Does self-employment provide a bridge to retirement? (2024)

,

Article history

Abstract

1. Introduction

The percentage of the European Union’s (EU) over-65 population will rise from 21% in 2020 to almost 30% in 2050, causing the dependency ratio to increase from 32 to 52% (OECD/European Commission, 2021). Although labour force participation for workers over 55 has risen in the past 30 years, the increase is insufficient to counterbalance the increase in economically inactive older individuals who need support from the working economy (Grigoli et al., 2022). Given the increasing pressure on pension systems and related welfare payments as life-expectancy increases, there is increased focus on designing policies to prolong working life using alternative working practices. In recent years, senior entrepreneurship—starting a business after the age of 50—has received policy attention as a possible measure to address the challenges caused by socio-demographic changes, such as a rising old-age dependency ratio (OECD, 2019). Senior entrepreneurship may stimulate employment rates among older individuals by providing additional employment opportunities for those facing age-related discrimination, extending active employment at older ages and thereby reducing the pressure on public pension and related welfare payments (Van Solinge, 2014). Considering the recent policy discourses about reforms aimed at increasing retirement ages (Komp, 2018) and the promotion of self-employment among seniors (Stypinska, 2018; OECD/European Commission, 2021), we know very little about the decision-making process and circ*mstances of the individuals that decide to switch to self-employment at older ages, prior to making this switch.

The antecedents of entrepreneurship are often categorised based on the underlying motivational factors that have led to the decision of individuals to become self-employed (Tyrowicz, 2011), namely pull and push motives (Pilkova et al., 2014). Older individuals with higher levels of experience, skills or accumulated wealth might choose self-employment as a late-career route to enjoy higher flexibility or autonomy, actualise an opportunity they have recognised in the market or apply their prior skills, knowledge and experience in a relevant field (Zissimopoulos and Karoly, 2007). On the other hand, push factors may force older individuals into self-employment, such as the absence of better job alternatives, money shortages at older ages, age discrimination in the job market or workplace and poor job security or prospects (Kautonen et al., 2014). In many cases, senior entrepreneurship is not solely driven by either pull or push factors and a combination of both types of motives could be at play in shaping one’s decision (Zissimopoulos and Karoly, 2007; Kerr and Armstrong-Stassen, 2011). While the role of pull factors could be more dominant in the entrepreneurial decision of the more advantaged groups, the less privileged groups might be pushed into self-employment as a result of a lack of a better option (Van Solinge, 2014). This includes the so-called precarious workers (Standing, 2011, 2014) who may opt for self-employment as a last resort. Finally, switches into self-employment are determined not only by the ambition to become self-employed and its different motivations, but also by the ability to realise such an ambition, that is, the ability to accumulate sufficient resources to start a business and become a self-employed worker (Clough et al., 2019). All in all, the decision-making process of older individuals to switch to entrepreneurship as a late-career employment option is complex and still not well understood.

Nevertheless, senior entrepreneurship has been proposed as a policy solution to extend the working lives of older workers who tend to be portrayed as underrepresented and disadvantaged (OECD/European Commission, 2021). However, it is still unknown whether self-employment serves as a bridge to retirement and, if it does, we know very little about the individuals involved. In this regard, it is important to realise that large heterogeneity exists within the population of self-employed, and even within the population of solo self-employed (i.e. self-employed without employees), ranging from precarious self-employed workers facing in-work poverty (Séphier Conen, 2021) to highly skilled freelancers providing high-quality solutions to complex challenges faced by incumbent firms, for instance in the area of Information and Communication Technology (ICT) (Burke and Cowling, 2020). So, when discussing self-employment as a bridge to retirement, it is important to know which type of self-employment is being considered. Heterogeneity in self-employment exists on several dimensions, including start-up motive (Larsson and Thulin, 2019), education (Van Stel and Van der Zwan, 2020) and age (Kautonen et al., 2014).

Addressing the gap in our lack of knowledge as to what extent employees are likely to switch to senior entrepreneurship, the present article examines labour market transitions of wage workers aged 50 plus to investigate the job-related, family-related and individual characteristics of those individuals who switch to self-employment to ascertain the circ*mstances in which self-employment is a viable employment option for older individuals. Using a longitudinal data base in European countries over the period 2005–2015, we explore antecedents of switches to other wage-employment, self-employment or non-employment options compared to staying in the same wage job.

