copy the linklink copied! 5. The outcomes of investments in skills

This chapter looks at how proficiency in literacy, numeracy and problem solving in technology-rich environments makes a difference to the outcomes experienced by individuals – and how these differ among the six countries that participated in the third round of the Survey of Adult Skills. It finds that proficiency is positively linked to a number of important economic and social outcomes – not just employment and wages, but also aspects of well-being such as health, volunteering and political efficacy. It also considers the impact of wages on mismatches between workers’ qualifications and skills and those needed for their jobs.

A note regarding Israel

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

Previous chapters of this report have compared the level, distribution and use of information-processing skills among countries and socio-demographic groups. This chapter focuses on the relationship between proficiency and skills use and labour-market and social outcomes including employment, earnings, skills mismatches, self-reported health, participation in associative or volunteer activities, and individuals’ sense of influence over the political process. It discusses the results with a particular focus on numeracy proficiency but very similar results would be obtained using literacy proficiency.

The main findings of this chapter are:

Educational qualifications and proficiency in information-processing skills reflect different aspects of an individual’s human capital and are separately identified and valued in the labour market. After the effects of educational attainment have been taken into account, an increase of one standard deviation in an individual’s numeracy proficiency (56 score points) is associated with a 1.6 percentage-point increase in the probability of being employed as opposed to being unemployed. The same increase in numeracy proficiency is also associated with a 7% increase in hourly wages, on average across the OECD countries and economies participating in the Survey of Adult Skills (PIAAC).
Proficiency and years of education appear to play a very small role in the employment outcomes of adults in all Round 3 countries, with the exception of Hungary. The relationship between proficiency and hourly wages is also relatively weak in these Round 3 countries (excluding Hungary) and below the OECD average, while years of education are more strongly correlated with hourly wages than the OECD average, particularly in Kazakhstan and the United States. In Hungary, on the other hand, proficiency in numeracy and years of education are comparatively strong predictors of employment and wages. This is likely to reflect differences in institutional arrangements (such as wage-setting mechanisms) as well as the relative weight given to educational qualifications and other factors in employers’ hiring, promotion and wage-setting decisions.
Mismatches between workers’ qualifications and skills and what they report as required or expected in their jobs are pervasive in most countries and economies participating in PIAAC.
On average across the OECD countries and economies that participated in the Survey of Adult Skills, about 22% of workers report that they are overqualified – that they have higher qualifications than required to get their jobs – and 12% report that they are underqualified. Moreover, 11% have higher levels of numeracy skills than those typically required in their job, and 4% are underskilled. Finally, 40% of workers are mismatched by field of study: they work in an occupation that is unrelated to their field of study. These forms of mismatch overlap; it is common for workers who are mismatched by field of study to also be overqualified, for example.
In Hungary, Kazakhstan and the United States, the overall incidence of skill mismatch is at or below the rate observed in the OECD on average. In contrast, the Latin American countries in Round 3 – Ecuador, Mexico and Peru – stand out along with Chile from Round 2 for their very high incidence of overskilling. Although measured differently, this is in line with the relatively low use of literacy and numeracy skills in the workplace in these countries (see Chapter 4).
Chile, Ecuador and Mexico, along with the United States, also have a relatively high incidence of mismatches by field of study: 10 percentage points higher than the OECD average in Chile, 17 percentage points higher than average in Ecuador, 12 percentage points higher than average in Mexico and 8 percentage points higher than average in the United States. Latin American countries may be more likely to lack training systems that provide relevant skills and are aligned with the needs of the economy (OECD, 2018[1]). However, the difference could also be explained on statistical grounds, in countries with a large population of graduates from general programmes. Finally, Ecuador stands out in that underqualification is more common than overqualification. This could reflect the rapid growth in the demand for higher qualifications, which has not been matched by an equivalent increase in graduate numbers. The incidence of qualification mismatch is lower than average in Peru and Mexico.
Qualification and skills mismatches may both have distinct effects on wages, even after adjusting for both qualification level and proficiency scores, because jobs with similar qualification requirements may have different skill requirements. This may happen because employers can evaluate qualifications but they cannot measure skills directly. Overqualification has a stronger negative association with real hourly wages than overskilling or field-of-study mismatches. On average across participating OECD countries, overqualified workers earn about 17% less than well-matched workers with the same qualification and proficiency levels and in the same field. The equivalent wage penalty is 7% less for overskilling, and 3% less for field-of-study mismatch. Among the Round 3 countries, Peru and the United States stand out for having one of the largest wage penalties for overqualification. Ecuador is unique in that none of the forms of mismatch analysed in this study is associated with differences in hourly wages.
Proficiency in literacy, numeracy and problem solving in technology-rich environments is positively associated with several aspects of well-being identified using PIAAC. On average in OECD countries, proficiency in these information-processing skills is positively associated with trust, volunteering, political efficacy and self-assessed health. The relationships with political efficacy and self-assessed health hold even after accounting for a range of socio-demographic characteristics. On the other hand, the association with trust becomes very small and that with volunteering is no longer statistically significant once individual characteristics are accounted for. The strength of the association varies across countries. With the exception of Hungary and the United States, the countries in Round 3 have weaker relationships between numeracy proficiency and non-economic outcomes than most of the other countries included in PIAAC. At the other end of the spectrum, all relationships are positive and statistically significant in the United States.

