How earning capacity moderates the age-self-employment relationship

Self-employment is an inherently risky career choice, where investing money and time does not always result in favourable outcomes (Kautonen et al., 2017; Zhao et al., 2021). An individual's decision to become self-employed depends on a myriad of factors, including their prior financial situation and their current age (Beiler, 2017; Schulz et al., 2021; Soto-Simeone and Kautonen, 2021). The self-employment decision can be viewed as an evolving and ever-changing process during an individual's life (Brieger et al., 2021; Perenyi et al., 2017). Throughout their life, different incentives, such as financial gain or increased autonomy, draw individuals towards self-employment (Gielnik et al., 2018; Kautonen et al., 2017). In addition, various disincentives discourage individuals from entering self-employment, such as the high start-up costs and the resultant career instability (Soto-Simeone and Kautonen, 2021; Zhao et al., 2021). Similarly, the financial situation of an individual has long been identified as a primary factor when contemplating the self-employment decision (Kihlstrom and Laffont, 1979; Evans and Leighton, 1990).

Research on entry into self-employment in the context of an individual's age has become increasingly popular in recent years (Syed et al., 2024; Viljamaa et al., 2024). However, while these studies consider an individual's age as a factor in the self-employment process (Hoogendoorn et al., 2022), few consider the changing nature of self-employment entry mechanisms that occur due to the age of the individual. Much of the literature focuses on individuals at a particular life stage, with studies examining younger individuals (Fubah et al., 2025; Krafft and Rizk, 2021) or those in their later years (Soto-Simeone and Kautonen, 2021; Viljamaa et al., 2024). While these studies help us understand how different influences impact the decision to become self-employed for different age groups separately, they do not show if these impacts are consistent or differ across cohorts. Our study addresses this by investigating the decision to become self-employed across the entire lifespan, considering young, middle-aged and older individuals, as opposed to only focusing on a certain life stage. It unlocks important insights into our understanding of how individuals enter self-employment through highlighting whether the influences on the decision to become self-employed are consistent across the age spectrum or if they change due to the age of the individual when they enter self-employment.

Another issue that is absent from the current literature is how moderating factors influence an individual's decision to enter self-employment across an individual's life (Kautonen et al., 2017; Stirzaker and Sitko, 2018). One such moderating factor, an individual's earning capacity, may aid in furthering our understanding of the relationship between age and entry into self-employment. The impact of earning capacity as a moderating factor between age and the decision to enter self-employment is not clearly understood across an individual's lifespan (Jiao et al., 2021; Zhao et al., 2021). In particular, Jiao et al. (2021) highlight that despite considerable research conducted on the decision to enter self-employment, it is not understood how an individual's access to financial resources at different ages plays a role in their decision to migrate from being an employee in the labour market to becoming self-employed. Zhao et al. (2021) further emphasise that while the impact of mediating factors on the self-employment decision has been investigated in certain contexts, their impact has not been considered with respect to the individual's age and whether their impact changes. This article addresses this gap and significantly extends the existing literature by analysing an individual's pay from employment prior to entering self-employment as a moderating factor, as well as its potentially heterogeneous impact on the decision to enter self-employment across different life stages. Understanding how an individual's financial circumstances in conjunction with their current age, impact their decision to enter self-employment is important because of how it can help to further unlock an individual's motives and circumstances when entering self-employment. While an individual's pay from employment is just one aspect of their financial resources, it is a key indicator of their financial capabilities and may help to understand how an individual chooses to enter self-employment.

Logistical regression models are employed to investigate the relationship between an individual's earning capacity and age cohort and their propensity to enter self-employment, utilising data from the “UK Household Longitudinal Study” (Postel-Vinay and Sepahsalari, 2023). The latter provides a total sample of 61,763 observations over the period 2009–2020 (N = 15,218 unique individuals).

The remainder of this article is structured as follows: firstly, the literature regarding the decision to enter self-employment, in the context of their age and earning capacity, is discussed. The theoretical hypotheses are also presented. Then, an overview of the data used alongside the methodology to test the hypotheses is provided. After this, the results of econometric estimations and a discussion of their implications are presented. Finally, the article is concluded.

There has been considerable development in the understanding of how an individual's age impacts their decision to transition from paid employment to self-employment (Zhao et al., 2021). It has been conceptualised that the motivation to enter self-employment changes throughout the lifespan of an individual (Kautonen et al., 2017). Certain incentives are thought to be more prominent for younger individuals, such as monetary reasons and their ability to be more creative thinkers (Gielnik et al., 2018; Kraft et al., 2022). Whereas other incentives, such as increased quality of life and using acquired skills, are more prominent for older individuals (Jiao et al., 2021; Perenyi et al., 2017).

Financial gain is theorised to be a primary motivating factor to enter self-employment for younger individuals (Zhao et al., 2021). Younger individuals are more likely to form new businesses by identifying market gaps and offering products and services to exploit these opportunities (Gielnik et al., 2018). This opportunity exploitation can lead to higher firm profits once trading begins. These increased profits incentivise younger individuals to become self-employed.