The remainder of the paper is structured as follows. We review the literature in Section 2, describe our data and methods in Section 3, our results in Section 4 and finally we conclude with a discussion of our findings in Section 5.

2. Review of the literature

Given the socio-demographic pressures such as a rising old-age dependency ratio (OECD, 2019) created by longer life expectancies in Europe, there is an increased policy focus on prolonging the length of individuals’ working life. The life-course perspective studies transitions in an individual’s work–life cycle (Elder et al., 2003) and a consensus has arisen among researchers that retirement is not a single event but rather a process that older individuals go through over a variable period of years (Szinovacz, 2003; Shultz and Wang, 2011), with dozens of possible combinations of paid work and time out from the labour force (Pleau and Shauman, 2013). The term ‘bridge employment’ is used to refer to jobs that follow career or full-time employment and precede complete labour market withdrawal or retirement from work (Feldman and Kim, 2000; Shultz, 2003; Cahill et al., 2015). The transitions characterising bridge employment occur both within the individual’s own profession and in other types of occupations, and they can take the form of (full- or part-time) salaried work, permanent or temporary jobs and self-employment (Beehr and Bennett, 2015).

In recent years, self-employment has received policy attention as a possible measure to prolong active labour participation. Senior entrepreneurship is believed to stimulate employment rates among older individuals and hence benefit the national welfare and economy in three ways. Firstly, senior self-employment can reduce the pressure on public pension and welfare institutions by increasing labour market participation and reducing early withdrawals from the labour force, including early retirement (Van Solinge, 2014). Transitions into unemployment at older ages have been found to be more likely to evolve into long-term unemployment (Figueiredo and Paiva, 2019); therefore keeping individuals in active employment is particularly important at older ages. Secondly, late-career entrepreneurship provides an employment option for older individuals who due to age-related discrimination in the labour market face reduced wage-employment opportunities or the threat of redundancy (Soto-Simeone and Kautonen, 2021). Thirdly, self-employed senior workers tend to retire later than their wage employee counterparts, leading to more prolonged use of their human capital in the economy (Kautonen et al., 2017, 2023).

2.1 Antecedents of senior entrepreneurship: push motives

The antecedents of entrepreneurship are often categorised into pull or push motives (Pilkova et al., 2014; Henley, 2021). The distinction between pull and push motivations in entrepreneurship is captured through the concepts of opportunity and necessity entrepreneurship (Reynolds et al., 2002). Although there are different ways to measure these concepts, it is generally agreed that opportunity entrepreneurs are motivated by pull factors like seizing a business opportunity, while necessity entrepreneurs are driven primarily by push factors like a lack of other employment options. Indeed, Reynolds et al. (2002), who first introduced these concepts, define necessity entrepreneurship as starting a business because there are no better job options available and opportunity entrepreneurship as starting a business to take advantage of an opportunity. De Vries et al. (2020) report that in the Netherlands, necessity entrepreneurs tend to be outperformed by opportunity entrepreneurs in terms of sales volume and income. Such performance differences may be explained by different levels of education. In a study of older American workers, Abraham et al. (2020) finds education levels to be the decisive factor driving opportunity entrepreneurship, which would be consistent with the finding by Fossen and Büttner (2013) that returns to education are significantly lower for necessity entrepreneurs than for opportunity entrepreneurs and wage employees.

An important category of push factors is formed by precarious work conditions. In his seminal work, Standing (2011) argues that a new social class is emerging in society, the precariat, who are characterised by their insecurity and vulnerability due to precarious work conditions. Precarious employment is associated with unfavourable conditions such as a lack of job security, poor job opportunities, high chance of job loss or inadequate salary (Raymo et al., 2011). The employment characteristics inherent in such job conditions may impact retirement decisions differently as compared to more favourable conditions. For example, Wahrendorf et al. (2013), in a study of individuals’ retirement intentions, found that employees in lower social positions were more likely to be in precarious employment and report intentions to retire. Moreover, Standing (2014) suggests that the traditional retirement model, which assumes stable and secure employment until retirement age, is no longer feasible for many precarious workers. He proposes the development of a new social contract that recognises the changing nature of work and provides a basic income for all citizens, as well as access to healthcare, education and affordable housing. Nevertheless, in the absence of such a new social contract, the availability and feasibility of working alternatives such as self-employment remain especially relevant for the group of precarious workers, as they may experience a higher risk of losing their job (Büchtemann and Quack, 1990).