The results suggest that, independent of policies designed to increase participation in education and training, improvements in adults’ skill levels may provide considerable economic and social returns for individuals and society as a whole. Improvements in adults’ skill levels can be brought about by the teaching of literacy and numeracy in schools and by programmes for adults with poor literacy and numeracy skills and limited familiarity with information and communications technology (ICT), through training in the workplace, and greater use of skills in and outside work to avoid their deterioration.

copy the linklink copied! Skills proficiency, labour-market status and wages

To the extent that workers’ productivity is related to the knowledge and skills they possess, and that wages reflect such productivity, albeit imperfectly, individuals with greater skills should expect higher returns from their participation in the labour market and thus be more likely to participate. Most studies use individuals’ past educational qualifications as a proxy for their current productive potential when investigating the returns to investments in human capital; until the release of the Survey of Adult Skills (PIAAC), only a few studies examined the return on actual skills (Leuven, Oosterbeek and van Ophem, 2004[2]; Tyler, 2004[3]). Since the release in 2013 of the first round of results, PIAAC has provided an opportunity to test, with validly comparable data, how information-processing skills influence individuals’ employment chances and wages. Based on the countries and economies that participated in the first two rounds of PIAAC, an increase of one standard deviation in an individual’s literacy proficiency (48 score points) is associated with a 0.8 percentage-point increase in the probability of being employed. An increase of one standard deviation in literacy proficiency is also associated with a 6% increase in hourly wages in these countries (OECD, 2016[4]). Other researchers have confirmed the labour-market value of skills (Hanushek et al., 2015[5]; Vignoles, 2016[6]).

Since three of the five countries that implemented the Survey of Adult Skills for the first time in 2018 are from Latin America, it is worth noting that several studies have looked at returns to education and skills in this region. They exploited skill surveys such as the World Bank Skills Towards Employment and Productivity survey (STEP; 2012) and the Peruvian Skills and Labor Market Survey (ENHAB; 2010) or simply data on educational attainment and wages or employment status. The findings generally suggest a decline in returns to educational attainment over time, particularly for upper secondary graduates (i.e. those scoring at Level 3 of the International Standard Classification of Education, ISCED), due to supply factors such as an increase in the number of graduates at this level, and demand factors such as a shift in demand towards higher education graduates (Manacorda, Sánchez-Páramo and Schady, 2010[7]). Using data for Colombia, Acosta, Muller and Sarzosa (2017[8]) found that cognitive skills are strongly related to higher earnings while socio-emotional skills are strongly correlated with participation in employment. In Peru, cognitive skills are found to increase wages, after controlling for educational attainment and socio-emotional skills (José Díaz, Arias and Tudela, 2014[9]). Consistent results are reported for Bolivia in Cunningham, Acosta and Muller (2016[10]).

This section reviews the relationship between skills proficiency, employment status and wages with particular focus on the Round 3 countries: Ecuador, Hungary, Kazakhstan, Mexico, Peru and the United States. Throughout the chapter, the OECD average refers to the OECD countries and economies that have implemented the survey, either in 2012 or 2015 or 2018.

Proficiency, education and employment

Findings from previous rounds of PIAAC have confirmed that proficiency in literacy and numeracy plays an important and independent role in determining success in the labour market, over and above the role played by formal education, although it is hard to identify how much this statistical association captures the true causal effect of skills on wages.1

Among the OECD countries and economies that have implemented the Survey of Adult Skills in any of the three rounds, an individual who scores one standard deviation higher than another on the numeracy scale (around 56 score points) is 1.6 percentage points more likely to be employed than unemployed (Figure 5.1). An increase in one standard deviation in the number of years in formal education (around 3.3 years) is associated with a 2.4 percentage-point increase in the chances of being employed. Among the countries participating in PIAAC in the third round in 2018, only Hungary had similar results, with a positive association between employment rates and both numeracy proficiency and educational attainment. In Ecuador, Kazakhstan, Mexico, Peru and the United States there are low or negative returns to proficiency and education, which in most cases are not statistically significant.