Psychological and cognitive reasons, such as risk perception and creativity, are also reasons why younger individuals have a greater propensity to engage in firm start-ups, as they view risk differently than their older counterparts (Camelo-Ordaz et al., 2020). This is highlighted by younger individuals often being endowed with overconfidence and an overestimation of their capabilities in relation to self-employment (Kraft et al., 2022). This can lead to younger individuals being more inclined to transition from paid employment to self-employment, despite the potential obstacles they may face, such as a lack of skills or work experience (Jiao et al., 2021). Put simply, the risk of firm failure does not impact the choice of younger individuals to the same extent as that of older individuals when considering the transition to self-employment.

Older individuals also have incentives to enter self-employment (Kautonen et al., 2017; Perenyi et al., 2017). However, these differ from those of younger individuals (Perenyi et al., 2017). While monetary gain is less of an incentive (Kautonen et al., 2017), other aspects of self-employment attract older individuals. These aspects include allowing them to utilise the skills that they have accumulated as well as the increased quality of life (Hoogendoorn et al., 2022), which is primarily derived from the greater levels of autonomy within self-employment (Dawson et al., 2009; Zhang et al., 2025). Older individuals tend to have greater levels of human capital due to more life experience (Curran and Blackburn, 2001; Lee, 2017). Therefore, self-employment can offer them an avenue to utilise their accumulated skills in creating a business (Agostinho et al., 2025; Gielnik et al., 2018).

However, there are also some barriers to self-employment that older individuals face (Maalaoui et al., 2020). One disadvantage highlighted in the literature is the lower cognitive abilities of older individuals when compared to their younger counterparts, which can result in older individuals being less able to think as creatively or critically about decisions regarding their firm (Weinberger et al., 2018), which can inhibit them in the early stages of self-employment (Gielnik et al., 2018). Additionally, transitioning to self-employment may have a social stigma attached to it for older individuals, as they may not align with stereotypes of an individual who begins their own firm and perceive that doing so is not an appropriate occupation for them due to their age (Maalaoui et al., 2020).

As discussed, several factors attract both younger and older individuals to transition from paid employment to self-employment. However, the reasons for starting a new firm differ (Kautonen et al., 2017). This creates an interesting question: What does this mean for those individuals who are middle-aged and fall between the younger and older cohorts?

It is widely accepted within the current literature on age and self-employment that middle-aged individuals face weaker incentives to enter self-employment than their older and younger counterparts (Lee, 2017; Zhao et al., 2021). While younger individuals often benefit from greater creativity, higher risk tolerance and longer time horizons (Gielnik et al., 2012; 2018), these advantages tend to diminish by mid-life, when financial and family responsibilities may increase risk aversion. At the same time, although older individuals may experience decreasing returns to self-employment and opportunities for growing the business (Brieger et al., 2021; Zhang et al., 2025), these drawbacks may be outweighed by the autonomy and flexibility it offers older individuals. In fact, the well-being benefits associated with autonomy in self-employment appear to rise with age (Kautonen et al., 2017; Stirzaker and Sitko, 2018).

Moreover, other aspects may discourage self-employment entry for middle-aged individuals. Compared to their older and younger peers, they likely have increased responsibilities such as caring for children and elderly parents, as well as increased monthly expenses (Zhao et al., 2021). Therefore, the security of paid employment may be more preferable over the inherently risky career choice of entering self-employment (Block et al., 201a). Furthermore, this cohort is not yet at the point where the employment market becomes limited to them (Lee, 2017; Perenyi et al., 2017), as such, they may not enter self-employment due to unfavourable employment conditions.

This creates a situation where the incentives and circumstances to transition into self-employment may not be as great for middle-aged individuals as they are for younger or older individuals.

Much of the recent literature on the relationship between pay level and self-employment has been viewed through the lens of opportunity and necessity entrepreneurship (Bilan and Apostoaie, 2023; Fossen, 2021). While entrepreneurship is only one type of self-employment (Carter, 2011), it gives an insight into why individuals transition from paid employment to self-employment.

Opportunity entrepreneurship occurs when an individual enters self-employment when faced with multiple appealing options to sustain their livelihood (Audretsch et al., 2022; Estrin et al., 2024). Necessity entrepreneurship occurs when an individual enters self-employment due to an absence of employment choices, or all other options are considerably less appealing than self-employment (Block et al., 2015b). Thus, opportunity entrepreneurship involves an individual being “pulled” into self-employment, while necessity entrepreneurship occurs when an individual is “pushed” into self-employment (Cowling and Dvouletý, 2023; Weber et al., 2022).

Many studies have discovered that individuals with high pay levels, or individuals who share traits with high earners, are more likely to become self-employed (Weber et al., 2022). In addition, the majority of individuals who become entrepreneurs can be classified as “opportunity entrepreneurs” (Block et al., 2015b; Fairlie and Fossen, 2020). These individuals are not inherently on high pay levels, but often share similarities with individuals who are high earners in paid employment, such as the ability to exploit opportunities, possess managerial skills and have higher educational attainment (Block et al., 2015b; Fossen, 2021). Such individuals are also often more willing to take risks with their employment situation and are open to attempting to become self-employed (Block et al., 2015b). It has been conceptualised that having access to financial capital, which will be greater for individuals on higher levels of pay before starting a business, enhances the likelihood of an individual becoming an entrepreneur (Estrin et al., 2024; Weber et al., 2022). The ability to utilise financial resources has several benefits that allow individuals to become self-employed, such as offering a “buffer” if the business fails or requires investment that was not foreseen by the individual (Chliova et al., 2015). Additionally, increased levels of access to finance prior to the formation of the business allow for more innovation to occur at the early stages of the business (Nakara et al., 2021). Having a monetary “buffer” to rely on and jointly having the ability to invest in innovative projects early in the business's life can encourage individuals to become self-employed. In turn, this can reduce the perceived risk that the individual is taking when making the decision to become self-employed.

