Productivity consequences of workforce aging: Stagnation or Horndal effect?

A growing population in which every successive cohort is larger than the preceding one will, in general, be characterized by a youthful, progressive age structure. This relation between cohort growth and age structure is not only valid for the total population but also for the working-age population. Positive cohort growth rates lead to an age profile dominated by young adults, whereas negative cohort growth rates give a dominance of older adults. In this chapter we address the question of how such changes in the age composition of the working-age population affect productivity growth. Two hypotheses are relevant. The first one is based on productivity measurement at the individual level. Here, most studies indicate that labor productivity peaks somewhere between 30 and 50 years of age. This suggests that a primeage workforce would be more productive than either a youngor old-age workforce. The second hypothesis is based on the experience of the Horndal steel plant in central Sweden. Between 1927 and 1952 this plant experienced a mean annual growth rate in productivity of 2.5 percent in spite of the fact that no major investments were undertaken (Genberg 1992). The first economist to discuss the Horndal effect was Erik Lundberg (Lundberg 1961). Later, it formed an important part of Kenneth Arrow’s learning-by-doing argument (Arrow 1962). A key aspect of the Horndal story is that the steelwork had an older workforce. In 1930, more than a third of the workers at the Horndal steelworks were above age 50 years, compared to only one-in-five of all nonagricultural male workers in Sweden. In 1950 almost half of the workers at Horndal were above 50, compared to one-in-four among nonagricul-