts for people with disabilities in a SEC doubled and represented more than 50 percent of all contracts signed by people with disabilities, although the increase was in temporary contracts [13].
Using data from the Spanish Survey on Disabilities to evaluate the impact of different employment promotion measures for people with disabilities (among them sheltered employment centres) on the quality of the job match, results indicate there is not any effect (either positive or negative) of employment programs on the quality of job [10]. In general, although empirical evidence is scarce, there is a consensus on the limited impact of sheltered employment as a method of integration into the regular labour market
The regulation of sheltered employment centres in Spain
As noted, the Spanish Act for Social Integration for People with Disabilities (Ley de Integración Social del Minusválido, LISMI) defined the Sheltered Employment Centres as a measure to promote employment of people with disabilities. Labour market integration was a prominent objective in promoting the social and economic integration of people with disabilities. The framework for this objective was threefold: financial subsidies and fiscal incentives to promote the hiring of people with disabilities by firms; sheltered employment centres where, initially, only people with disabilities could be hired; and occupational centres, where people with disabilities engage in occupational therapy and personal and social adjustment in lieu of productive work 4 .
The LISMI defined sheltered employment centres as organizations promoted by public or private agencies that have a primary goal of performing productive work, are regularly involved in market operations and pursue the objective of ensuring profitable employment and providing personal and social adjustment services for their workers with disabilities. This set of organisations is the most common means of integration into the regular labour market for people with disabilities (see Royal Decree 2273/1985). They were defined as a class of ‘special’ or ‘protected’ firms due to their objective of integrating people with disabilities into the labour market.
LISMI presented a step-by-step design for labour market integration [11]. Occupational centres help to deal with the limitations created by impairments and disabilities. People with disabilities with a basic level of employability (sometimes as a result of the occupational centres) that would face difficulty integrating into ‘ordinary’ firms would be hired by SECs. People would accumulate human capital and basic knowledge about some occupations via learning-by-doing at SECs. This process would lead to these workers entering the ‘non-protected’ labour market where, thanks to the financial subsidies and fiscal incentives to hire workers with disabilities, they would become truly independent from public assistance and face the same conditions as the rest of the workers.
However, SECs were established by the LISMI as means of employing people with disabilities unable to obtain and hold a job under normal conditions (i.e., being hired by an ‘ordinary’ firm). From this perspective, the SECs would be a ‘protected’ labour market for people with disabilities with a low level of employability that they would not have been hired by an ‘ordinary’ firm [5]. Thus, the SECs would usually be both the beginning and the end of the labour market trajectory for people with severe disabilities.
Initially, the LISMI stated that the entire workforce of a SEC should be formed by workers with disabilities 5 . However, since the passing of the 66/1997 Act, an SEC must only be composed of at least 70 percent workers with disabilities. SECs must hire workers using the contractual arrangements provided by the Spanish Workers Charter. In its original wording, the Royal Decree of 1985 assumed the employment contract to be open-ended, although it provided for the possibility of a fixed duration and allowed for a maximum trial period of 6 months. After the reform of this Royal Decree in 1999, the employment contract may be open-ended or fixed-term and have a trial period set by collective agreement (not exceeding six months).
Employment in sheltered employment centres in Spain
Estimating the number of SECs in Spain and the number of workers with disabilities employed by this type of firm is not a simple matter. Data are scarce, although several authors have stressed that there has been an increase in the number of SECs in recent years. From 1996 to 2006, this number has risen from 562 to more than 1,500. The number of employees in SECs in 2006 was over 47,000, which is a 100 per cent increase since 1999. According to the Federal Association of SEC in Spain (Federación Empresarial Española de Asociaciones de Centros Especiales de Empleo 6 , FEACEM), more than 59,000 persons with disabilities worked in 1,871 SECs during 2010.
One of the most important aspects in the analysis of employment in SECs is the type of disability of hired workers. The general perception is that SEC workers often have more severe disabilities or disabilities that are mainly psychological, which hinder their integration into the regular labour market. In 2001, 50.5 percent of people with disabilities working in SECs had an intellectual disability, 39 per cent had a physical disability, and 10.5 per cent had a sensorial disability 7 .
There are two specific databases that can be used for the study of workers with disabilities and their relationship with the labour market to quantify the magnitude of employment in SECs: EDDES-1999 and EDAD-2008 (Survey on Disability, Personal Autonomy and Dependency Situation). According to the EDDES, in 1999, 22.5 per cent of people with disabilities between 16 and 64 were working. Within this group, 3.6 per cent were working in SECs (0.8 per cent of the disabled population). In 2008, the EDAD provides the following analogous figures: 27.3 per cent of the working-age disabled were employed, and 7.1 per cent of the workers with disabilities were employed in SECs. Clearly, there has been a notable increase in the proportion of people with disabilities working in SECs. Another way to approximate the quantitative importance of SECs as a way of integrating people with disabilities is based on procurement data 8 .