The primary incentive for lower-paid individuals to become self-employed is that it can offer them a path out of poverty (Henley, 2021). Also, self-employment does give lower-paid individuals similar benefits to those on higher pay levels, such as increased autonomy and freedom in their career progression, despite the different initial situations that both types of individuals face (Santos et al., 2022). The broader economic conditions are also a factor in the propensity of a lower-paid individual to become self-employed (Fairlie and Fossen, 2020). Necessity entrepreneurship is more prominent during times of lower economic activity, as it is counter-cyclical in nature (Fairlie and Fossen, 2020). This increases the push factor due to reduced alternative labour market options to earn during such times (Fossen, 2021). However, while some push factors do exist for lower-paid individuals (Fossen, 2021; Santos et al., 2022), it appears that these individuals do not have the same level of incentives as their higher-paid counterparts to become self-employed (Chliova et al., 2015; Weber et al., 2022).

Early studies on the transition into self-employment conceptualised the choice that individuals make when they enter self-employment as similar to that when they changed from one employment situation to another (Lucas, 1978; Kihlstrom and Laffont, 1979). Kihlstrom and Laffont (1979) claimed that income was the primary driver behind an individual's decision to transition to self-employment from paid employment, as when their expected income from self-employment exceeded their current pay, they would become self-employed. Recent research has also shown that these monetary-related incentives do exist, particularly for younger individuals and for those coming from low-paid backgrounds (Gielnik et al., 2018; Henley, 2021; Zhao et al., 2021).

Younger individuals may make the switch from paid employment to self-employment as they have a greater time horizon to reap the benefits of the potential earnings. They are also more adept, compared to older individuals, at identifying market opportunities in which a potential business would be profitable (Camelo-Ordaz et al., 2020). Individuals with a lower earning capacity also enter self-employment, as it can offer them an improved financial situation (Henley, 2021; Santos et al., 2022). Also, necessity entrepreneurship is seen to increase during recessionary periods (Fossen, 2021). As entrepreneurship is a form of self-employment, this may indicate that individuals switch from paid employment to self-employment to increase their income beyond what is offered to them by the current labour market they are in.

Kihlstrom and Laffont (1979) also argued that as an individual's propensity for risk increases, so too does the expected value of their returns from self-employment. Again, recent studies have shown that this relates to younger individuals (Kraft et al., 2022; Zhao et al., 2021). Younger individuals have been shown to possess overconfidence in their entrepreneurial capabilities and overestimate their potential success, compared to older individuals (Kraft et al., 2022). This, coupled with their lower risk aversion related to their career choices (Zhao et al., 2021), can lead younger individuals to opt into self-employment despite the potential obstacles they may face. Given that younger individuals often possess greater creative capacities, in terms of idea generation, experimentation and lateral thinking, than their older peers (Weinberger et al., 2018), they are more likely to identify market opportunities (Hoogendoorn et al., 2022; Zhao et al., 2021). As such, younger individuals may perceive themselves as capable of establishing a firm that fulfils these market gaps, perceiving the venture as worthwhile due to their longer time horizons to realise the potential benefits. Therefore, it appears that the importance of having a higher pay level before entering self-employment is reduced when younger individuals are considering the switch to self-employment. Having an increased earning capacity provides a financial buffer for prospective self-employment (Chliova et al., 2015). However, having a financial buffer may not be as important for younger individuals when considering opting into self-employment, due to highlighted reasons (Hoogendoorn et al., 2022; Kraft et al., 2022). Therefore, it appears that younger individuals from lower-paid employment may have an increased propensity to enter into self-employment:

Evans and Leighton (1990) brought additional factors to the discussion, such as how individuals are drawn to self-employment for reasons beyond monetary gain, such as greater autonomy, and also showed how liquidity constraints can explain why some individuals do not enter self-employment. This idea of self-employment providing individuals with more than just increased levels of income became a recurring theme when discussing the pull factors of self-employment (Kuvshinikov and Kuvshinikov, 2023). This body of literature argued that high profits were not the primary incentive to enter self-employment for many individuals and that the non-pecuniary benefits it provides are more important (Hamilton, 2000; Soto-Simeone and Kautonen, 2021).

Self-employment can also offer an attractive alternative for older individuals, as they face push factors from the labour market due to lower internal and external employment opportunities (Perenyi et al., 2017). As per Jovanovic (1979), when there is a skills mismatch between a worker and their job requirements, there is an increased propensity for the individual to leave their current employment and seek an alternative. For older individuals who face both this skill mismatch and a labour market with limited opportunities, self-employment may offer an appealing avenue. Adding a further aspect to this, Lazear (2005) highlighted that individuals who opt into self-employment do not have specialist knowledge in any specific subject but a broad range of knowledge in many different areas. This is due to them needing a broad range of knowledge as they must be competent at several things to run their business, particularly in the early stages of their firm. Older individuals tend to have greater levels of human capital due to more experience in employment (Curran and Blackburn, 2001; Lee, 2017), thus transitioning into self-employment would allow them to utilise their gathered skills more productively than if they were to remain in paid employment.