Figure 5.1. Effect of education and numeracy proficiency on the likelihood of being employed

These relationships are calculated holding everything else constant. For instance, the relationship between proficiency and the probability of being in employment is computed by comparing the likelihood of being employed among adults with different proficiency but who have spent the same number of years in education and share the same socio-demographic characteristics. Such a calculation is possible because of the imperfect overlap of education and proficiency, as discussed in previous chapters.

The finding, in most countries including Hungary, that educational attainment is a better predictor of employment than numeracy proficiency suggests that it is harder for employers to judge workers’ actual numeracy proficiency. As a result, employers are more likely to rely on readily available, albeit imperfect, signals such as educational qualifications. However, skills become a stronger predictor of labour-market outcomes as tenure in the job increases, a phenomenon called “employer learning”, referring to the fact that employers learn about their employees’ skills once they have been hired (OECD, 2014[11]).

The lack of a relationship between employment status and education and proficiency in Latin American countries is striking. It is, however, in line with previous studies on Latin American countries that have shown a stronger correlation of cognitive skills with earnings than with employment status (Cunningham, Acosta and Muller, 2016[10]; Acosta, Muller and Sarzosa, 2017[8]). The absence of strong social protection systems in these countries may lead to most adults engaging in any employment they can find, possibly in the informal sector (OECD, 2015[12]). More education and greater proficiency could therefore translate into higher-quality jobs, rather than a greater chance of being employed.

Numeracy proficiency, education and wages

Hourly wage levels are strongly associated with numeracy proficiency. This relationship is explored in Figure 5.2 after adjusting for several individual characteristics, including years of education. As with the likelihood of employment, it is difficult to determine the direction of causality. For instance, higher wages may be characteristic of occupations that favour workers acquiring skills through formal education. This section uses linear regression analysis to distinguish years of education from skills proficiency to help determine whether returns to education reflect the fact that highly educated individuals tend to have greater proficiency in information processing skills, or the fact that employers value their credentials.

Proficiency and schooling have significant and distinct effects on hourly wages. Across the OECD countries that have implemented the Survey of Adult Skills in any one of the three rounds, an increase in one standard deviation in numeracy proficiency is associated with a 7% increase in hourly wages, keeping years of education and other socio-demographic characteristics constant. An increase in years of education by one standard deviation brings about a bigger increase in hourly wages of about 18%, all else being equal. Returns to proficiency are above average in Hungary, while they are below average in Ecuador, Kazakhstan, Mexico, Peru and the United States. The relationship is weakest in Ecuador, where it is not statistically significant. Returns to years of education exceed the OECD average in all Round 3 countries, with the exception of Peru. Hungary shows the third highest returns to years of education of all participating countries, after Singapore and Slovenia. The results for Peru are strikingly close to those obtained by José Díaz, Arias and Tudela (2014[9]) using the ENHAB survey. In that study the authors showed that one standard deviation increase in cognitive skills was associated with a 9% increase in wages, while for years of education the increase was 15%. The figures for PIAAC are 7% and 14%.

Figure 5.2. Effect of education, numeracy proficiency and numeracy use at work on wages

Part of the effect of proficiency on hourly wages may be based on the type of tasks and responsibilities workers are expected to carry out in their job. In addition to years of education and skills proficiency, the analysis behind Figure 5.2 considers the use of numeracy skills at work. Workers in jobs that require more intense use of numeracy also earn higher wages. Including skills use at work also serves to show that the effect of skills proficiency is not driven by selection effects. It is not that more proficient workers earn more because they are selected into more skill-intensive jobs. They earn higher wages than less proficient workers who hold jobs with similar skill requirements. Interestingly, while the use of numeracy at work is correlated with hourly wages in Hungary and the United States, this is not the case in Ecuador, Kazakhstan, Mexico and Peru.

Overall, the number of years of education tends to have a smaller impact on wages in countries with a more compressed wage distribution, such as the Nordic countries, Italy and Flanders (Belgium) (OECD, 2015[13]; OECD, 2015[12]). In contrast, greater educational attainment is associated with significantly higher wages in Germany, Chile, Mexico and Turkey, all of which have relatively high earnings inequality. However, this only suggests a link between the earnings distribution and returns to education, as other factors affect the ranking of countries. For instance, Slovenia – where earnings inequality is relatively low – shows relatively high returns to education.