Data and descriptive analysis
In this article, a database provided by the Spanish Ministry of Employment and Social Security is used: the Muestra Continua de Vidas Laborales (MCVL) or Longitudinal Sample of Working Lives includes information about the labour market trajectory of Spanish workers 9 It includes any labour-based relationship with the Social Security System (a job, a pension or a contributory unemployment benefit), and the entire labour market trajectory since their first employment. It is a representative sample of the working population affiliated with Social Security each year, and it provides the whole working life of these workers. Data are available from 2005 to 2011, so a panel of every disabled worker’s employment status for each year in that period can be constructed.
Using an administrative dataset from Social Security records, the information about the disability status of workers comes from the firms’ reports once the worker has been hired (the worker can present his/her Disability Certificate to the firm). Although this requirement is not compulsory, employers will presumably report this information when they are aware of it, and it is mandatory to obtain financial incentives provided by the Public Administration.
All individuals with at least one employment spell where they are identified as ‘disabled’ were selected for the analysis. The entire labour trajectory of these individuals is considered. Spells corresponding to receipt of unemployment benefits or pensions have been excluded. Until 2000, labour contracts corresponding to SECs were identified as a different type of contract in the database. This specific contract for SECs allows us to identify employers that are SECs. This is important because from 2001 onwards, this specific type of contract is not separately identified in the database: it is included with all other open-ended contracts. As the dataset includes a unique identification number for the employer for each employment spell, SECs have been identified using the contracts specifically for SECs that were signed through 2000. All the employment spells linked to employers identified as SECs were selected 10 .
For this study, workers with disabilities are considered to be those with at least one employment spell in 2005. The final sample for the empirical analysis consisted of 9,006 individuals who reported disability in at least one employment spell. The career paths of these people before and after 2005 are included in the dataset. This implies that information about their previous experience, the duration of the selected employment period and any subsequent episodes of work is available.
Socio-demographic features
The sample mainly consists of males (66.8 per cent) because of their higher employment rate in the Spanish labour market. The majority of the sample is between 30 and 44 years old, with a low proportion of older workers (9.1 per cent are over 55). The sample consists mainly of low-skilled workers.
The existence of a disability as well as the corresponding degree of disability is accredited by means of a certificate from the Institute for the Elderly and Social Services. A worker’s degree of disability must be at least 33 per cent to receive subsidies or support. There is a greater proportion of workers with a 33 per cent level of disability in the ordinary firms than in SECs. However, the percentage of workers with a high degree of disability is similar in both kinds of firms.
Existing contracts in 2005 are mainly open-ended contracts, although temporary contracts account for 39.6 per cent of the total. A large proportion of employment spells start during 2005. Spells starting prior to 2001 are less than 20 per cent of the sample. Employment spells in Sheltered Employment Centres are 8.9 per cent of the total. The characteristics of both kinds of employment spells can be observed in Tables 1 and 2. The proportion of men is greater in ordinary firms than in SECs, and workers are younger. In both types of firms, workers are predominantly low-skilled, although the percentage varies and is greater in the case of SECs.
The size of SECs is greater than the size of ordinary firms. Almost fifty per cent of ordinary firms where workers with disabilities are employed have less than eleven employees, while this size of firm represents less than 20 per cent of SECs. On the contrary, firms over 50 employees account for almost two out of three SECs but less than 20 per cent of ordinaryfirms.
With respect to the distribution of workers by sector, there are some interesting differences between SECs and ordinary firms. Workers in SECs are mainly in the Industry, Health and Services to firms sectors. For ordinary firms, the concentration of workers in these sectors is lower. The three sectors cited are the ones with the highest percentages of workers, although Retail trade and Construction employ approximately 10 per cent of workers.
In Table 4, data regarding the career paths of workers with disabilities are displayed. As it is done in the previous section, the sample of workers with disabilities is divided into two groups: those who are working in SECs and those who are working in ordinary firms in 2005. They begin working their first job at around age 22 (a bit older in the case of workers with disabilities who are employed by SECs). Prior to the existing employment spell in 2005, workers with disabilities in ordinary firms and SECs have worked for an average of 9.5 years and 8.4 years, respectively. Although the proportion of workers withunemployment experience is lower among those working at SECs, those who have been unemployed have experienced longer periods of unemployment. A worker’s previous experience in SECs is higher for workers with disabilities currently working in SECs than those working in ordinary firms. As informed by the previous sections, this is an expected result because SECs can be a ‘protected’ labour market for people with disabilities; it can also be the beginning and end of their career path.