The role of an older individual's pay from employment may also influence their decision to transition into self-employment. As discussed previously, older individuals are more risk-averse than their younger counterparts (Maalaoui et al., 2020; Zhao et al., 2021). Considering this alongside higher pay levels providing increased financial capital that an individual can rely on, the risk associated with becoming self-employed is reduced (Nakara et al., 2021). Therefore, this indicates that having a higher pay level in paid employment may encourage older individuals to transition into self-employment:

These influencing factors on the decision to enter self-employment from paid employment are synthesised as a conceptual framework, see Figure 1. The difference in earning capacity, where younger individuals enter from lower-paid employment and older individuals enter from higher-paid employment is highlighted, in addition to influencing factors that incentivise them to enter self-employment. Here, it can be seen that the decision to enter self-employment for both cohorts is expected to occur in differing earning situations for varying reasons.

The dataset used in this study is “Understanding Society”, or the “UK Household Longitudinal Study”, which is conducted by the Institute for Social and Economic Research (ISER) at the University of Essex, UK The first wave of the survey was distributed in 2009 and the latest wave of the survey available for this research was in 2020. All years from 2009 to 2020 are included in the analysis. The purpose of the dataset is to understand the short and long-term effects of social and economic change in the United Kingdom at the household and individual levels. The data has a strong emphasis on domains of employment, financial resources, education as well as family and social ties.

The dataset contains annual data on a nationally representative sample of households where each person aged 16 or older are surveyed, where an annual individual interview and self-completed questionnaire is undertaken. Only respondents aged 18 or older were included in this research; this allows our findings to be comparable with other studies in the area (Camelo-Ordaz et al., 2020; Gielnik et al., 2018). Other recent studies have utilised this dataset when focusing on an individual's age (Sacker et al., 2017) and on their earnings (Postel-Vinay and Sepahsalari, 2023).

The dependent variable of this study is the change in the primary employment status of an individual from employed in paid labour to self-employment. The variable takes a value of “1” if an individual becomes self-employed at some point between time period t and time period t+1. The variable takes a value of “0” otherwise. This is similar to the methods adopted by other studies that have used the self-employment decision as the dependent variable in their study (Biehl et al., 2014; Yu and Artz, 2019). As per the dataset, a self-employed individual is an individual who is running a business or a professional practice, a partner in a business or a professional practice, a sub-contractor, doing freelance work or has otherwise defined their employment situation as “working for myself”. It should be noted that only individuals who transitioned from full-time employment to self-employment are included in our analysis.

An individual's pay level from employment prior to entering self-employment is a key independent variable of this study. It is measured as the total net pay amount, which is the take-home pay that an individual received in their last paycheque prior to taking part in the survey ^[1]. This measure is akin to the measurement that Kaiser and Malchow-Møller (2011) applied when they considered the impact of switching from self-employment to paid employment on an individual's pay level.

The other key independent variable is the age of the individual at the time the survey took place. However, to understand how an individual's prior pay impacts self-employment entry for different age cohorts, individuals are categorised into three distinct groups: younger, middle-aged and older individuals. Younger individuals are those aged from 18 to 30 years of age. This is similar to the definition of a young individual used by others in the literature when studying age and self-employment (Minola et al., 2014). Those from 31 to 50 years of age are categorised as “middle-aged” individuals. Other studies in the area have defined “older” as an individual over the age of 50 years who starts a business (Perenyi et al., 2017; Soto-Simeone and Kautonen, 2021). This study uses the same definition, defining “older” self-employed as those aged 51 years of age to the UK pensionable age of 66 years of age.

The descriptive statistics given in Table 1 show some interesting differences between those individuals who became self-employed and those who choose to remain in paid employment.

More middle-aged individuals become self-employed than remain in employment, 58% compared to 55%. Whereas, fewer older individuals become self-employed, 23% versus 27%. There is a marginal difference in the younger cohort, where 19% of the sample were younger individuals who switched from employment to self-employment, whereas 18% were those who remained in employment.

Similarly, there are differences between self-employed individuals and individuals in paid employment in terms of their pay levels. Those who become self-employed have higher pay in the previous time period in comparison with individuals in paid employment. Those who are self-employed earned a mean of £4,779 per month from their paycheque in the time period before becoming self-employed, whereas employed individuals earned considerably less, £2,719. However, there is an increased variance between the two groups. Self-employed individuals have a standard deviation of £11,447, compared to £5,696 for those in paid employment. Thus, while on average those who become self-employed have comparably higher average pay levels before entering self-employment, this is not the case for all.

The objective of the econometric approach is to understand the impact of prior pay levels on an individual's decision to enter self-employment at different life stages. To conduct this, a pooled logit model is estimated. The methods used to test the hypotheses are similar to those used in other studies that investigate the self-employment decision (Biehl et al., 2014; Yu and Artz, 2019).

Firstly, the model is constructed with age as a categorical variable, AgeCat_ijt-1, where three categories exist, 18 to 30, 31 to 50 and 50+ years of age, with 31–50 as the reference category. This is outlined as Equation (1):

SEmp_it is the dependent variable and is the binary self-employment decision for individual i in time period t. Pay_it-1 is an independent variable that measures an individual's income from employed work in time t-1.

