the time, place, and approach, studies indicate that anywhere from half to virtually all of the growth in output per hour worked or per capita income is due to advances in knowledge, and that the balance is due to increases in capital and other inputs per unit of labor.
RESEARCH ON THE IMPACTS OF AGING
AT THE INDIVIDUAL LEVEL
There is a substantial literature on behavioral measures of productivity or proxies for it over the life cycle. Studies include psychometric ones (such as ones that measure verbal or quantitative reasoning), ratings (such as those of supervisors), productivity measures (such as in piece rates or baseball scores), and statistical studies at the company level. Useful surveys are those by Prskawetz and Lindh (2006) and Skirbekk (2004).
On the whole, the literature on individual productivity measures shows great diversity across age, individuals, and measures. As one of the pioneers in the field, Salthouse (1991) found that the relationship between age and cognition varies considerably across different cognitive tests.
Many psychometric measures show a clear relationship to age. Verhaegen and Salthouse (1997) provide a meta-analysis of cognitive studies (p. 246). They compare the performance of individuals over and under 50 years of age and conclude as follows:
[M]eta-analyses of correlations between age and different measures of cognition revealed that the age relations in this literature are somewhat stronger with measures of speed than with measures of reasoning, spatial abilities, and working and episodic memory and that primary memory has a smaller age relation than do the latter variables.
Avolio and Waldman (1994) examine a series of studies that measure work-related skills using the General Aptitude Test Battery (GATB) for more than 25,000 workers from 16 to 74 years old. They conclude that age accounted for a relatively small percentage of the variance in ability test scores once experience, education, and occupational type were controlled. Differences in performance across age groups were relatively small until at least age 65. However, unlike the earnings data reviewed below, job experience has little value in predicting the maintenance of abilities over the long run except for complex jobs.
Literature using other metrics for individual productivity also shows divergent results by age and metric. One survey concludes that supervisors’ ratings typically do not find any clear systematic relationship between age and productivity. The evidence on productivity as measured by piece rates is mixed.
An important new approach is cross-sectional employer-employee