The relative contribution of education, proficiency and other factors to the variation in individual wages

As shown in Figure 5.2, educational attainment and proficiency in information-processing skills contribute independently to explaining individuals’ wages. To compare the size of their contribution, the analysis conducted above looked at how one standard deviation in educational attainment or skills proficiency relates to wages. A better way to compare is to look at how much of the variance in wages each variable explains (OECD, 2014[11]). Figure 5.3 does precisely this, comparing the relative importance of proficiency and years of education and other variables reflecting job- and field-specific knowledge such as work experience and field of study. Together these variables explain about 26% of the variance in wages while individual characteristics like gender, migrant status, marital status and the language spoken at home contribute an additional 4%, on average across OECD countries. Information-processing skills contribute 4.5%, educational attainment explains 12%, field of study 1% and experience 9%. In Ecuador, Mexico and Peru, these human capital variables account for about 20% of the variation in hourly wages, below the OECD average. In Kazakhstan, these factors account for only 13% of the variation in hourly wages. On the other hand, they account for about one-third of the variation in Hungary and the United States, almost 8 percentage points more than the OECD average.

Figure 5.3. Contribution of education, literacy and numeracy to the variation of hourly wages

The contribution of information-processing skills to the variance of hourly wages is greatest in England (United Kingdom) and Singapore where it exceeds 10%. The contribution of literacy and numeracy proficiency is also high in Hungary, where it is close to 8%, while it is low in Ecuador, Kazakhstan, Mexico, Peru and the United States, which are all at the bottom end of the scale. Overall, years of schooling are more important in understanding the returns to human capital than proficiency. Hungary is one of five countries where years of education account for 20% or more of the variance of wages. The variance in hourly wages explained by years of education in Kazakhstan is relatively small while is it slightly above average in Ecuador and the United States and very close to the average in Peru and Mexico. Comparing the share of variance explained by proficiency and years of education, only in Israel and England (United Kingdom) does proficiency contribute more to the variance of wages than years of schooling. In all Round 3 countries, years of education appear to play a bigger role in explaining returns to human capital than proficiency, although, with the exception of Hungary, both factors explain a relatively small portion of the variance compared with other countries. Finally, Mexico stands out in the analysis as one of the countries where field of study contributes the most to the variation in hourly wages. Differences among countries in the magnitude of these effects are likely to be influenced by how wages are distributed across occupations and, in turn, by the labour-market institutions, such as minimum wages and unions, that affect that distribution.

The relative importance of different human capital factors across age groups and gender is presented in Figure 5.4. Information-processing skills explain a larger share of the variance in wages among 30-49 year-old and 50-65 year-old workers than among younger ones (16-29 year-olds), on average across participating OECD countries. Across all countries participating in the survey, and net of differences between countries, proficiency in numeracy and literacy explains 3% of the variance in wages among 16-29 year-olds, 6% among 30-49 year-olds and 5% among 50-65 year-olds. This is in line with the concept of “employer learning” (OECD, 2014[11]; Pinkston, 2009[15]). Overall, human capital components (proficiency, education, field of study and experience) explain a larger portion of the variance in hourly wages for 30-65 year-olds than for the youngest workers.

Interestingly, credentials are found to play a bigger role in explaining returns to human capital for women than for men. Years of education and field of study account for 14.5% of the variance in hourly wages for women, compared with 12% for men. On the other hand, experience and proficiency play a bigger role for men than for women.

Figure 5.4. Contribution of education, literacy and numeracy to the variation of hourly wages, by age group and gender

copy the linklink copied! Mismatches between workers’ skills and job requirements and the impact on wages

Ensuring a good match at the aggregate level between the skills acquired in education and on the job and those required in the labour market is essential if countries want to make the most of their investments in human capital. It is also a desirable outcome for individuals who have themselves invested in education. A mismatch between workers’ skills and the demands of their job has potentially significant economic implications. For individuals, it affects job satisfaction and wages. For employers, it increases the rate of turnover and may reduce productivity. At the macroeconomic level, mismatches increase unemployment and reduce growth through the inefficient use of human capital and/or a reduction in productivity (McGowan and Andrews, 2015[16]). That said, some level of mismatch is inevitable. Requirements for skills and qualifications are never fixed. The task content of jobs changes over time in response to technological and organisational change, the demands of customers, and in response to the evolution of the supply of labour. Young people leaving education and people moving out of unemployment, for example, may take jobs that do not necessarily fully match their qualifications and skills. Thus, for a number of reasons, some workers are likely to be employed in jobs for which they are too highly qualified and others may be in jobs, at least temporarily, for which they lack adequate schooling.