While push factors may thus include poor wage-employment prospects (Walker and Webster, 2007), financial necessity (Cahill et al., 2015) and the threat of unemployment, the importance of these factors may vary at different ages. For older individuals wage-employment prospects may be lower and the threat of unemployment higher due to age-related discrimination in the job market or workplace (Karpinska et al., 2013). In considering the role of age in studies of opportunity and necessity entrepreneurship, Block and Sandner (2009), Van der Zwan et al. (2016) and Calderon et al. (2017) find that on average necessity entrepreneurs tend to be older than opportunity entrepreneurs. In contrast, Abraham et al. (2020) find that self-employment among older Americans is mostly opportunity-driven, although in their study education level was found to be the decisive factor driving entrepreneurial opportunities. In a study of older US workers’ post-career employment choices, Kerr and Armstrong-Stassen (2011) find that a combination of push factors such as financial necessity and pull factors such as personal fulfilment and independence affect post-career transitions into self-employment. However, financial necessity may be less of a factor at older ages as at older ages individuals tend to have fewer financial pressures than those at younger ages with dependent children and large financial commitments such as mortgages and other debt. Figueiredo and Paiva (2019) suggest that the incidence of long-term unemployment at older ages indicates a reluctance and reduced motivation for necessity entrepreneurship.

2.2 Antecedents of senior entrepreneurship: pull motives

Pull factors that make self-employment a desirable option include flexible work schedules and hours (Quinn, 1999; Zissimopoulos and Karoly, 2007), job autonomy and independence (Bond et al., 2005; Walker and Webster, 2007) and greater opportunities for learning and development (Bond et al., 2005). Flexible work schedules may be particularly attractive for older individuals who may select self-employment as a late-career route to transition to retirement, while providing an opportunity to apply prior skills, knowledge and experience in a relevant field. Moreover, older people are commonly drawn to self-employment for socially motivated reasons such as the intention to give back to their community (Djebali et al., 2023). Soto-Simeone and Kautonen (2021) study the primary social motivations of older workers to engage in entrepreneurial activity. They identify three underlying motivations to specifically play a role at this age: to remain active and valuable despite older age; to have greater control over one’s life; and to help others within the community.

However, there are also reasons why opportunity entrepreneurship may decrease with age. While self-efficacy or confidence in one’s ability is a positive attribute for entrepreneurship success, overconfidence can have a detrimental effect on entrepreneurial performance, in the case where individuals overestimate their skills and ability to assess the risks associated with self-employment (Trevelyan, 2008). At older ages individuals may have a more accurate assessment of their suitability of self-employment, through many years of work experience and may be less prone to unrealistic and overconfident expectations; while at younger ages, nascent and early career entrepreneurs may be overoptimistic regarding their entrepreneurial skills (Bolger et al., 2008) and more likely to engage in opportunity entrepreneurship, regardless of their suitability.

Using data from a survey of entrepreneurial motivations amongst UK respondents, Stephan et al. (2015) find that necessity and opportunity motivations for entrepreneurship have differing relationships with age. They find that opportunity motivation is higher for individuals at an early stage of their career, but that this motivation declines between the mid-30s and mid-40s and increases again thereafter. They suggest that this may be due to family circ*mstances, as those in their 30s and 40s, who often have family responsibilities and financial obligations such as mortgages, are less likely to pursue entrepreneurship for reasons of self-actualisation and may only engage in entrepreneurship if another attractive employment option is not available. At older ages (50+), family and childcare obligations tend to become less time intensive giving individuals more time to seek out opportunities for entrepreneurship. For necessity motivation they find the opposite effect, with necessity motivated entrepreneurship lowest for entrepreneurs at the start of their career, highest for entrepreneurs in their mid-40s and then declining again for older age groups. Given that they tend to have fewer financial obligations and family responsibilities, individuals both at the start and end of their working career are less likely to undertake necessity entrepreneurship, that is, only in the case when there are no other attractive employment opportunities.

The incidence of either necessity or opportunity entrepreneurship may decrease with age, due to increasing opportunity costs of self-employment. The choice of self-employment may be weighed up against retirement benefits and unemployment benefits which may be higher than at younger ages if unemployment benefits are linked to salary, which tends to increase with age. Therefore, the opportunity costs of self-employment vs. non-employment may be higher at older ages. In a study of labour force participation in Europe, Engelhardt (2012) notes that financial incentives to keep older people in the labour market are outweighed by the monetary incentives of the pension systems both in case of employment maintenance and employment exit.

