Age, Cohort & Cognitive Skills
Thematic Focal Person: Vegard Skirbekk
Overall vision
Foreseeing the development in skills and capacities among those aged 50 and above is paramount for societies that are ageing more profoundly than ever before. Postponed retirement is the most important policy action to maintain stability between the economically active and inactive populations (OECD 2006; UN 2007). The ability to raise employment at older ages effectively depends on a thorough understanding of the levels, trends and determinants of age-specific productivity potential. Several studies (McEvoy and Cascio 1989; Skirbekk 2008) conclude that on average the age-productivity curve tends to be hump¬shaped, wherein initial years of labour market experience boosts working capacities, while experiencing declines in certain cognitive skills (such as short term memory and processing speed) and health can lower productivity towards the end of the working life. Although the importance of cognitive abilities in the labour market is increasing, little is known about the prevalence of these skills in the coming decades. If later-born cohorts are increasingly healthier and more cognitively capable, this will imply that older individuals are likely to remain productive in the workforce to a higher age. Although cohort effects were presented years ago by Quetelet (1842), the cohort dimension has been ignored in most contemporary investigations of productivity and demographic change. Several important productivity determinants among seniors do change along cohort lines, including mental ability levels, health, educational attainment and work capacities (Romeu Gordo 2009; Skirbekk 2008b). We study the changes in these productivity determinants in an integrated common framework that allows us to project the composition of skills by age, sex and birth cohort.
The shape of the age-productivity profile differs according to age variation in the skills relevant to the work (Skirbekk 2008b; Warr 1994). Changes in the industrial structure and less physically intense work in traditionally physically demanding jobs have given way to non¬routine interactive and analytical work. The magnitude of computerization, automatization of work tasks, trade, specialization and skill-biased technical change may imply a continued decrease in the demand for low skilled workers and an increasing relevance for cognitive performance and higher education in the coming decades (Elliot 2007; Kavanagh and Doyle 2006; Juhn et al. 1993; Schmidt and Hunter 1998). Using data on the changing importance of skills combined with projections of actual skills, we study plausible future developments in the labour market.
Specific work plans for mid-2011 to mid-2013
To understand the relevance of changes in skills by age and gender, we will examine the changing importance of skills in the labour market. We will extend a model developed by Skirbekk (2008b) to take into account changes in both demand for cognitive abilities and the age-specific supply of skills. We will analyse which kinds of jobs allow older workers to have a better chance in the labour market given the differential development of their abilities, using Germany as an example. The main objective of the analysis is to answer the following questions: What types of skills are used in jobs where older workers are more likely to be hired or fired? In what kinds of jobs are older workers present? In addition, we will develop and investigate various scenarios reflecting plausible future demands based on current observations and reflecting rapid changes in technological development. We will also look at changes in demand for the abilities, including the need to learn, as induced by technological shocks, and how this may affect workers of various ages, cohorts and skill levels.
Measures of cognitive abilities, including verbal and numerical skills, have been found to be increasingly related to work performance over time (e.g., Juhn et al. 1993; Schmidt and Hunter 1998). This can be explained by the fact that the work tasks that have become prevalent over the last few decades, such as manual work, have given way to non-routine interactive and analytical work (Spitz-Oener 2006).
Although the importance of cognitive abilities in the labour market is increasing, little is known about the prevalence of these skills in the coming decades. We will present new skill projections, based on lifetime and generational variation in cognitive skills over time. Important data sources include data on nationally representative ability scores at younger ages, such as FIMS and PISA, and ability scores at older ages, including SHARE and SAGE. We will also use data from the International Adult Literacy Survey (IALS), which describes standardized scores on literacy by gender across all ages (Statistics Canada 2008). The age distributions of literary skills for Canada, Italy, Norway, Switzerland and the United States are available in both the 1994-1998 and the 2003 surveys. As we have multiple waves, we aim to use Age-Period-Cohort (APC) model approaches (see the Methodology section) to identify age, period and cohort changes. These estimates will be used to project the skill composition of the population by age and sex until 2050.
Evidence on age trajectories of cognitive abilities suggests that these abilities tend to follow relatively similar patterns in different cultures and for both genders (Avolio and Waldman 1994; Baltes and Mayer 1999; Deary et al. 2000; Maitland et al. 2000; Park et al. 1996). These documented life cycle changes could be the basis for scenarios regarding future literacy development. Supporting evidence by cohort-age changes in these types of mental abilities, will be presented from new longitudinal data available from the SHARE (Survey of Health, Aging and Retirement in Europe) survey, which presents longitudinal data on age changes among the 50+ population in European countries (Börsch-Supan 2005; Engelhardt et al. 2008).
