This chapter examines whether individuals who make greater use of their skills at work are rewarded differently in the labour market. The analysis focusses on two dimensions. First, the relationship between skills use and wages is assessed in order to identify whether more intensive deployment of skills is associated with higher wages. Second, the potential spillovers onto job satisfaction and the risk of burnout are investigated, recognising that the frequency of skills use may shape not only monetary returns but also wider dimensions of worker well-being.
How Workers Use, or Don’t Use, their Skills in the Workplace
Abstract
More frequent skills use at work is associated with better economic and non-economic outcomes
Higher frequency of skills use is associated with higher wages and the relationship is statistically significant even after accounting for individual and job characteristics. Greater use of influencing skills and information-processing skills such as ICT and numeracy is linked to the highest wages, while physical skills are negatively linked to wages.
The magnitude of the association between wages and skills use differs substantially across countries, suggesting that it might be influenced by institutional arrangements, occupational patterns and the structure of the demand for skills.
Although women earn lower wages than men with comparable characteristics, for most skills, the returns to use do not differ significantly by gender. This suggests that, once men and women are sorted in jobs that make frequent use of one’s skills, they typically receive comparable rewards.
The association between wages and skills use varies significantly by age, suggesting a relationship between skills use, career progression and seniority wages. This highlights the importance of a life‑cycle perspective when assessing the value of skills in the labour market.
Countries with more unequal distribution of skills use tend to exhibit higher wage inequality. By contrast, greater use of skills is positively correlated with labour productivity, although other unobserved drivers – such as industry structure and institutional quality – may play a larger role.
Deployment of workers’ skills generates intrinsic motivation, with greater skills use linked to higher levels of job and life satisfaction. Use of skills also mitigates job burnout, except in the case of physical skills, whose intensive use is associated with increased risk.
The analysis of the returns to skills use at work provides a valuable perspective on the functioning of labour markets and the allocation of human capital. While much research has focussed on the returns to education and formal qualifications (see Psacharopoulos and Patrinos (2018[1]) and Gunderson and Oreopolous (2020[2]) for recent reviews of the literature), the actual utilisation of skills in the workplace may yield more direct insights into the link between skills, wages and productivity. The mere accumulation of skills does not guarantee economic returns if these are underutilised or misallocated relative to job tasks. An analysis of the relationship between skills use and wages, hence, provides a more precise measure of the channels through which individual skills translate into wages, productivity, and subjective well-being.
The expected relationship between skills use and wages rests on the notion that workers who deploy their skills more frequently at work contribute more significantly to outputs and are therefore compensated accordingly. The productivity channel is central, since firms that enable employees to apply their skills would achieve a better match between skills and job tasks and more experimentation. At the same time, skills use is likely to influence job satisfaction and lower turnover, as workers tend to derive greater intrinsic motivation and engagement when their skills are effectively deployed. This may further reinforce productivity outcomes through reduced turnover and stronger commitment to organisational objectives.
Examining these outcomes is relevant for policy and practice. From a policy perspective, understanding the returns to skills use can inform education and training strategies, particularly in relation to skill mismatch and underutilisation of human capital. From a firm-level perspective, it highlights the importance of work organisation and management practices in fostering both performance and employee well-being.
More intensive use of skills in the workplace is associated with higher wages and the association is statistically significant (Figure 2.1). Even after controlling for individual and job characteristics, workers who deploy skills more frequently in their daily tasks at work receive higher wages (see Box 2.1 to examine how skills use and skills proficiency are associated differently to wages).1 This association holds for all skill categories except dexterity and physical skills – which is in line with the findings of previous studies, such as De La Rica, Gortazar and Lewandowski (2020[3]) and Agasisti, Johnes and Paccagnella (2021[4]). This could be partially due to the fact that jobs that rely heavily on physical or manual skills tend to be concentrated in specific sectors with relatively low productivity growth and limited opportunities for wage progression, in contrast to knowledge‑intensive sectors (Acemoglu and Autor, 2011[5]).
The largest association between skills use and wages is observed for influencing skills and information-processing skills such as reading, ICT and numeracy.2 For instance, across countries in the PIAAC sample, a one‑standard-deviation increase in the use of influencing skills at work is linked to a 6% increase in hourly wages (everything else, including occupation, being equal). In contrast, the frequency of learning at work or collaborating with colleagues is associated with only a 2% and 1% increase, respectively.
Association of a one‑standard-deviation increase in the use of each skill and hourly wages, OLS coefficients
Note: The Figure represents the coefficients of OLS regressions of skills use on wages, weighted by sampling weights. Each bar represents a separate regression. Wages are gross hourly earnings for employed and self-employed individuals, including bonuses, in PPP-adjusted 2022 USD. The wage distribution was trimmed to eliminate the 1st and 99th percentiles. The sample includes only full-time workers aged 25‑65. Coefficients are adjusted for gender, immigration background, age, age squared, whether one lives with a partner or has children, educational attainment, literacy proficiency, occupation, industry, firm size, permanent contract, public employment, and country fixed effects. All coefficients are statistically significant (at least at the 5% significance level). Korea is not included in the analysis due to unusual response patterns on wages.