Although the existing evidence in the literature is mixed, overall it seems that senior entrepreneurship is somewhat more associated with pull motives rather than push motives. Therefore, our working hypothesis is that at older ages (50+), workers in precarious job conditions (those who would be pushed to choose entrepreneurship out of necessity) are less likely to switch to self-employment. This working hypothesis resulting from the literature is noteworthy as it is at odds with the earlier mentioned policy intentions of self-employment acting as a bridge to retirement for precarious workers (Stypinska, 2018).

3. Data and methods

This study uses data from the Survey of Health, Ageing and Retirement in Europe (SHARE) project, which conducted interviews with 123,000 individuals aged 50 plus and their spouses across 21 European countries in 2005, 2007, 2009, 2011, 2013 and 2015, collecting data on health, employment, socio-economic status, social networks and other demographic factors (Boersch-Supan, 2017). Our outcome variable is a categorical variable measuring changes in employment status over a two-year period between consecutive waves of the SHARE survey. It takes on a value of 0 for respondents who are in wage-employment at time t and are in the same wage job at time t + 2 (reference category), a value of 1 if they have switched to a new wage job at time t + 2, a value of 2 if they have switched to self-employment at time t + 2 and a value of 3 if they have switched to non-employment (retirement, unemployment or out of the labour force) at time t + 2.

Using multinomial logit regression, we explain the probability of switching to each of these three outcome categories, using a number of independent variables which we measure at the start of each two-year period, time t, as we are measuring the effect of antecedents on changes in employment status that have occurred by time t + 2. We categorise our independent variables into three groups: Job-Related Characteristics, Individual Characteristics and Family-Related Characteristics. Our job-related characteristics can be further categorised into those relating to the type of job (Time Pressure due to Heavy Workload, Lack of Job Control, Physically Demanding Job) and those relating to the job conditions (Non-permanent contract, Inadequate Salary, Lack of Job Security, Lack of Job Prospects) which relate to the following survey questions.

• (1) Time pressure due to heavy workload. Survey participants were asked to indicate how much they agree or disagree with the following statement. ‘I am under constant time pressure due to a heavy workload. (Would you say you strongly agree, agree, disagree or strongly disagree?)’ The responses were originally coded into a 4-point Likert-type scale from 1 (strongly agree) to 4 (strongly disagree). They were re-coded in the current study so that higher values indicate heavier time pressure and workload.

• (2) Lack of job control. Participants were asked whether they think they have very little control in their job. ‘I have very little freedom to decide how I do my work. (Would you say you strongly agree, agree, disagree, or strongly disagree?)’ The responses were initially coded from 1 (strongly agree) to 4 (strongly disagree). They were re-coded for this study so that the higher values indicate less job control.

• (3) Physically demanding job. Participants were asked the following question. ‘My job is physically demanding. Would you say you strongly agree, agree, disagree, or strongly disagree?’ The responses were originally coded from 1 (strongly agree) to 4 (strongly disagree). They were re-coded so that higher values indicate that the job is more physically demanding.

• (4) Non-permanent contract. Respondents are asked ‘In this job, do you have a short-term or a permanent contract? By short-term we mean less than 3 years 1. Short-term 2. Permanent’. We re-coded this variable so that a higher value indicates a short-term contract.

• (5) Inadequate salary. Respondents are asked about whether they think their earnings are adequate. ‘Considering all my efforts and achievements, my [salary is/earnings are] adequate. (Would you say you strongly agree, agree, disagree or strongly disagree?)’ The responses are coded from 1 (strongly agree) to 4 (strongly disagree).

• (6) Lack of job security. Respondents are asked to respond to the following statement. ‘My job security is poor. (Would you say you strongly agree, agree, disagree or strongly disagree?)’ The responses were originally coded from 1 (strongly agree) to 4 (strongly disagree). We re-coded this variable so that the higher values indicate less job security.

• (7) Lack of job prospects. Respondents are asked to respond to the statement, ‘My [job promotion prospects/prospects for job advancement] are poor. (Would you say you strongly agree, agree, disagree, or strongly disagree?)’ The responses were originally coded from 1 (strongly agree) to 4 (strongly disagree). We re-coded this variable so that the higher values indicate less job prospects.