Mismatches in the Survey of Adult Skills

The Survey of Adult Skills (PIAAC) offers a unique source of data regarding aspects of skills and qualifications mismatches as it includes information on workers’ qualifications and experience, their perceptions of the qualification requirements of their jobs, the task composition of their jobs, and their proficiency in key information-processing skills. This section examines three types of mismatches: in qualifications, field of study and skills. These are defined in Box 5.1 below. While these measures focus on different aspects of mismatches, they overlap to some extent, just as education levels, fields of study and skills do. For instance, graduates who face difficulties finding work in their field of study may accept a job in a different field and below their level of qualification because they lack some of the specific knowledge required by that job. In this case, they would be mismatched by field of study and overqualified.

copy the linklink copied!

Box 5.1. Measuring mismatches in qualifications, skills and fields of study in the Survey of Adult Skills

In general, and for every type of mismatch, there are several measurement strategies. Surveys can ask respondents about their own appraisal with regards to mismatch (subjective measures), or compare respondents to what is common in their country (statistical approaches) or what is, by definition, appropriate (normative approaches). Each type of measure has its advantages and disadvantages.

Qualification mismatches arise when workers’ educational attainment levels are higher or lower than required for their jobs. If they are more highly educated than their job requires, workers are classified as overqualified; if the opposite is true, they are classified as underqualified. In PIAAC, workers are asked what would be the usual qualifications, if any, “that someone would need to get (their) type of job if applying today”. The answer to this question is used as each worker’s qualification requirement and compared to their actual qualification to identify mismatches. Although they can be biased by individual perceptions and period or cohort effects, these kind of self-reported qualification requirements have the advantage of being job specific rather than assuming that all jobs with the same occupational code require the same level of qualification.

Skills mismatches arise when workers have skills levels that are higher or lower than required for their jobs. If they have greater skills than the maximum their job requires, workers are classified as overskilled; if their skills are below the minimum, they are classified as underskilled. For the purpose of this chapter, skill requirements at work, the key term in the measurement of skills mismatch, are derived following Pellizzari and Fichen (2013[17]). The maximum and minimum skill score required for each occupation are defined based on the proficiency of respondents who are classified as well matched to their job. The well matched are workers who replied that they do not feel they “have the skills to cope with more demanding duties than those they are required to perform in their current job” and who also replied that they do not “need further training in order to cope well with their personal duties”. This measure of skills mismatch is robust to reporting bias, such as overconfidence, and it does not impose the strong assumptions needed to directly compare proficiency and skills use. However, this approach does not measure all forms of skills mismatch; rather, it focuses on mismatches in the proficiency domains assessed by the Survey of Adult Skills, leaving out mismatches related to job-specific skills or involving generic skills.

Field-of-study mismatches arise when workers are employed in a different field from the one they have specialised in. The matching is based on a list of occupations, narrowly defined using the International Standard Classification of Occupations (3-digit ISCO classification) that are considered an appropriate match for each field of study. Workers who are not employed in an occupation that is considered a good match for their field are counted as mismatched. The list of fields and occupations used in this chapter can be found in Annex 5 of the 2014 edition of the OECD Employment Outlook (OECD, 2014[11]). The list is largely based on that developed by Wolbers (2003[18]) but has been adapted to the ISCO 08 classification (Montt, 2015[19]).

Sources:

OECD (2017[20]), Getting Skills Right: Skills for Jobs Indicators, https://dx.doi.org/10.1787/9789264277878-en; OECD (2014[11]), OECD Employment Outlook 2014, https://dx.doi.org/10.1787/empl_outlook-2014-en; Montt (2015[19]), “The causes and consequences of field-of-study mismatch: An analysis using PIAAC”, https://dx.doi.org/10.1787/5jrxm4dhv9r2-en; Pellizzari and Fichen (2013[17]), “A new measure of skills mismatch: Theory and evidence from the Survey of Adult Skills (PIAAC)”, https://dx.doi.org/10.1787/5k3tpt04lcnt-en; Wolbers (2003[18]), “Job mismatches and their labour-market effects among school-leavers in Europe”, www.socsci.ru.nl/maartenw/esr03-3.pdf.

The main piece of information needed to determine whether workers are over- or underqualified is to measure the level of education required in their jobs. PIAAC asks workers what qualification they consider would be necessary to get their job today. The comparison between workers’ qualifications and this self-reported requirement shows that, on average, 22% of workers are overqualified while about 12% are underqualified (Figure 5.5). The incidence of qualification mismatch varies significantly across countries. In all Round 3 countries except Ecuador and Kazakhstan, the overall qualification mismatch rate is lower than in the OECD average. Kazakhstan has an overall rate very close to the OECD average, although the composition is slightly different with overqualification playing a bigger role than on average. Ecuador has a relatively high overall rate and is one of only five PIAAC countries, where being underqualified is more common than being overqualified. This could reflect rapid growth in the demand for higher qualifications not matched by an equivalent increase in graduate numbers.