Source: 2023 Survey of Adult Skills.
To examine whether proficiency in literacy and numeracy affects wages more than the actual use of these skills in the workplace, a Mincer-style wage regression can be estimated (similar to that of Kawaguchi and Toriyabe (2022[6])):
Equation 2.1.
where is the natural logarithm of gross hourly earnings for individual i (including bonuses, in PPP-adjusted 2022 USD), is the test score of individual i, is the index of skills use at work, is a vector of controls (including gender, immigration background, age, age squared, whether one lives with a partner or has children, educational attainment, literacy proficiency, occupation, industry, firm size, permanent contract, public employment), and is the error term. To make coefficients directly comparable, wages are regressed on one standard deviation increase in both and .1 To check whether the effects of proficiency and use are statistically different from each other, a Wald test of the following linear hypothesis is then performed:
Equation 2.2.
Figure 2.2 shows that proficiency has a consistently stronger impact on wages than use. For example, a one‑standard-deviation increase in literacy proficiency is associated with an average wage increase of 7%, compared with a 5% increase for reading use at work. According to the Wald test, these coefficients differ significantly, which implies that proficiency and use make distinct contributions to wages. A similar pattern emerges for writing, where proficiency carries a larger premium than use. The divergence is most pronounced for numeracy: proficiency in numeracy yields an 8.5% wage premium, double the 4.3% associated with numeracy use at work.
Several explanations may help account for this gap. First, proficiency reflects the underlying stock of cognitive skills that individuals can draw upon across a wide range of tasks and contexts, including those not explicitly captured by self-reported skills use. Employers may reward proficiency more strongly because it signals adaptability and learning potential, which are critical in dynamic labour markets. Second, use at work is partly determined by job design and task allocation, which may not fully exploit an individual’s abilities. Workers with high proficiency but limited opportunities to apply these skills may still secure higher wages, particularly if proficiency is recognised by credentials or through occupational sorting. Finally, wage‑setting institutions may attach higher value to proficiency because it aligns with formal qualifications and educational attainment, which are often central to pay scales and career progression.
Association between a one‑standard-deviation increase and hourly wages, OLS coefficients
Note:
1. Standardising both proficiency and skill-use variables allows coefficients to be expressed on a comparable scale, but this approach has limitations, as the standard deviation of a frequency-based skill-use index may not represent an equivalent “distance” or effort as a one‑standard-deviation increase in proficiency scores. Hence, while standardisation facilitates comparison, it should not be interpreted as implying equivalence in what it means to move one standard deviation along each dimension.
The Figure represents the coefficient of OLS regressions of skills use and proficiency on wages, weighted by sampling weights. Wages are gross hourly earnings for employed and self-employed individuals, including bonuses, in PPP-adjusted 2022 USD. The wage distribution was trimmed to eliminate the 1st and 99th percentiles. The sample includes only full-time workers aged 25‑65. One model is estimated for each skill, with the corresponding skills use and proficiency as independent variables (literacy scores for reading and writing use at work, and numeracy scores for numeracy use at work). Coefficients are adjusted for gender, immigration background, age, age squared, whether one lives with a partner or has children, educational attainment, occupation, industry, firm size, permanent contract, public employment, and country fixed effects. All coefficients are statistically significant (at least at the 10% significance level). Korea is not included in the analysis due to unusual response patterns on wages.
Source: 2023 Survey of Adult Skills.
Looking only at cross-country averages, however, masks substantial variation in the wage returns to skills use. Figure 2.3 illustrates this point by comparing the association between wages and the use of those skills that on average show the largest and smallest association with wages – i.e. influencing skills and physical skills – across countries. The relationship between the use of influencing skills and wages is positive everywhere, but their magnitude differs sharply. In Spain, workers who use influencing skills more intensively earn wages that are 10% higher, while in France the difference in wages is just over 2%. Such variation suggests that institutional arrangements and the demand for influencing skills differ significantly between labour markets.
In contrast, returns to physical skills are consistently negative throughout the sample, but again there are notable cross-country differences. In New Zealand and Spain, intensive use of physical skills is associated with an earnings penalty of more than 9%, whereas in Lithuania and Italy the effect is close to zero and not statistically significant. These results likely reflect technological intensity and wage‑setting systems. In countries where manufacturing, agriculture or manual services still represent a significant share of employment, the wage penalty seems to be less pronounced.
Association of a one‑standard-deviation increase in the use of each skill and hourly wages, OLS coefficients
Note: The Figure represents the coefficients of OLS regressions of skills use on wages, weighted by sampling weights. Each bar represents a separate regression. Wages are gross hourly earnings for employed and self-employed individuals, including bonuses, in PPP-adjusted 2022 USD. The wage distribution was trimmed to eliminate the 1st and 99th percentiles. The sample includes only full-time workers aged 25‑65. Coefficients are adjusted for gender, immigration background, age, age squared, whether one lives with a partner or has children, educational attainment, literacy proficiency, occupation, industry, firm size, permanent contract, and public employment. Bars with a checkered pattern are not statistically significant (at the 10% significance level).