To control for Individual Characteristics we include measures of Education, Good Health, Money Shortage, Job Dissatisfaction, Age and Gender.

• (1) Education. Respondents are asked ‘What is the highest school leaving certificate or school degree that you have obtained?’ for which the range of answers is country specific. The SHARE dataset uses the ISCED education level classification system to classify the survey information on education into a six-level education variable.1 The application of ISCED facilitates the international comparison of education levels of various countries, which would not be possible otherwise.

• (2) Good health. Participants are asked to rate their health on a five-point scale. ‘Would you say your health is... 1 = excellent to 5 = poor)’. We re-coded this variable so that the higher values indicate better health.

• (3) Money shortage. Respondents are asked the following question. ‘How often do you think that shortage of money stops you from doing the things you want to do? (Often, sometimes, rarely or never?) 1. Often 2. Sometimes 3. Rarely 4. Never’. We re-coded this variable so that higher values indicate greater levels of money shortage.

• (4) Job dissatisfaction. Respondents are asked to respond to the statement, ‘All things considered I am satisfied with my job. Would you say you strongly agree, agree, disagree, or strongly disagree?’ The responses are coded from 1 (strongly agree) to 4 (strongly disagree).

• (5) Age. We include the respondents’ age and age squared. We restrict our sample to include only those aged 50–70.

• (6) Gender. 1 indicates male and 2 indicates female.

Finally, we include two Family-Related Characteristics, one that relates to marital status and one that relates to family responsibilities.

• (1) Living with partner. Using survey data we create a categorical variable that takes a value of 1 if the respondent lives with a partner and 0 if not.

• (2) Heavy family duties. Participants were asked ‘How often do you think that family responsibilities prevent you from doing what you want to do? (Often, sometimes, rarely or never?)’ The responses are originally coded from 1 (strongly agree) to 4 (strongly disagree). We re-coded this variable so that a higher value indicates greater limitations due to family responsibilities.

• In Table 1 we present the descriptive statistics of the variables included in the regression models for our estimation sample size of 5,628 observations.2 The mean values for most of the job-related characteristics are slightly below average, with the exception of Lack of Job Prospects which is above the average value of 2.5. Non-Permanent Contract is below the average value of 0.5, the average respondent is more likely to have a permanent work contract at the start of the two-year period. The average level of education is above upper secondary education. Good Health is slightly above average, Money Shortage is close to its average value, while Job Dissatisfaction is relatively low. Age is skewed towards younger participants with an average age of 55.7 and there are more female respondents than male. Most respondents live with a partner and on average respondents have a lower than average amount of family duties. In Table 2 we display the pairwise correlation coefficients between our variables. The highest correlation is between Inadequate Salary and Money Shortage at 0.342, and the highest Variance Inflation Factor (not displayed) is 1.29, therefore multicollinearity is not considered to be an issue of concern.

4. Results

We use a multinomial logit regression with standard errors clustered by country and report the marginal effects as well as the average predicted probability for each category of outcome variables in Table 3. For job-related characteristics, Time Pressure due to Heavy Workload has a negative association with switching to non-employment. This variable refers to employees being under constant time pressure which may indicate that these employees are essential to the functioning of the company, and whom a firm would try to retain. A Physically Demanding Job has a positive association with switching out of active employment. This may indicate that those working in physically demanding jobs choose to retire early, possibly because of a lower retirement age being legislated for in the case of physically demanding jobs. For the remaining job-related characteristics, we find that workers with worse job conditions are more likely to switch to non-employment, those with Non-Permanent Contract, Lack of Job Security and Lack of Job Prospects. Although some wage workers without a permanent contract or good job security are also found to switch to another wage job, in both cases the estimated marginal effect is over three times stronger for switches to non-employment. None of these job conditions is found to have a positive association with a switch to self-employment, but we find that a lack of job prospects is negatively associated with a switch to self-employment. In other words, workers with good job prospects in their current wage job are more likely to switch into self-employment. Finally, we find that an Inadequate Salary has no association with switches from a wage job, suggesting that most people who switch their employment status do so for reasons other than economic.