The figures for the workers with permanent contracts are similar to those for the whole sample; the differences between workers in ordinary firms and in SECs are similar.
It is examined whether the contracts existing in 2005 are still in force in the following years through 2011. The proportion of contracts still in force is 29.6 per cent in the case of SECs and 17.9 per cent in ordinary firms. Workers who are not in the same employment situation mainly moved to other jobs, but this group is larger for workers previously employed in ordinary firms than in SECs (70.6 per cent vs. 54.7 per cent). The percentage of workers who are out of the labour force in 2011 is 36.4 per cent for workers in SECs in 2005 and 47.2 per cent for workers in ordinary firms in 2005. This result can be related to the economic crisis. SECs have increased their hiring during the crisis, while ordinary firms have decreased their hiring [13]. This is an advantage for workers with disabilities in SECs because the activity of these firms seems to be less affected by the recession; thus, they have better opportunities for staying in the labour market.
Therefore, SECs seem to offer more stable jobs than ordinary firms. However, it is difficult for workers in SECs to find a job with an ordinary firm. When these workers in ordinary firms lose their jobs, they generally move to work in other SECs or leave the labour market altogether. In the next section, these questions are studied in depth.
Duration of employment spells
In this section, the duration of employment spells in 2005 is analysed. The database allows to follow these spells until 2011. As fixed-term contracts are more unstable, we will focus on permanent contracts and analyse if contracts in SECs are more stable than contracts in ordinary firms.
In Fig. 1, the Kaplan-Meier survivor functions are presented. They are defined as the probability of surviving during a given length of time. In this case, it is the probability of being in the same employment spell that existed in 2005. In Fig. 1, the survivor functions depending on whether the firm is a SEC or an ordinary firm are showed. The stability of employment spells is higher in SECs than in ordinary firms 11 .
Although there are differences in the stability of employment spells in SECs and ordinary firms, we have to take into account that different kinds of firms employ different kinds of workers, as it has been established in the previous section. These differences in the characteristics of employees could lead to higher stability in SECs than in ordinary firms. Therefore, SECs per se would not necessarily provide more stability to workers with disabilities. In the next section, duration models are used to estimate the determinants of survival in employment.
Duration models
Determinants of survival in employment were analysed using duration models. This kind of model allows the analysis of a dependent variable that is the time spent from the start of the employment spell with censored observations (information referred to the time that the employment spell began is available; however, in the last period of the sample, several spells have finished while others continue).
To take into account the nature of the data, the so-called “hazard rate” measures the probability of making a transition out of the current employment spell at each instant, conditional on survival up to that point. If T is the length of a completed spell, the failure function is:
Thus, the survivor function is:
The probability density function is defined as the slope of the failure function:
With these elements, the hazard rate is defined as:
θ (t) Δt, can be understood as a conditional probability of having a spell of length t exactly;that is, it gives the probability that the spell finishes in the interval (t, t + Δt), conditional on survival up to time t[2].
Cox’s proportional hazard model is used to study the determinants of duration of employment. Its advantage is that it is not necessary to make any assumptions about the shape of the baseline hazard function, unlike parametric duration models. The model is specified as follows:
h(t) is the risk function, and h(t/X) is the risk function taking into account the information contained in the independent variables X (sex, age, industry, qualification, etc.). In the case of h0(t), this information is not considered. It is assumed that the risk function depends separately on time and the independent variables. In the Cox regression, g(X) = e z,where
Results
The independent variables included in the model are the socioeconomic characteristics of workers (sex, age, degree of disability, level of qualification and region of residence), their career path (working time, experience in unemployment and seniority), and the characteristics of firms (industry and size). It is also included whether the firm is a SEC or not because this is the main variable of interest.
Results for the Cox model are presented in Table 5 for the full sample and separately for men and women. The magnitude of the hazard rates is similar for both men and women, so comments will focus on the whole sample. With respect to the variable of interest – being employed in a SEC – the hazard rate is less than one; this estimate implies that employment in SECs is more stable than in ordinary firms, even after controlling for the effects of other variables.
In general, the effects of the socioeconomic and firm characteristics are of the expected signs. Men have a higher likelihood of surviving in the same employment situation. Young workers have a lower likelihood of surviving in the same employment situation. The degree of disability has a negative effect, indicating that those with higher degrees of disability have higher probabilities of non-survival in employment.