Source: 2023 Survey of Adult Skills.
Returns to skills use may also vary between socio-demographic groups. Understanding whether the association between skills use and wages differs between groups is important for at least two reasons. First, it sheds light on whether skills are valued differently in the labour market depending on who uses them, which speaks directly to issues of equity and discrimination. Second, it allows for a better understanding of the mechanisms underlying the pay gaps. For instance, if women receive lower returns to the same skills use levels, the gap cannot be explained solely by differences in skills use, skill proficiency or occupational choices, but also by how labour markets reward skills use.
Table 2.1 presents estimates of wage returns to skills use, and includes an interaction term that indicates whether returns differ for men and women. The results show that most skills (except manual and physical ones) are associated with higher wages when used more frequently (column 1). The interaction term in column 3 is generally not statistically significant for many skills, which suggests that, in most cases, men and women obtain similar returns once they use a given skill with the same frequency. Using PIAAC data, Battisti, Fedorets and Kinne (2023[7]) find similar non-significant results when looking at the impact of the interaction term between a female dummy and numeracy proficiency on wages.
However, Table 2.1 also shows a few important exceptions. For learning at work and self-organising skills, the interaction coefficients are positive and significant, indicating that women receive higher wage returns than men for the use of these skills at work. Women who report no use of self-organising skills at work earn on average 14% less than men. Yet when these skills are more intensively utilised, the wage gradient is steeper for women. A one‑standard-deviation increase in their use corresponds to 2% higher wages for men and 5% higher wages for women, implying that greater opportunities for self-management could reduce or even reverse the gender pay gap.
The overall message is therefore twofold. On the one hand, women face a systematic wage disadvantage across the board, as shown by the large and negative baseline coefficients for women in column 2 (which is in line with much of the literature exploiting PIAAC data to estimate the gender wage gap, such as Christl and Köppl – Turyna (2020[8]), Rebollo-Sanz and De la Rica (2022[9]) and Komatsu (2023[10])). On the other hand, the interaction terms show that for a limited set of skills uses – particularly learning at work and self-organisation – the returns are somewhat more favourable to women compared to men. These patterns may reflect the fact that either women have specific non-observable characteristics that are rewarded differently, or employers perceive that skills use by women in these areas is especially valuable. It may also reflect occupational segregation: women who succeed in highly autonomous jobs may occupy positions where they are positively discriminated, resulting in higher relative rewards.
Association between a one‑standard-deviation increase in the use of each skill and hourly wages, OLS coefficients
Note: Each row represents a separate regression. Controls are not shown, but include immigration background, age, age squared, whether one lives with a partner or has children, educational attainment, literacy proficiency, occupation, industry, firm size, permanent contract, public employment, and country fixed effects. The interaction term highlights whether part of the wage gap is due to differences in how skills use is valued for men versus women. If the interaction coefficient is significant and negative (positive), it suggests that women earn less (more) than men for the same level of skills use, while if it is not significant, the return to skills use is the same for both genders. Korea is not included in the analysis due to unusual response patterns on wages.
Source: 2023 Survey of Adult Skills.
Similarly, analysing the association of wages with skills use by age group provides important insights into how skills use is valued at different stages of a worker’s career. Age not only reflects accumulated work experience but also shapes occupational positions, task allocations and opportunities to apply skills. Younger workers may be in jobs that reward skills use differently than older workers, either because they are still in career-building stages, because they are concentrated in entry-level positions with limited bargaining power, or because employers perceive their skills use as a signal of potential. Younger and older workers may be in occupations where skills use is rewarded differently depending on whether it complements or substitutes for experience, and whether it is aligned with evolving technological and organisational demands. Disaggregating the relationship by age therefore helps identify whether labour markets reward skills use consistently throughout the life cycle or whether wages and skills use interacts with career trajectories and institutional arrangements in distinctive ways.
Figure 2.4 presents the association between wages and skills use separately for three age groups. Two patterns stand out. First, across several skill domains (and especially for influencing and reading skills), younger workers aged 25 to 34 receive lower wage premia for skills use than mid-career and older workers. This is consistent with labour markets in which responsibility and pay rise with tenure, so the same skill used by a junior employee affects less strategic decisions and generates less “measurable” value. Selection and survivorship mechanisms could reinforce the pattern, since those who remain in roles that intensively use these skills may tend to be the workers for whom the market pays higher premia.
Second, older workers either see the highest or the lowest wage association with skills use depending on the skill considered. For instance, influencing skills, task discretion and self-organising skills are associated with the highest wages for workers aged 55 to 65. By contrast, for skills such as learning at work, co‑operative skills and dexterity, older workers receive the lowest premia, with coefficients that are negative. This polarisation may reflect the dual role of age in the labour market. On the one hand, experience and tenure may amplify the value of certain high-level skills that build on authority and autonomy. On the other hand, older workers may face occupational constraints in domains where physical skills are critical, which reduce or even reverse the wage premium associated with using these skills.