To capture the true impact of an individual's age and pay on their decision to enter self-employment, other determinants of self-employment highlighted by the literature are included, such as gender, ethnicity, education level and family situation (Parker, 2018). These controls are denoted by X_it-1. The unobserved time effects within the model are accounted for by δ_t-1, which is a series of time dummies, and the error term of the model is denoted by μ_it.

A second model is then constructed that is identical to Equation (1) but includes an interaction term for the age and pay variables, thus capturing the joint distribution of these variables. This is outlined in Equation (2), which is expressed as:

Where AgeCat_ijt-1*Pay_it-1 are the interaction terms of the age category variables with the pay variable. This will allow for an understanding of the importance of an individual's pay level on the decision to enter self-employment at different life stages, thus allowing for the testing of H1 and H2.

Each explanatory variable is observed in the year prior to potential entry into self-employment, i.e. each explanatory variable is lagged by one period. This accounts for changes in an individual's age and pay to be controlled for within the estimations. A similar strategy was used by Fossen (2021) when analysing the self-employment decision.

Table 2 presents the results of the logit model estimations of Equations (1) and (2). Both models hold explanatory power with 99% confidence, Wald-χ² p-value <0.01. Heteroscedastic and autocorrelative robust standard errors are implemented to reduce any potential bias in the significance of the model's variables.

When age is considered independently from prior pay levels (see Table 2, Column 1), age is not a deciding factor on an individual's choice to enter self-employment. Interestingly, when the interaction term between age and prior pay levels is considered (Table 2, Column 2), the age categories become statistically significant. Consequently, the differences in results between Equations (1) and (2) signal a misspecification issue with Equation (1) captured by the omitted interaction term. A likelihood-ratio test confirms that the extended model specified by Equation (2) has greater explanatory power ^[2]. Thus, the estimates of the extended model in Table 2, Column 2, are discussed below.

From Table 2, Column 2, the younger cohort of individuals aged from 18–30 years of age is more likely to enter self-employment relative to those aged 31–50 years of age (i.e. the middle-aged cohort) as β_1,1 > 0. However, older individuals aged from 51 to 66 are less likely to enter self-employment compared to those who are middle-aged β_1,3<0. Taking both results together, it appears that the propensity to become self-employed decreases as an individual ages. This finding aligns with other studies on entrepreneurial entry that consider an individual's age (Brieger et al., 2021; Zhao et al., 2021).

In Table 2, Column 2, the interaction terms of both age cohorts and prior pay are significant. This implies that the effect of an individual's prior pay on their propensity to enter self-employment is different for the younger age cohort and the older age cohort relative to the middle-aged cohort.

As the coefficient is negative for the younger age cohort, β_3,1 <0, it appears that higher prior pay levels decrease a younger individual's propensity to enter self-employment to a greater extent than that of the higher prior pay middle-aged individuals, which supports H1.

The reverse effect appears to be true for the older cohort, β_3,3 >0, indicating that higher prior pay levels increase an older individual's propensity to enter self-employment to a greater extent than that of a higher-paid middle-aged individual, supporting H2.

Regarding the base category, 31–50 year olds, the results situate them between the older and younger age groups. The estimations show that they are more likely to enter self-employment than the older age cohort, but less likely than the younger age cohort. Similarly, the estimations place the effect of prior pay levels on an individual in the middle-aged cohort between the older and younger age groups. The results, in Table 2, Column 2, show that they are more likely to enter self-employment from higher levels of prior pay than individuals aged 18–30 years old, but less likely to enter from higher levels of prior pay than individuals aged 51–66 years old.

This is shown visually below in Figure 2, where a negative relationship exists between prior pay from employment and the decision to enter self-employment for 18–30 year olds, whereas a positive relationship is observed for 51–66 year olds.

Figure 2 outlines the interaction between an individual's prior pay and their propensity to enter self-employment. Displayed is the changing relationship between an individual's prior pay level and the propensity to enter self-employment for both the younger and older cohorts in our sample relative to middle-aged individuals as prior pay increases.

A negative relationship exists for the younger cohort, showing that as pay levels increase, younger individuals are less likely to become self-employed, relative to middle-aged individuals. The reverse relationship appears for the older cohort. As prior pay increases, their propensity to enter self-employment relative to middle-aged individuals also increases.

The estimates support research that found younger individuals are more likely to become self-employed when they have lower levels of pay prior to engaging in self-employment (Gielnik et al., 2018; Zhao et al., 2021). They also support studies that have discovered that older individuals who become self-employed have an increased propensity when they enter higher-paying employment (Maalaoui et al., 2020). As distinct from these other studies, our findings considered both younger and older individuals and show that a difference in the circumstances exists when they engage in entry into self-employment.