PIAAC also identifies workers who are overskilled or underskilled by comparing their proficiency score in a given domain to the maximum and minimum score required by their occupation (see Box 5.1). Workers are overskilled in a domain if their score is higher than the maximum score required and they are underskilled if their score is lower than the minimum score required. In Hungary, Kazakhstan and the United States, the overall incidence of skill mismatch is at or below the rate observed in the OECD on average. By contrast, Latin American countries stand out, with incidences that are well above average. This applies to Ecuador, Mexico and Peru from Round 3 but also to Chile from Round 2 and is mostly due to an above-average incidence of overskilling. One possibility is that skill requirements are weighed down by the low skill levels of workers on average. This would make highly skilled individuals stand out even in occupations that would normally require higher-level qualifications and competencies. Although measured differently, this finding is in line with the relatively limited use of literacy and numeracy in the workplace in these countries.

Chile, Ecuador, Mexico and the United States have a relatively high incidence of mismatches by field of study, whereby workers are in jobs that are not related to their field of study. The incidence of field-of-study mismatch is 10 percentage points above the OECD average of 40% in Chile, 17 percentage points higher in Ecuador, 12 percentage points in Mexico and 8 percentage points in the United States. This could be due to a poorer alignment of education choices with labour-market needs as well as to the predominantly general nature of secondary education. Findings by the Inter-American Development Bank (Novella et al., 2019[21]; Rucci, 2017[22]) suggest that, in Latin American countries, the alignment between the content of education and training and labour-market requirements may be particularly poor. Statistically, a high incidence of field-of-study mismatch could also be due to a relatively small sample size, since these countries have a very large share of adults with only a general upper secondary education who by definition are excluded from the field-of-study analysis (Montt, 2015[19]).

Figure 5.5. Mismatches in qualifications, numeracy and fields of study

The effect of mismatches on wages

Overqualification has a stronger negative effect on real hourly wages than mismatches in skills or field of study, when workers are compared with equally qualified and equally proficient well-matched counterparts (Figure 5.6). On average across OECD countries and economies, overqualified workers earn about 17% less than well-matched workers with the same qualification and proficiency levels and in the same field of study. The equivalent wage penalty for overskilling is 7%, and for field-of-study mismatches it is 3%. While the negative correlation between overqualification and wages is consistent and statistically significant across the vast majority of countries, this is not the case for overskilling and field-of study mismatch. In Kazakhstan, Mexico, Peru and the United States, the wage penalties related to overqualification are above average, particularly in Peru and the United States where the hourly wages of overqualified workers are more then 30% lower than the hourly wages of well-matched workers who have the same level and field of qualification and the same proficiency in numeracy. None of the mismatch variables are associated with changes in hourly wages in Ecuador. Finally, the wage penalty associated with overqualification in Hungary is below the OECD average, while over-skilling and field-of-study mismatch do not have any statistically significant association with hourly wages.

Figure 5.6. Effect of mismatches in qualifications, numeracy and fields of study on wages

This evidence should not be interpreted as suggesting that having skills in excess of those required at work is not valued at all on the labour market. On average across countries, overqualified workers earn about 4% more than well-matched workers in similar jobs. In other words, tertiary graduate who hold jobs requiring only an upper secondary qualification will earn less than if they were in jobs requiring a tertiary qualification, but more than upper secondary graduates in jobs requiring upper secondary qualifications.

Qualification and skills mismatches may both have distinct effects on wages, even after adjusting for both qualification level and proficiency scores, because jobs with similar qualification requirements may have different skill requirements. This may happen because employers can evaluate qualifications but they cannot measure skills directly. In addition, the kinds of mismatches in skills captured by the two indicators are different: the survey’s indicators of skills mismatch are based on numeracy, literacy and problem solving, while skills mismatches captured by qualification-based indicators may be interpreted as more general and may be based, for example, on the level of job-specific skills.

copy the linklink copied! Non-economic outcomes of information-processing skills

While employability and wages are important for individual well-being, non-economic factors also contribute both to individual well-being and to the smooth functioning of societies. PIAAC collects information on four non-economic outcomes: the level of trust in others; participation in associative, religious, political, or charity activities (volunteering); the sense of being able to influence the political process (i.e. political efficacy); and self-assessed health conditions.