Association between a one‑standard-deviation increase in the use of each skill and hourly wages, OLS coefficients
Note: The Figure represents the coefficients of OLS regressions of skills use on wages, weighted by sampling weights. Each marker represents a separate regression. Wages are gross hourly earnings for employed and self-employed individuals, including bonuses, in PPP-adjusted 2022 USD. The wage distribution was trimmed to eliminate the 1st and 99th percentiles. The sample includes only full-time workers aged 25‑65. Coefficients are adjusted for gender, immigration background, whether one lives with a partner or has children, educational attainment, literacy proficiency, occupation, industry, firm size, permanent contract, public employment, and country fixed effects. Korea is not included in the analysis due to unusual response patterns on wages.
Source: 2023 Survey of Adult Skills.
The evidence so far suggests that inequalities in how workers are able to apply their skills on the job reinforce disparities in pay arising from different skill levels, as workers with greater opportunities to deploy skills may command wage premiums, while those in roles with limited skills use remain in lower pay brackets. Unequal skills use may also reflect discriminatory managerial practices, which can exacerbate inequalities in pay.
Understanding the broader relationship between skills and wage inequality has been a central theme in recent labour economics, but the debate has largely focussed on proficiency rather than the use of skills at work. For instance, using data from the first cycle of PIAAC, Paccagnella (2015[11]) finds that while proficiency in cognitive skills is associated with positive wage returns, the correlation between the dispersion of proficiency and the dispersion of wages across countries is weak, and in some cases even negative. This perspective is reinforced by Broecke, Quintini and Vandeweyer (2019[12]), which emphasised how differences in wage inequality across countries are driven primarily by differences in the return to skills and by how well the supply of skills meets the demand.
However, the existing literature has not fully leveraged the information on how skills are used in the workplace, even though this margin may be critical in shaping how proficiency translates into wage outcomes. Indeed, skills use provides a way of observing the match between workers’ capacities and the frequency of their use at work, which mediates the transmission of proficiency into wages. Only OECD (2015[13]) exploited information on skills use at work to simulate the impact that improving numeracy skills use would have on wage inequality. They find that in most of the countries included in the first cycle of the Survey of Adult Skills, wage inequality measured by the Gini coefficient would indeed decline if numeracy skills were more frequently used in the labour market.
Exploiting the new cycle of PIAAC data, Figure 2.5 plots the relationship between wage inequality and inequality in the use of reading skills at work, using the gap between the 90th and 10th percentiles of both distributions.3 The positive slope of the fitted regression suggests that countries where the use of reading skills is more unequally distributed among workers also tend to display higher levels of wage inequality. The explanatory power of this relationship is non-trivial, with an R-squared of 0.41, suggesting that about 40% of the cross-country variation in wage inequality can be accounted for by inequality in skills use.4 While the relationship is purely descriptive and does not account for other structural factors that may also influence wage dispersion, the findings still underscore the importance of considering not only the availability of skills in the workforce but also how they are used at work as a determinant of wage dispersion.
Nordic countries such as Norway, Sweden, Denmark, and Finland cluster in the lower-left corner of the graph, combining both relatively equal skills use and low wage inequality. This reflects their institutional settings, with compressed wage structures (Mogstad, Salvanes and Torsvik, 2025[14]) and workplace practices that distribute skill-intensive tasks more evenly between similar workers and jobs. At the other extreme, countries like Spain and Chile appear in the upper-right corner, where both skill-use inequality and wage inequality are high, consistent with labour markets characterised by dual structures and segmented opportunities for high- versus low-skill tasks. Anglo-Saxon countries such as the United States, Canada and the United Kingdom fall in the mid-range of both wage inequality and skills-use inequality, suggesting that other unobserved institutional factors – such as weak collective bargaining or high returns to top-end occupations – are also at play.
Note: Wage inequality is measured as the difference between the top and the bottom deciles of the (log) wage distribution. Wages are gross hourly earnings for employed and self-employed individuals, including bonuses, in PPP-adjusted 2022 USD. The wage distribution was trimmed to eliminate the 1st and 99th percentiles. Similarly, inequality in the use of reading skills at work is calculated as the difference between the top and bottom deciles of the skills use distribution. The sample includes only full-time workers aged 25‑65. For consistency, the country name is reported for Belgium and the United Kingdom, even if the Survey of Adult Skills is conducted only at subnational level – namely in the Flemish region and in England, respectively. Korea is not included in the analysis due to unusual response patterns on wages.
Source: 2023 Survey of Adult Skills.
The efficiency with which skills are used in different jobs and firms is also associated with cross-country differences in labour productivity. Adalet McGowan and Andrews (2017[15]) show that when the skills workers possess are misaligned with the requirements of their jobs – whether through over-skilling or under-qualification – aggregate labour productivity is lower. Their cross-country, industry-level estimates attribute the drag primarily to two margins: a loss of allocative efficiency as high-productivity firms are less able to attract appropriately skilled workers (a between-firm reallocation failure), and a within-firm shortfall where under-qualified workers depress establishment-level productivity. Exploiting 2023 Survey of Adult Skills data, Andrews, Égert and de la Maisonneuve (2025[16]) document a strong positive association between the average level of adult skills and industry labour productivity across countries. Crucially, they also show that allocation matters independently of average skill: industries with lower labour-market mismatch and with a higher propensity for high-skilled workers to be employed in larger, more dynamic firms exhibit higher productivity.