For our individual characteristics, we find that more highly educated workers are less likely to switch into non-employment but we do not find that education has a significant association with switches into self-employment. We also find that healthier workers are less likely to switch into non-employment and more likely to switch to self-employment. Money Shortage is found to have no effect on switches from a wage job, consistent with the finding for Inadequate Salary. Job Dissatisfaction is positively associated with switches into another wage job and into non-employment but not with switches into self-employment. Finally, for our family-related characteristics, Heavy Family Duties have no significant association with any type of switch, while Living with Partner makes a switch to another wage job less likely. Overall, we find that individuals switching to self-employment are those with good job prospects and good health, while individuals with precarious job conditions, such as poor job prospects, poor job security and a non-permanent contract are most likely to switch into non-employment, which suggests that the transitions into self-employment that we observe in our dataset by senior workers are opportunity-based (indicating pull factors) rather than necessity-based (indicating push factors).

5. Discussion and conclusion

Low rates of labour market participation among older workers in the EU and increasing life-expectancy, have led to an increased focus on designing policies to prolong the working life of older workers (OECD, 2006) including flexible retirement options (e.g. the UK government allows individuals to access their pension savings prior to the normal pension age to provide older workers more flexibility to continue working while drawing a pension3), stimulating training and skills development (European Commission, 2016) and improving employer practices to retain older workers (e.g. the Danish ‘senior starter kit’ offering support and subsidies for health and safety at work as well as flexible work arrangements; OECD, 2015).

5.1 Pull vs. push factors for senior entrepreneurship

In addition, senior entrepreneurship has received policy attention as a way to prolong the working lives of older people, reduce older-age unemployment and enhance the social inclusion of older individuals (Kautonen et al., 2008). Older individuals with higher levels of experience, skills or accumulated wealth might choose self-employment for different reasons than younger workers. Antecedents of self-employment are often categorised into pull or push factors, leading to opportunity or necessity entrepreneurship. Pull factors for senior entrepreneurship may include its attractiveness as a late-career option which may offer higher flexibility or autonomy, a means to actualise an opportunity they have recognised in the market or an opportunity to apply their prior skills, knowledge and experience in a relevant field. Alternatively, push factors such as the absence of better job alternatives, money shortages at older ages, age discrimination in the job market or workplace and poor job security or prospects, may force older individuals into self-employment (Kautonen et al., 2014). We consulted the literature on pull and push motivations for entrepreneurship amongst older workers and based on this we formulated the working hypothesis that precarious workers (those who would be pushed to switch to self-employment out of necessity) would be less likely to transition into entrepreneurship at older ages. Our working hypothesis was confirmed by our empirical analysis.

In our study for 21 European countries, we observe that switches from wage-employment to self-employment at older ages are preceded by pull factors related to opportunity entrepreneurship and not by push factors related to necessity entrepreneurship, such as poor job-related or individual characteristics. We find that wage workers switching into self-employment are in good health and have good job prospects which indicates that they could also easily have chosen to continue in wage-employment or switch to another wage job (opportunity entrepreneurs). Moreover, we do not find that transitions to self-employment from full-time employment is a likely outcome for precarious workers who might otherwise end up non-active in the labour market (necessity entrepreneurs). Instead, we find that several factors related to precarious job conditions precede a switch to non-employment and while some of these conditions are also linked to switches to another wage job, the probability of switching to non-employment is far higher.

We also found important differences between switches within the waged labour market (from one wage job to another) and switches from wage-employment to self-employment, as these two types of switches were determined by different factors. We found that switches to another wage job were determined by precarious conditions like a non-permanent contract, a lack of job security and dissatisfaction with one’s current wage job, whereas switches to self-employment were associated with positive circ*mstances like good job prospects and good health. This shows that older wage workers in precarious working conditions actively try to improve their labour market situation, but they only seem to consider switching to another wage job as a realistic option to achieve this goal, not switching to self-employment. Furthermore, we found that the probability of switching to other wage-employment was lower for individuals living with a partner, perhaps indicating less flexibility to change their labour market situation in the first place. Finally, we found a lower probability of switching to self-employment for women, confirming a stylised fact in entrepreneurship literature that, at least in more developed continents like Europe, entrepreneurial propensity is indeed lower for women (Reynolds et al., 2002; McAdam, 2023).