Trust, volunteering and political efficacy are variables collected in many surveys, such as the World Value Survey (www.worldvaluessurvey.org) and the European Social Survey (www.europeansocialsurvey.org). They are often used as proxies to measure social capital in the extensive economic and sociological literature that has investigated the link between social capital (and cultural traits) and long-term economic development (OECD, 2016[4]). The Survey of Adult Skills offers a unique opportunity to better understand the relationships between education, skills proficiency and widely used measures of social capital and individual well-being. Depending on the subjective value attached to the various non-economic outcomes, they can be seen as either interesting outcomes in themselves, or, in light of the vast literature on the relationship between social capital and economic growth, as mediating variables in studying the relationship between skills proficiency and economic outcomes.

As Figure 5.7 illustrates, on average in the OECD, proficiency in information-processing skills is positively associated with trust, volunteering, political efficacy and self-assessed health. The relationships with political efficacy, self-assessed health and volunteering hold even after accounting for the usual range of socio-demographic characteristics. On the other hand, the association with trust becomes very small and, in many instances, is no longer statistically significant once individual characteristics are accounted for.

Figure 5.7. Effect of numeracy proficiency on positive social outcomes

The strength of the association differs among countries. With the exception of Hungary and the United States, countries in Round 3 have weaker relationships overall between proficiency in numeracy and non-economic outcomes than most of the other countries included in PIAAC. In Ecuador and Peru, only political efficacy and health outcomes are positively associated with proficiency, while in Mexico only volunteering and self-reported health bear a positive association with proficiency. In Kazakhstan, only participation in volunteering activities bears any relationship with numeracy proficiency, once controls are applied. On the other hand, in Hungary, only political efficacy is not correlated with numeracy proficiency and all relationships are positive and statistically significant in the United States.

copy the linklink copied! Summary

Proficiency in literacy, numeracy and problem solving in technology-rich environments is positively and independently associated with the probability of participating in the labour market and being employed, and with higher wages. Proficiency in these information-processing skills is also positively associated with other important aspects of well-being, notably health and beliefs about one’s impact on the political process. The findings in this chapter, along with those in Chapter 3, also highlight the distinction between qualification and skills: some workers have lower proficiency in skills than would be expected given their educational level, either because they performed poorly during their initial education or because their skills have declined over time. This can lead to significant mismatches, particularly as skills are difficult for employers to gauge and qualifications are routinely used as signals of individual ability. The resulting mismatch between the skills a worker possesses and those required at work is associated with a sizeable reduction in wages.

The relationship between numeracy proficiency and labour market outcomes is weaker than average in several Round 3 countries, notably in Latin America. In these countries, years of education are a better predictor of wage outcomes than numeracy proficiency but are weakly correlated to the likelihood of being employed. Overskilling and mismatches in fields of study are also more common in these countries than on average, suggesting poor alignment between their education systems and labour-market needs. The picture is more mixed for Kazakhstan, where some indicators point to a weak association between skills and labour market outcomes while others are in line with the OECD average. Of all the Round 3 countries, Hungary and the United States tended to perform the closest to the OECD average.

Overall, the results suggest that investments in improving adults’ proficiency in literacy, numeracy and problem solving in technology-rich environments may have significant benefits. Independent of policies designed to increase participation in education and training, improvements in the teaching of literacy and numeracy in schools and programmes for adults with poor literacy and numeracy skills and limited familiarity with ICTs may result in considerable economic and social returns for individuals and for society as a whole.

copy the linklink copied!

A note regarding the Russian Federation

The sample for the Russian Federation does not include the population of the Moscow municipal area. More detailed information can be found in the Technical Report of the Survey of Adult Skills, Third Edition (OECD, 2019[23]).

References

[8] Acosta, P., N. Muller and M. Sarzosa (2017), Beyond Qualifications: Returns to Cognitive and Socioemotional Skills in Colombia, Institute of Labor Economics, http://conference.iza.org/conference_files/Statistic_2018/sarzosa_m23031.pdf (accessed on 7 May 2019).

[10] Cunningham, W., P. Acosta and N. Muller (2016), Minds and Behaviors at Work: Boosting Socioemotional Skills for Latin America’s Workforce, World Bank Publications, http://documents.worldbank.org/curated/en/290001468508338670/Minds-and-behaviors-at-work-boosting-socioemotional-skills-for-Latin-America-s-workforce (accessed on 7 May 2019).

[14] Fields, G. (2004), Regression-Based Decompositions: A New Tool for Managerial Decision-Making, Cornell University, ILR school, http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.494.9450&rep=rep1&type=pdf.

[5] 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.