Figure 2.6 illustrates the cross-country relationship between labour productivity and the frequency of reading skills at work. The positive slope of the regression line indicates that economies where workers more frequently use reading skills at work tend to achieve higher levels of GDP per hour worked, with the estimated correlation (R² ≈ 0.31) suggesting that skills use explains a non-trivial share of the variation in productivity across OECD countries.5 High-productivity economies such as Ireland, Norway, Switzerland and the United States combine above‑average productivity with high reported use of reading skills. By contrast, countries in Central and Eastern Europe (e.g. Hungary, Lithuania, Croatia) cluster at the lower end of both productivity and skills use, reinforcing the view that insufficient utilisation of available skills constrains aggregate performance. Some outliers such as New Zealand – high reading skills use but relatively modest productivity – or Chile – low productivity despite average skills use – highlight that while the skills-use channel is important, other structural factors including capital deepening, industry mix and institutional quality are at play.
Note: Labour productivity is expressed as the natural logarithm of GDP per hour worked in PPP-adjusted 2020 USD. For consistency, the country name is reported for Belgium and the United Kingdom, even if the Survey of Adult Skills is conducted only at subnational level – namely in the Flemish region and in England, respectively.
Source: 2023 Survey of Adult Skills; OECD Productivity Database (2023).
At its core, skills use captures whether job design and production technologies actually allow workers to use the skills they possess – what scholars often call skill discretion or skill utilisation (Fujishiro and Heaney, 2017[17]). When workers are placed in roles where their skills are better used, the job generates intrinsic motivation and satisfaction; when they are not, the job becomes a source of disutility, increasing the risk of turnover (Dopeso-Fernández, Giusti and Kucel, 2023[18]).6 For example, Boxall, Hutchison, and Wassenaar (2015[19]) show that high-involvement work processes (discretion, participation, developmental practices) improve employee outcomes not merely through direct effects but because they raise skills use and intrinsic motivation, thereby improving job satisfaction and well-being. The returns to skills use seem to extend even beyond the workplace. Fujishiro and Heaney (2017[17]) document that higher skills use – operationalised as “doing what I do best” – is associated with better self-rated physical health, with part of the effect mediated by healthier behaviours.
Figure 2.7 shows the marginal probability of being extremely satisfied at work as a function of skills use, using data from the 2023 Survey of Adult Skills. Coefficients are adjusted for gender, immigration background, whether one lives with a partner or has children, educational attainment, literacy proficiency, occupation, industry, firm size, permanent contract, public employment, wages and country fixed effects. A consistent pattern emerges across most domains: higher frequency of skills use – whether in information-processing or generic skills – is positively and significantly associated with the probability of reporting the highest level of job satisfaction. The positive relationships are not trivial in magnitude. For example, the predicted probability of being extremely satisfied almost doubles between the lowest and highest levels of use of learning-at-work skills, underscoring the importance of learning opportunities and continuous skill engagement as key determinants of job satisfaction.7 The use of task discretion at work is also strongly associated with job satisfaction. This should come as no surprise since task discretion connects directly to the dimension of autonomy, a core component of self-determination theory: having control over how tasks are performed allows workers to align work processes with their own problem-solving styles, pace, and preferences, effectively increasing the “consumption value” of the job (Hackman and Oldham, 1976[20]). It also signals trust from employers, reducing perceived monitoring costs and fostering organisational commitment.
By contrast, dexterity use at work exhibits a notably flat relationship with job satisfaction, suggesting that manual precision tasks – unlike cognitive or autonomy-related skills – do not generate the same intrinsic motivational returns. The use of physical skills displays a slightly negative slope, indicating that jobs requiring more physically demanding tasks are associated with lower likelihoods of being extremely satisfied, all else (including wages) equal. This pattern is consistent with the idea that, while greater skills use generally entails effort costs, those associated with cognitive skills tend to be offset by intrinsic rewards (such as autonomy or pro-social preferences), whereas the disutility from physically strenuous work appears less compensated. Box 2.2 shows similar results for the relationship between skills use at work and life satisfaction.
Marginal probability change of being extremely satisfied at work
Note: The Figure displays the predicted probability of being extremely satisfied at work (job satisfaction equal to 5 in a 1‑5 scale) by skills use in the workplace, based on a weighted regression model. Each graph represents a separate regression. The sample includes only full-time workers aged 25‑65. Coefficients are adjusted for gender, immigration background, whether one lives with a partner or has children, educational attainment, literacy proficiency, occupation, industry, firm size, permanent contract, public employment, wages and country fixed effects. All coefficients are statistically significant (at the 1% significance level).