Turning to the control variables of the study. Men are more likely to become self-employed than women. This is in line with other self-employment studies (Brush et al., 2020; Marlow, 2014). However, there is an ongoing discussion within the literature that aims to understand if this is due to males and individuals with male-associated traits being more suited to self-employment, or if the structures currently in place that support self-employment are designed in a manner that encourages males to enter self-employment and discourages females (Laguía et al., 2022). Ethnicity also impacts their propensity to become self-employed, which was also found by other studies (Aldrich and Waldinger, 1990; Kerr and Mandorff, 2023) and is due to many factors, such as increased risk tolerances (Brieger and Gielnik, 2021; Dutta et al., 2021) or opportunity identification (Angulo-Guerrero et al., 2017; Gielnik et al., 2018). The results from the education variable show that individuals with higher levels of education are more likely to enter self-employment. This is interesting, as increased education has been seen to discourage self-employment (Patel, 2024) and only training in certain disciplines, such as management or economics, increases self-employment entry (Piva and Rovelli, 2022). Therefore, this finding implies that the highly educated and the self-employed may share similar traits as opposed to increased education being a driver of self-employment entry. The family situation of an individual also does not impact their likelihood of becoming self-employed. Again, this finding is interesting as the familial situation of an individual has been shown to impact much of the self-employment experience for an individual within the literature (Gabaldon et al., 2015; Obschonka, 2016).

There are a few possible issues that may impact the validity of the results shown, the first of which is bias caused by multicollinearity between the independent variables. To test if this would cause potential problems, the correlation between the pay and age category variables was investigated ^[3]. As there is only minimal correlation between the independent variables, we can assume that this does not cause substantial problems surrounding the significance and magnitude of the coefficients within our results.

The model may also suffer from endogeneity problems, as an individual's decision to enter self-employment and their prior pay from employment may exhibit reverse causality. To overcome this issue, the independent and control variables were observed one period prior to the outcome variable. This is a similar approach to Bellemare et al. (2017) where the lagging of independent variables was recommended to reduce the impact of endogeneity in social science research.

Finally, bias may arise due to the overrepresentation of one age cohort switching to self-employment in our data. To investigate this, a sensitivity analysis was conducted to understand if this occurred ^[4]. Paid employment to self-employment switches appear proportionally similar across age cohorts, thus indicating that no bias due to overrepresentation occurs. The results are also consistent when age-cutoffs are implemented ^[5].

Moreover, additional estimations were made to ensure the validity of the results derived in this study ^[6]. Firstly, age was included as a continuous measure, as opposed to categorical. The results of this exhibit similar outcomes to what is presented above, with a direct negative relationship with self-employment entry but a positive interaction with the earnings variable. This provides support for the presented results, with younger individuals being more likely to enter self-employment from lower-earning backgrounds, while older individuals are more likely to enter from higher-earning situations. Secondly, a household earnings variable is substituted for an individual's pay. When done so, similar results are obtained to what is presented. Measures of savings ^[7] and investments were also used to understand if their inclusion altered the results. Again, the results remain consistent with those presented.

Further non-linearities within the independent variables were explored through additional estimations. Models with squared terms of the age and pay variables were estimated. The results of these indicate that the inclusion of the non-linear independent variable does not change the interpretation provided above. However, due to the increased complexity in interpretation that occurs when multiple non-linear independent variables are included within a model, the results are presented as above to allow for a more straightforward understanding of the role of an individual's age and earning capacity on their propensity to enter self-employment. Similarly, models with pay expressed as deciles were estimated; again, these exhibited comparable results to what is presented above in this article.

Our findings expand the understanding of how the current life stage of an individual impacts their choice to become self-employed. Individuals in the older age cohort were less likely to enter self-employment compared to the middle-aged group. This may be due to reasons highlighted in previous studies, where older individuals have been found to be comparatively less able to identify opportunities in markets and form new businesses due to reduced creative capacities (Gielnik et al., 2018; Weinberger et al., 2018). However, the opposite occurs for younger individuals, who are more likely to become self-employed than middle-aged individuals. This is in line with previous studies that have found that certain incentives towards self-employment are greater for younger people (Brieger et al., 2021; Zhao et al., 2021). Traits such as being more risk accepting (Kraft et al., 2022), more financially driven (Gielnik et al., 2018) and having a greater ability for creative thinking (Hoogendoorn et al., 2022) may be the underlying mechanisms why younger people are more inclined to become self-employed.

In addition, it contributes to the body of literature on the role of age and earning capacity in becoming self-employed by considering both factors in conjunction with one another (Jiao et al., 2021; Zhang and Acs, 2018). This analysis adds to the literature by displaying that an individual's propensity to enter self-employment can be viewed when age and earning capacity are considered jointly rather than as two separate effects. This has several implications for the current understanding of how individuals make the self-employment entry decision.

In the case of younger individuals, the findings suggest that this cohort values the payments earned from their current employment situation more than the potential earnings they may receive from being self-employed. This finding does not align with other studies that have found that younger individuals value the potential earnings from self-employment over their current earnings from employment (Gielnik et al., 2018; Zhao et al., 2021). Taken further, this indicates that younger individuals may be more inclined to become self-employed out of necessity. This is different for older individuals, as the results suggest that higher prior pay levels from employment incentivise them to enter self-employment. This indicates that individuals in the older age group may engage in self-employment for reasons beyond monetary gain. This complements other findings that have found that older individuals value other aspects of self-employment than profits (Perenyi et al., 2017; Soto-Simeone and Kautonen, 2021).