[9] José Díaz, J., O. Arias and D. Tudela (2014), Does perseverance pay as much as being smart?: The returns to cognitive and non-cognitive skills in urban Peru, 9th IZA/World Bank Conference on Employment and Development 25-26 June, Lima, http://conference.iza.org/conference_files/worldb2014/arias_o4854.pdf (accessed on 7 May 2019).

[2] Leuven, E., H. Oosterbeek and H. van Ophem (2004), “Explaining international differences in male skill wage differentials by differences in demand and supply of skill”, Economic Journal, Vol. 114/495, pp. 466-486, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=526835 (accessed on 7 May 2019).

[7] Manacorda, M., C. Sánchez-Páramo and N. Schady (2010), “Changes in returns to education in Latin America: The role of demand and supply of skills”, Industrial and Labor Relations Review, Vol. 63/2, pp. 307-326, https://doi.org/10.1177%2F001979391006300207 (accessed on 7 May 2019).

[16] McGowan, M. and D. Andrews (2015), “Labour market mismatch and labour productivity: Evidence from PIAAC data”, OECD Economics Department Working Papers, No. 1209, OECD Publishing, Paris, https://doi.org/10.1787/5js1pzx1r2kb-en (accessed on 28 February 2018).

[19] Montt, G. (2015), “The causes and consequences of field-of-study mismatch: An analysis using PIAAC”, OECD Social, Employment and Migration Working Papers, No. 167, OECD Publishing, Paris, https://dx.doi.org/10.1787/5jrxm4dhv9r2-en.

[21] Novella, R. et al. (2019), Encuesta de habilidades al trabajo (ENHAT) 2017-2018: Causas y consecuencias de la brecha de habilidades en Perú, Inter-American Development Bank, Washington, D.C., https://doi.org/10.18235/0001653.

[23] OECD (2019), Technical Report of the Survey of Adult Skills, Third Edition, OECD, http://www.oecd.org/skills/piaac/publications/PIAAC_Technical_Report_2019.pdf.

[1] OECD (2018), Getting Skills Right: Brazil, Getting Skills Right, OECD Publishing, Paris, https://dx.doi.org/10.1787/?9789264309838-en.

[20] OECD (2017), Getting Skills Right: Skills for Jobs Indicators, Getting Skills Right, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264277878-en.

[4] OECD (2016), Skills Matter: Further Results from the Survey of Adult Skills, OECD Skills Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264258051-en.

[13] OECD (2015), In It Together: Why Less Inequality Benefits All, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264235120-en.

[12] OECD (2015), OECD Employment Outlook 2015, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2015-en.

[11] OECD (2014), OECD Employment Outlook 2014, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2014-en.

[17] Pellizzari, M. and A. Fichen (2013), “A new measure of skills mismatch: Theory and evidence from the Survey of Adult Skills (PIAAC)”, OECD Social, Employment and Migration Working Papers, No. 153, OECD Publishing, Paris, https://dx.doi.org/10.1787/5k3tpt04lcnt-en.

[15] Pinkston, J. (2009), “A model of asymmetric employer learning with testable implications”, Review of Economic Studies, Vol. 76/1, pp. 367-394, https://doi.org/10.1111/j.1467-937X.2008.00507.x.

[22] Rucci, G. (2017), Skills Mismatches in Latin America and the Caribbean, Skills for Employment, https://www.skillsforemployment.org/KSP/en/Details/?dn=WCMSTEST4_189325 (accessed on 11 May 2019).

[3] Tyler, J. (2004), “Basic skills and the earnings of dropouts”, Economics of Education Review, Vol. 23/3, pp. 221-235, https://doi.org/10.1016/j.econedurev.2003.04.001.

[6] Vignoles, A. (2016), “What is the economic value of literacy and numeracy?”, IZA World of Labor 229, https://doi.org/10.15185/izawol.229.

[18] Wolbers, M. (2003), “Job mismatches and their labour-market effects among school-leavers in Europe”, European Sociological Review, Vol. 19/3, pp. 249-266, http://www.socsci.ru.nl/maartenw/esr03-3.pdf (accessed on 10 May 2019).

Note

← 1. For example, employment may itself favour the acquisition of skills or prevent the depreciation of workers’ skills that are not put to use whilst unemployed.

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

https://doi.org/10.1787/1f029d8f-en

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at http://www.oecd.org/termsandconditions.

╳

OECD Skills Studies: Skills Matter: Additional Results from the Survey of Adult Skills

Do you want to open the publication page on OECD-ILIBRARY for more details on:

available formats

purchase of a print copy

or to publication metadata?

Cancel

5. The outcomes of investments in skills