5.2 The roles of preference and ability in switching to self-employment

We have explained our main findings in terms of push and pull motivations to transition into self-employment, which are associated with poor respectively favourable job conditions and personal circ*mstances. However, it is important to recognise that besides motivation, the ability to select into self-employment may play a role in determining labour market transitions of individuals. In this regard, in the spirit of Hayward et al. (1994),4 the selection into self-employment may be viewed as a two-step process, where step 1 entails a preference for self-employment over wage-employment and step 2 entails the ability to realise one’s preference and actually shift into self-employment.5 While push and pull motivations may explain step 1 (a preference for self-employment, for different motivations), they cannot explain step 2. That is, even if individuals have a preference to shift into self-employment, not everyone will be able to realise their preference and actually become self-employed. This is because (successfully) running a business as a self-employed worker requires various forms of capital, such as human capital (entrepreneurial skills; Unger et al., 2011), financial capital (to make necessary investments, e.g. to buy equipment) and social capital (to build and maintain a customer base). Empirically we are not able to distinguish between these two steps of the selection process into self-employment, and it may well be the case that our findings reflect this selection effect, in which workers in a precarious position on the waged labour market also encounter difficulties in finding suitable self-employment opportunities and where these precarious positions (e.g. a lack of job prospects in the current job or a poor health) possibly reflect a lack of (human, financial and/or social) capital or, more generally, a lower ability to successfully run a business. Individuals in precarious positions, even if they have a preference for self-employment, may recognise their lower entrepreneurial ability and refrain from switching into self-employment. Conversely, individuals in strong labour market positions may also have a higher entrepreneurial ability, and if they have a preference for self-employment, it will be easier for them to realise those preferences.

5.3 Policy implications

Overall, we find that self-employment may serve as a bridge to retirement but mostly for older wage workers in positive circ*mstances, with good health conditions and job prospects and less so for older wage workers in precarious working conditions. The reasons for these differences may be related to a higher financial risk for precarious workers compared to workers in more positive circ*mstances. First, precarious workers may have lower savings accumulated over their working lives (which may in part define their precarious position), so if their business fails, they have less savings to rely on compared to opportunity-based entrepreneurs. Second, precarious workers more often have practical skills rather than analytical skills (Standing, 2011), which tends to make their earning capacity as an entrepreneur lower. In particular, precarious workers less often qualify for high-earning freelancing jobs in the knowledge-intensive services sector as these jobs typically require analytical rather than practical skills (Burke and Cowling, 2020). Third, due to the difference in skill sets, business options open to precarious workers more often require higher capital investments. For example, precarious workers with a practical skill set might start businesses with physical business premises (e.g. to run a cafe or restaurant), while workers with analytical skills could offer their knowledge-intensive services as freelancers with no additional capital investments. Fourth, by voluntarily switching from wage-employment to self-employment, precarious workers often give up their right to unemployment benefits in case their business fails. This is so because in many countries the self-employed are not part of a collective unemployment insurance system, while private unemployment insurance is often too expensive. Although this also holds for non-precarious workers who become self-employed, the financial risk of foregoing unemployment benefits is higher for precarious workers given their often smaller savings to fall back on. Similarly, switching to self-employment may make it more difficult to continue building up pension rights, as current pension policies across the EU disadvantage self-employed workers, particularly precarious self-employed workers. For example, in some EU countries there is a minimum income threshold in place to be allowed to join self-employment pension schemes (Christie et al., 2022).

This makes the financial risk for individuals entering into self-employment from precarious conditions fourfold. First, they have lower savings. Second, they have lower earning capacity. Third, their entrepreneurship endeavours may require higher capital investments and fourth, they may forego unemployment benefits and pension rights. Older precarious wage workers may recognise these higher financial risks and therefore consider other wage jobs to improve their labour market situation but not the option of self-employment. Indeed, they may well be aware that a switch to self-employment may result in precarious self-employment (Conen and Schippers, 2019) and therefore, they hang on in the waged labour market for as long as they can.