Middle-aged individuals fall between their younger and older counterparts in how prior employment earnings influence their propensity to enter self-employment. The middle-aged cohort may engage in both necessity self-employment entry (Block et al., 2015b; Kim and Rizzi, 2020) and opportunity-driven entry (Lougui and Broström, 2021). Accordingly, this age group appears more heterogeneous in both the pathways into self-employment and the associated opportunity costs of entry (Audretsch et al., 2022; Lougui and Broström, 2021).

When findings from the literature are considered, labour market constraints are likely to manifest differently across age groups (Darougheh, 2026; Kim and Rizzi, 2020). Middle-aged individuals are more likely to face caring for older family members (Kim and Rizzi, 2020; Oladipo et al., 2023) and dependent children (Gabaldon et al., 2015). Therefore, middle-aged individuals may be pushed into self-employment, as labour market employment cannot accommodate their needs, whereas self-employment can due to the increased flexibility it enables (Van Der Zwan et al., 2020).

At the same time, previous studies indicate that opportunity-driven self-employment is also conceivable among the middle-aged cohort. The opportunity cost of entry may be reduced through “spousal insurance” (Darougheh, 2026; Stephens, 2002), where the partner's earnings provide stable support through the initial stages of self-employment. Additionally, a more rounded skillset developed by several years in employment and accompanying industry knowledge, both attributes contribute to encouraging entry (Dutta et al., 2021; Lazear, 2005). Together, these factors may increase engagement in opportunity-driven self-employment among middle-aged individuals whose household wealth and skillset enable such transitions. Accordingly, self-employment entry within this cohort is highly contingent on the circumstances in which a middle-aged individual is. Compared to older and younger individuals, middle-aged individuals' entry decision may therefore be influenced by their own employment situation, but also the financial and interdependencies of their family unit. Policy Implications.

The results provide valuable insights for policymakers and for other parties who aim to improve self-employment rates. One key insight for policymakers relates to how younger individuals with lower levels of prior pay are more likely to become self-employed. It was hypothesised that this is due to how lower risk perception and monetary gain influence younger people to start their own businesses (Kraft et al., 2022; Zhao and Yang, 2021). This may signal that targeting financial assistance towards low-paid younger individuals for self-employment projects could encourage these individuals to enter self-employment and support them in the early stages of firm development, while maintaining the underlying incentives of high-risk propensity and desire for monetary gain.

This is useful as it has been highlighted that the self-employment rates of younger individuals should be higher in developed countries (OECD/European Commission, 2023). Understanding that directly targeting policy towards incentivising younger individuals from lower-paid employment to become self-employed may be an effective strategy. This should be encouraged by policymakers due to the potential increases in innovation within an economy. Given that younger individuals are often more adept at identifying market opportunities and possess greater creative capacities (Gielnik et al., 2012; Hoogendoorn et al., 2022), this could lead to greater innovation and development within the economy (Esteve-Pérez et al., 2018; Nakara et al., 2021). As such, encouraging youth self-employment is potentially an avenue for sustained economic growth (Audretsch et al., 2022; Lenihan et al., 2019). The findings of this study show that younger individuals are more likely to enter self-employment from relatively weaker earnings positions and for necessity-driven reasons (Audretsch et al., 2022). This suggests that policy should offer support for younger individuals to assist them in self-employment and avoid firm failure in the early stages. This would aid in reducing avoidable exit rates of nascent firms and encourage economic growth. This could be achieved through sustained targeted policy, which provides financial assistance and also through other supports such as networking opportunities, skills development and administrative support (OECD/European Commission, 2023). Financial support would provide assistance with the development of initial capital needed to overcome early-stage challenges (Fubah et al., 2025; Hoogendoorn et al., 2022), while the other supports highlighted will provide the younger self-employment individual with greater access to markets and a larger customer base. Crafting targeted policy such as this would not only encourage a higher level of younger individuals to become self-employed but also provide support and reduce exit rates.

Moreover, policymakers would see benefits from alleviating younger individuals from the circumstances that cause them to enter self-employment for necessity reasons. As necessity self-employment is less productive and has reduced spillover benefits on the economy than opportunity self-employment (Audretsch et al., 2022), if younger individuals were afforded the opportunity to enter self-employment in better economic circumstances, they would likely establish more productive firms. Therefore, policy could be crafted to provide younger individuals with better employment prospects, such as employment fairs or training courses, which would allow them to accumulate capital and develop their firm concepts and strategy prior to entering self-employment (Hoogendoorn et al., 2022; Williams, 2004).

In contrast, financial aid measures are potentially less effective at enticing older individuals into self-employment than for younger individuals. Policymakers could provide other incentives, however, such as providing assistance in developing networks and focusing on developing the entrepreneurial skills of older individuals, which are two areas that older individuals have found to be difficult when establishing their business (Maalaoui et al., 2020; Weinberger et al., 2018). In addition, a more holistic enterprise policy that targets social contributions and sustainable development is an avenue that is not widely considered, but should be pursued by policymakers to encourage self-employment rates (Lenihan, 2011). This could be taken a step further, and policy could encourage the pro-activity for innovative thinking of older workers in employed work, as this has been seen as a way to increase self-employment rates and tackle a key development area for older individuals with potential to become self-employed (Lenihan et al., 2019; Weinberger et al., 2018).