Source: 2023 Survey of Adult Skills.
Although the relationship between the deployment of skills at work and overall life satisfaction is unlikely to be entirely direct, as it may operate through intermediate channels such as job satisfaction, it remains valuable to investigate whether skills use is correlated with life satisfaction, since this would suggest that the effects of workplace experiences extend beyond the professional domain. The application of skills can reinforce a sense of competence, personal development and autonomy, all of which are dimensions that contribute to subjective well-being. If individuals derive fulfilment from exercising and expanding their capabilities, the benefits may spill over into their broader quality of life, shaping perceptions of purpose and achievement outside the workplace.
Figure 2.8 illustrates the estimated marginal effects of skills use at work on the probability of reporting extremely high life satisfaction. Similarly to the previous analysis on job satisfaction, the results indicate that greater use of skills in the workplace is associated with higher life satisfaction across almost all domains. These findings are robust to adjustments for gender, immigration background, family composition, education, wages, occupation, industry and employment conditions, which implies that the associations are not merely capturing compositional differences in the workforce. Taken together with the findings on job satisfaction, the evidence highlights that not all skills are equally relevant for job and life satisfaction, and that opportunities for cognitive challenge and decision making at work may be especially valuable for well-being outcomes.
Marginal probability change of being extremely satisfied in life
Note: The Figure displays the predicted probability of being extremely satisfied in life (life satisfaction equal to 9 or 10 in a 1‑10 scale) by skills use in the workplace, based on a weighted regression model. Each graph represents a separate regression. The sample includes only full-time workers aged 25‑65. Coefficients are adjusted for gender, immigration background, whether one lives with a partner or has children, educational attainment, literacy proficiency, occupation, industry, firm size, permanent contract, public employment, wages and country fixed effects. All coefficients are statistically significant (at the 1% significance level).
Source: 2023 Survey of Adult Skills.
Another risk to job quality and worker well-being that has so far been overlooked in the economic literature on the returns to skills and skills use is job burnout. Burnout represents a critical outcome to consider, as it reflects the potential costs of skills use when demands become excessive or when workers face persistent mismatches between their abilities and the conditions under which they are deployed. The relative scarcity of high-quality data on mental health and psychosocial risks in the workplace means that this aspect is often ignored.
Burnout (as defined by World Health Organization) involves emotional exhaustion, cynicism (feeling detached from or negative toward one’s job), and reduced professional efficacy.8 While the Survey of Adult Skills does not include standardised burnout inventories like the Maslach Burnout Inventory or Copenhagen Burnout Inventory (see Romo et al. (2025[21]) for a technical discuss of the existing burnout diagnostic tools), it is possible to create a proxy measure of burnout risk. In particular, a composite Burnout Risk Index can be computed aggregating together the following PIAAC variables:9
Exhaustion: pace of work, frequency of working to tight deadlines.
Cynicism: self-reported job satisfaction, tendency to feel depressed.
Reduced efficacy: perceived over-skilling, perceived overqualification.
These six variables are combined using factor analysis, a statistical method that identifies the common latent structure underlying a set of observed indicators and assigns weights according to the contribution of each variable to the extracted factor. The scores have been rescaled to range between 0 and 1 (using a min-max approach) in order to facilitate interpretation and comparability across units of analysis (see OECD/European Union/EC-JRC (2008[22]) for more details on the construction of composite indices). Clearly, any interpretation should be cautious, as this composite index is about risk factors and not actual burnout prevalence.
Figure 2.9 presents the estimated effects of the use of different skills at work on the Burnout Risk Index. The results show that, after controlling for individual and job characteristics, greater task discretion and the use of self-organising skills are strongly associated with a lower risk of burnout. In both cases the coefficients are negative and highly significant, with task discretion displaying the largest negative effect. This confirms the previous finding in this chapter that workers who have more autonomy in how they perform their tasks, and who rely on organisational skills to manage their workload are more likely to be satisfied at work. Reading, learning at work and writing also exhibit negative and significant coefficients, although the magnitude of these effects is smaller.
By contrast, the use of physical skills shows a positive association with burnout risk, with the coefficient being both statistically significant and relatively large. This indicates that occupations requiring a high degree of physical effort may contribute to higher levels of burnout, possibly due to physical strain, lower job satisfaction or the limited scope for autonomy in such roles. Dexterity and problem-solving skills display weaker and less consistent effects, with coefficients close to zero, while the use of numeracy, ICT, influencing, and co‑operative skills is associated with only modest reductions in burnout risk.
Effect of a one‑standard-deviation increase on Burnout Risk Index
Note: The Figure represents the coefficients of OLS regressions of skills use on the Burnout Risk Index, weighted by sampling weights. Each bar represents a separate regression. The sample includes only full-time workers aged 25‑65. Coefficients are adjusted for gender, immigration background, age, age squared, whether one lives with a partner or has children, educational attainment, literacy proficiency, occupation, industry, firm size, permanent contract, public employment, and country fixed effects. All coefficients are statistically significant (at the 1% significance level).
Source: 2023 Survey of Adult Skills.