Policy should therefore focus on this group of precarious workers and mitigate their financial risks, so that shifting to self-employment becomes more attractive to them. In this regard, it is worrisome that ‘Overall, senior entrepreneurship policies and programmes are under-developed in the EU’ (OECD/European Commission, 2021, p. 156). There are only a small number of entrepreneurship initiatives and schemes designed specifically to support older people in starting a business, and those that exist tend to be very small-scale (OECD/European Commission, 2021). Crucially, while there are not many senior entrepreneurship policy initiatives in the first place, such initiatives seem to be non-existent for precarious older workers (Stypinska, 2018). This is problematic because this group, when switching to self-employment, has a more than proportionate chance of ending up in precarious self-employment (Conen and Schippers, 2019), in which solo self-employed workers face hardships associated with underemployment and low paid work assignments, often obtained via digital platforms in the gig economy (Boeri et al., 2020). Hence, if policies are intended to offer help to precarious older workers in becoming self-employed, they must make sure that they do not end up in precarious self-employment but in a legitimate self-employment job enabling them to earn a decent living so that self-employment can indeed become a viable bridge to retirement. To make self-employment a realistic bridge to retirement for older wage workers, tailored policy programs should be designed that offer help in the areas of building entrepreneurship skills, facilitating access to start-up finance, expanding entrepreneurship networks and making social security regulations more attractive for (senior) entrepreneurs (OECD/European Commission, 2021).

5.4 Limitations and concluding remarks

Our paper has limitations. A limitation of our study is that our data do not allow us to investigate the sectors of economic activity that are associated with transitions to self-employment. Although sector data for self-employment occupations is available in the data set, this information is available for a minority of our switches to self-employment only. Future research using other data sets may look into the sector of economic activity in order to shed light on the type of self-employment activity that older-age switchers to self-employment embark on.

A second limitation is that, insofar as non-movements into self-employment are involved, we are unable to distinguish empirically to what extent these non-movements are due to a lack of preference for self-employment or an inability to act on those preferences and realise one’s ambition to become self-employed. Future research using more refined data may look further into this important distinction between entrepreneurial preference and entrepreneurial ability.

A third limitation is that, to get a more complete picture of the role of self-employment as a bridge to retirement, not only switches from wage-employment to self-employment by older labour force participants should be studied, but also the subsequent transition from self-employment to labour market exit. It is important to know how seniors switching from wage-employment to self-employment perform in their self-employment job, particularly in terms of survival as this determines the degree of prolongation of their working life. This question is left for future research.

We conclude that self-employment may indeed serve as a bridge to retirement but not to all wage workers equally. In spite of the relatively high danger of moving into non-employment for individuals in precarious job conditions, switching to self-employment does not seem to be a preferred or feasible option. When evaluating the success of policies designed to encourage senior entrepreneurship with the aim of reducing the dependency ratio, we find that only those workers who are less likely to rely on the government for pension or welfare benefits are switching to self-employment, thereby having little to no effect on reducing welfare supports. The policy implication of our study is that policies that are designed to encourage self-employment as a means to extend the active working lives of older workers may not be having the desired effect, especially when considering precarious older workers. The reluctance of precarious workers to transition into self-employment may indicate a lack of (analytical) skills which possibly explains both their precarious job conditions and a low entrepreneurial ability. A second possible explanation is related to financial risk including the high opportunity costs of switching from wage-employment toward self-employment, particularly in countries with high unemployment benefits (Koellinger and Minniti, 2009), which may act as a disincentive (Kautonen et al., 2014; Trlifajová, and Hurrle, 2019). Switching to self-employment will not only reduce access to unemployment benefits if the business fails, but also reduce the time in which full pension rights can be accumulated.

Conflict of interest statement. None declared.

Bibliography

1

The ISCED classification system takes the value of 0–6 where 0 represents pre-primary education, 1 is primary education or first stage of basic education, 2 is lower secondary education or second stage of basic education, 3 is upper secondary education, 4 is post-secondary non-tertiary education, 5 is first stage of tertiary education and 6 is second stage of tertiary education.

2

These 5,628 observations correspond with 5,593 individuals, where each observation in our estimation sample relates to an individual who participated in consecutive waves of the survey two years apart, t and t + 2. Hence, only 35 survey participants appear in more than one pair of consecutive waves of the survey or less than 1% of our observations. This makes clustering standard errors at the individual level unnecessary. We do cluster standard errors at the country level while also including a set of country dummies in our regression model.

4

Hayward et al. (1994, p. 84) model the re-entry to work after retirement as a two-step selection model, where step 1 entails retirees’ self-selection into employment and step 2 entails the market selection of retirees from the pool of available labour.

5

We are grateful to an anonymous referee for pointing this out.

© The Author(s) 2024. Published by Oxford University Press on behalf of the Cambridge Political Economy Society.

Issue Section:

Article

Email alerts

Citing articles via

More from Oxford Academic