Our findings do call into question the effectiveness of active aging policies, such as “Retaining Talent at All Ages” (OECD, 2023), which have been used by many developed nations in recent years to encourage older workers to remain in the labour force through becoming self-employed. This study provides evidence that these measures may not be effective in increasing self-employment rates of older workers, despite this cohort being identified as a section of the population where there is potential for increased self-employment entry rates (OECD/European Commission, 2023). Failing to facilitate the entry of older individuals into self-employment has been identified as a shortcoming of developed nations (Kupiainen et al., 2023). Doing so could provide greater employment opportunities within the economy, because self-employed individuals who are older are more likely to hire employees compared to younger self-employed individuals (Carter, 2011; Cowling and Wooden, 2021). Therefore, providing increased incentives for older individuals to become self-employed will possibly have positive impacts on labour market conditions and employment opportunities.

This study has highlighted that entry into self-employment is not a homogenous choice across an individual's life and that policymakers may benefit from utilising policy tools in the context of an individual's age. Through consideration of the age of an individual when entering self-employment, as opposed to treating all age groups as the same, particular sections of the population, such as older individuals, may see increased self-employment rates.

When making the decision to enter self-employment, a prospective individual should consider what life stage they are currently in and what role this can play in their decision.

For younger individuals, becoming self-employed has a variety of benefits for them over remaining in paid employment. As they are younger, they have two main advantages over their older counterparts: having more time to reap the benefits of self-employment and it can offer them a path out of lower income when they cannot find it through paid employment.

As younger individuals have more forthcoming years to spend in self-employment, they may see increased financial gains from their business due to its longer life cycle (Gielnik et al., 2012, 2018). However, if the business is not financially viable and the business has to be closed, the individual still has time to re-enter paid employment, as they are still seen as employable in the workforce (Perenyi et al., 2017). Also, as observed in the results, younger individuals with lower earning capacity being more likely to become self-employed. This indicates that self-employment can act as an avenue for younger individuals who are unable to find suitably high-paying jobs in employment to find comparative income in starting their own businesses.

However, the conditions and reasons that older individuals enter self-employment are more related to it offering a quasi-retirement (Perenyi et al., 2017; Soto-Simeone and Kautonen, 2021). If older individuals have the means, coming from high-paying employment backgrounds, and the desire to swap employment for self-employment, it can offer increased quality of life while also maintaining a certain level of financial stability (Kautonen et al., 2017; Soto-Simeone and Kautonen, 2021).

While this study showed that the life-stage of an individual does play a role in their likelihood of entering self-employment, and that their earning capacity from paid employment influenced younger and older individuals differently, the underlying mechanisms of this were not tested for. Moreover, the lack of data availability precluded the inclusion of innate personal traits of individuals, such as risk tolerance or ambition. It should be noted that these traits are likely to impact their salary from employment and their propensity to enter self-employment (Block et al., 2015a; Burchell and Coutts, 2019). In the theoretical framework, it was highlighted that younger individuals are more attracted to self-employment for financial gain reasons, whereas older individuals enter for quality-of-life related reasons (Gielnik et al., 2012; Soto-Simeone and Kautonen, 2021). Future research should investigate these underlying mechanisms and more closely investigate their relationship with an individual's life stage.

Naturally, this study also faced other data limitation issues. One such limitation is the definition of self-employment used within the UKHLS and the conclusions that can be drawn from the findings as a result. The definition encompasses a wide variety of self-employment types, such as individuals who are running professional practices, sub-contractors and freelance workers. As such, the results should be viewed with this in mind. However, this does open future research agendas to understand if the findings of this study hold for different types of self-employment.

Moreover, while this study focuses on individual earning capacity to understand self-employment entry, personal savings and assets are also likely to influence the decision to become self-employed (Lim, 2018; Robb and Robinson, 2014). While measures of savings and assets were used as robustness checks, the UKHLS lacks the detailed financial information necessary to examine this relationship comprehensively. Future research would therefore benefit from more comprehensive data to explore how wealth and asset accumulation share self-employment entry across different age groups.

Furthermore, an individual's prior pay from employment is one among many meditating factors that have the potential to impact an individual's decision to enter paid employment to varying extents at different life stages. Within the findings, some interesting future research avenues could be found in the control variables. As an individual's gender, among other factors, was seen to impact their propensity to enter self-employment (Marlow, 2014; Weiss et al., 2023). Future investigations could be done to see if this varies throughout an individual's lifespan.

Our research adds to the literature on entry into self-employment at different life stages and provides unique insights into how the influence of an individual's prior pay from employment changes as they age. The findings of this article show that younger individuals are more likely to enter self-employment when they come from low-paying employment relative to middle-aged individuals, whereas the reverse is true for older individuals. This is due to the differing reasons that individuals in these age cohorts become self-employed (Gielnik et al., 2012; Soto-Simeone and Kautonen, 2021).

This responds to two requests that have been made in the literature. An individual's financial capabilities impacted their decision to become self-employed, which was investigated in this article, which was requested by Jiao et al. (2021). This article also analysed the conditions under which an individual enters self-employment while considering all age cohorts, as opposed to only focusing on a particular age cohort. This responded to Zhao et al. (2021) through a closer investigation into how mediating factors, such as an individual's earning capacity, impact the self-employment decision across an individual's life.

The supplementary material for this article can be found online.