References
[5] Acemoglu, D. and D. Autor (2011), “Skills, Tasks and Technologies: Implications for Employment and Earnings”, Handbook of Labor Economics, Vol. 4/PART B, pp. 1043-1171, https://doi.org/10.1016/S0169-7218(11)02410-5.
[15] Adalet McGowan, M. and D. Andrews (2017), “Labor Market Mismatch and Labor Productivity: Evidence from PIAAC Data”, Research in Labor Economics, Vol. 45, pp. 199-241, https://doi.org/10.1108/S0147-912120170000045006.
[4] Agasisti, T., G. Johnes and M. Paccagnella (2021), “Tasks, occupations and wages in OECD countries”, International Labour Review, Vol. 160/1, pp. 85-112, https://doi.org/10.1111/ILR.12169;CTYPE:STRING:JOURNAL.
[16] Andrews, D., B. Égert and C. de la Maisonneuve (2025), “Adult skills and productivity: New evidence from PIAAC 2023”, OECD Economics Department Working Papers, No. 1834, OECD Publishing, Paris, https://doi.org/10.1787/12ac6e8c-en.
[7] Battisti, M., A. Fedorets and L. Kinne (2023), “Cognitive Skills among Adults: An Impeding Factor for Gender Convergence?”, IZA Discussion Paper, No. 16134, IZA Institute of Labor Economics, http://www.iza.org (accessed on 9 September 2025).
[27] Bénabou, R. and J. Tirole (2003), “Intrinsic and Extrinsic Motivation”, The Review of Economic Studies, Vol. 70/3, pp. 489-520, https://doi.org/10.1111/1467-937X.00253.
[19] Boxall, P., A. Hutchison and B. Wassenaar (2015), “How do high-involvement work processes influence employee outcomes? An examination of the mediating roles of skill utilisation and intrinsic motivation”, The International Journal of Human Resource Management, Vol. 26/13, pp. 1737-1752, https://doi.org/10.1080/09585192.2014.962070.
[12] Broecke, S., G. Quintini and M. Vandeweyer (2019), “Wage Inequality and Cognitive Skills: Reopening the Debate”, in Hulten, C. and V. Ramey (eds.), Education, Skills, and Technical Change: Implications for Future U.S. GDP Growth, University of Chicago Press, https://www.nber.org/books-and-chapters/education-skills-and-technical-change-implications-future-us-gdp-growth/wage-inequality-and-cognitive-skills-reopening-debate (accessed on 1 September 2025).
[8] Christl, M. and M. Köppl–Turyna (2020), “Gender wage gap and the role of skills and tasks: evidence from the Austrian PIAAC data set”, Applied Economics, Vol. 52/2, pp. 113-134, https://doi.org/10.1080/00036846.2019.1630707;WGROUP:STRING:PUBLICATION.
[3] De La Rica, S., L. Gortazar and P. Lewandowski (2020), “Job Tasks and Wages in Developed Countries: Evidence from PIAAC”, Labour Economics, Vol. 65, p. 101845, https://doi.org/10.1016/J.LABECO.2020.101845.
[18] Dopeso-Fernández, R., G. Giusti and A. Kucel (2023), “Only the smartest? Motivating job characteristics for all ability levels and their impact on job satisfaction”, Bulletin of Economic Research, Vol. 75/3, pp. 742-775, https://doi.org/10.1111/BOER.12379.
[25] Falck, O., A. Heimisch-Roecker and S. Wiederhold (2021), “Returns to ICT skills”, Research Policy, Vol. 50/7, p. 104064, https://doi.org/10.1016/J.RESPOL.2020.104064.
[17] Fujishiro, K. and C. Heaney (2017), ““Doing what I do best”: The association between skill utilization and employee health with healthy behavior as a mediator”, Social Science & Medicine, Vol. 175, pp. 235-243, https://doi.org/10.1016/J.SOCSCIMED.2016.12.048.
[2] Gunderson, M. and P. Oreopolous (2020), “Returns to education in developed countries”, in Bradley, S. and C. Green (eds.), The Economics of Education: A Comprehensive Overview, Academic Press, https://doi.org/10.1016/B978-0-12-815391-8.00003-3.
[20] Hackman, J. and G. Oldham (1976), “Motivation through the design of work: test of a theory”, Organizational Behavior and Human Performance, Vol. 16/2, pp. 250-279, https://doi.org/10.1016/0030-5073(76)90016-7.
[24] Hanushek, E. et al. (2015), “Returns to skills around the world: Evidence from PIAAC”, European Economic Review, Vol. 73, pp. 103-130, https://doi.org/10.1016/J.EUROECOREV.2014.10.006.
[6] Kawaguchi, D. and T. Toriyabe (2022), “Measurements of skill and skill-use using PIAAC”, Labour Economics, Vol. 78, p. 102197, https://doi.org/10.1016/J.LABECO.2022.102197.
[10] Komatsu, K. (2023), “Gender Gaps in Skill Use: The Case of Japan”, Japan Labor Issues, Vol. 7/45, pp. 43-55.