The majority of workers with disabilities are of low skill. The effect of qualification is small, and its significance is limited. However, a positive relationship between qualification and stability in employment is found. Labour experience has a negative relationship with survival: those with less experience have lower exit rates. This experience is not necessarily observed in 2005; thus, the interpretation of this result is that if workers lose their jobs after a long spell of employment, it would be difficult to go back to work in a different firm.
With respect to the different sectors, Restaurants and hotels, Services to firms and Construction have the highest hazard rates with respect to the omitted category (Industry). Workers with disabilities in these sectors have higher probability of losing their jobs. In contrast, the Agriculture and Health sectors are characterised by hazard rates greater than one: these results indicated that stability in employment is higher in these sectors than in the others. In analysing the effect of firm size, firms with less than eleven employees are used as reference group. The hazard rate estimates are below one for the rest of the categories. Thus, the probability of exiting the job observed in 2005 is highest in the smallest firms.
Discussion and conclusions
Sheltered employment centres are the main active labour market policy targeted at workers with disabilities in Spain. The aims of these centres are reducing unemployment and inactivity among workers with disabilities and fostering their integration into the ordinary labour market. The empirical evidence is not clear about their performance. Working in these centres provides a more stable career path with lower turnover than working in ordinary firms. Employees in these centres tend to have less favourable professional situations (lower wages and worse working conditions), so employment in SECs can lead to the immobility and instability of workers with disabilities.
The aim of this paper was to study whether employment in SECs is more stable than employment in ordinary firms using panel data ranging from 2005 to 2011. Results confirm that workers with disabilities in SECs have more stable jobs than workers with disabilities in ordinary firms. People with disabilities working in SECs in 2005 have a higher probability of being employed in the same job after six years than their counterparts employed by ordinary firms.
However, an open question is whether the public programmes targeted at SECs allow them to offer more stability to their workers. Given the support policies and the subsidies these firms receive when they hire workers with disabilities, we might think that this support contributes to the stability of these workers. Nevertheless, working conditions in SECs are not as good as in ordinary firms in terms of wages [12]. It is an open question whether poor working conditions with greater stability is better than good working conditions with less stability; and this is an important line of inquiry with respect to the labour and social integration of people with disabilities.
Available data do not allow for the analysis of SECs as a means to increase the participation of workers with disabilities in the labour market. However, it is possible to study SECs role in preventing such workers from leaving the labour market once they are employed. Results reinforce the image of the SECs as a method for keeping workers with disabilities participating in the labour market. Nevertheless, these results should be considered in determining the public resources SECs receive (with respect to ordinary firms). Therefore, the stability of employment in SECs can be associated with the public funding they receive or with other aspects of their operations. More research is needed to have sufficient information regarding these questions.
Conflict of interest
The authors have no conflict of interest to report.
Footnotes
Appendix
1
Vocational Rehabilitation and Employment (Disabled Persons) Recommendation (No. 168), 1983.
2
Regarding supported employment in the case of workers with mental disabilities, there are not significant differences between the quality of life of people in supported employment and people in SECs [6, 
]. However, they found that workers who benefit from supported employment enjoyed higher seniority and salaries than those who are in SECs, but workers in supported employment received lower levels of employment benefits (sick-pay, medical care, etc.).
3
At the international level, other forms of ‘protected’ employment for people with disabilities have been proposed as an alternative to increase the employability of these individuals in ‘ordinary’ firms. Among these, the most important type of program is supported employment. The use of supported employment in Spain is very limited (see [] for a survey on supported employment in Spain).
4
In fact, there are no labour contracts in occupational centres. In addition, the goods and services produced in occupational centres are not directed towards goods or services markets.
5
There were exceptions for specific job positions that were crucial for the normal economic and productive activity of the SEC.
7
8
Labour relations in the SECs are considered to be a special employment relationship, but it is subject to very different regulation than that which applies to ordinary workers, except for some peculiarities in certain contracts and dismissal for objective reasons. The biggest disparity is the different employment promotion measures applied to both groups and the related qualification requirements. They are generally more substantial and flexible, and access requirements are less stringent.
9
10
After 2001, the data on employment spells in SECs probably are biased downward because new SECs cannot be identified. However, if employment in SECs is identified using the Survey on Disability, Personal Autonomy and Dependency Situations 2008, the results show that this bias is of low magnitude.
11
Results could be biased given that seniority in the existing employment spell in 2005 is different in SECs and ordinary firms; in , the survivor function for permanent contracts beginning in 2004 is displayed. The conclusion is the same: the stability of employment spells is higher in SECs than in ordinary firms, especially after two years.