[14] Mogstad, M., K. Salvanes and G. Torsvik (2025), “Income Equality in the Nordic Countries: Myths, Facts, and Lessons”, IZA Discussion Paper, No. 17677, IZA Institute of Labor Economics, http://www.iza.org (accessed on 10 September 2025).
[23] OECD (2024), Do Adults Have the Skills They Need to Thrive in a Changing World?: Survey of Adult Skills 2023, OECD Skills Studies, OECD Publishing, Paris, https://doi.org/10.1787/b263dc5d-en.
[13] OECD (2015), OECD Employment Outlook 2015, OECD Publishing, Paris, https://doi.org/10.1787/empl_outlook-2015-en.
[26] OECD (2013), OECD Skills Outlook 2013: First Results from the Survey of Adult Skills, OECD Publishing, Paris, https://doi.org/10.1787/9789264204256-en.
[22] OECD/European Union/EC-JRC (2008), Handbook on Constructing Composite Indicators: Methodology and User Guide, OECD Publishing, Paris, https://doi.org/10.1787/9789264043466-en.
[11] Paccagnella, M. (2015), “Skills and Wage Inequality: Evidence from PIAAC”, OECD Education Working Papers, No. 114, OECD Publishing, Paris, https://doi.org/10.1787/5js4xfgl4ks0-en.
[1] Psacharopoulos, G. and H. Patrinos (2018), “Returns to investment in education: a decennial review of the global literature”, Education Economics, Vol. 26/5, pp. 445-458, https://doi.org/10.1080/09645292.2018.1484426.
[9] Rebollo-Sanz, Y. and S. De la Rica (2022), “Gender gaps in skills and labor market outcomes: evidence from the PIAAC”, Review of Economics of the Household, Vol. 20/2, pp. 333-371, https://doi.org/10.1007/S11150-020-09523-W/TABLES/21.
[21] Romo, L. et al. (2025), “Assessment of Burnout in the General Population of France: Comparing the Maslach Burnout Inventory and the Copenhagen Burnout Inventory”, Mental Health Science, Vol. 3/2, p. e97, https://doi.org/10.1002/MHS2.97.
← 1. Similarly to OECD (2024[23]), wages are measured as log hourly earnings for employed and self-employed individuals, including bonuses, expressed in PPP-adjusted 2022 USD.
← 2. Although much of the economic literature (for example, Hanushek et al. (2015[24]) and Falck, Heimisch-Roecker and Wiederhold (2021[25])), often refers to these associations as “returns to skills use”, this is not a rate of return in the strict sense, as it does not incorporate the costs of attaining a given level of skills use.
← 3. Results are confirmed looking at the association between wage inequality and inequality in the use of the other information-processing and generic skills computed in this report.
← 4. Note that an even slightly higher R-squared (R² ≈ 0.54) is found when estimates are adjusted by including a control for inequality in literacy proficiency score.
← 5. In line with the findings of OECD (2013[26]), an even larger R-squared (R² ≈ 0.39) is found when controlling for literacy proficiency score.
← 6. Intrinsic motivation refers to an agent’s engagement in an activity for its own sake – driven by the inherent interest or moral satisfaction associated with the task – whereas extrinsic motivation refers to incentives external to the agent (such as monetary rewards, punishments, or monitoring) (Bénabou and Tirole, 2003[27]).
← 7. It should also be noted that the association may operate in both directions, as individuals who are satisfied in their jobs may also be more inclined to seek additional learning and development opportunities.
← 8. The 11th Revision of the International Classification of Diseases (ICD‑11) of the World Health Organization (WHO) defines burnout as “a syndrome conceptualized as resulting from chronic workplace stress that has not been successfully managed. It is characterized by three dimensions: feelings of energy depletion or exhaustion; increased mental distance from one’s job, or feelings of negativism or cynicism related to one's job; and reduced professional efficacy. Burn-out refers specifically to phenomena in the occupational context and should not be applied to describe experiences in other areas of life.” See here: https://www.who.int/standards/classifications/frequently-asked-questions/burn-out-an-occupational-phenomenon (accessed on 26 November 2025).
← 9. More specifically, the Cycle 2 PIAAC variables used in the Burnout Risk Index are the following: (1) H2_Q08c “To what extent could you choose or change the speed or rate at which you work?”; (2) H2_Q12 “How often does your current job usually involve working to tight deadlines or at very high speed?”; (3) D2_Q13 “All things considered, how satisfied are you with your current work?”; (4) K2_Q02d “Indicate the extent to which you agree or disagree with the statement ‘I tend to feel depressed, blue’”; (5) H2_Q19a “Overall, which of the following statements best describes your skills in relation to what is required to do your job?”; and (6) D2_Q12a “If applying today, what would be the usual qualification, if any, that someone would need to get this type of job?”. In particular, over-qualification and over-skilling are included as components of reduced professional efficacy because, when individuals perceive that their skills or qualifications exceed those required by their job, they are more likely to experience under-utilisation, lack of challenge, and diminished sense of purpose or accomplishment at work.