69Ìþ«Æ·

This paper focuses on impact of computerization on the work and retirement decisions of the cohort of 51-61 year old individuals who entered the Health and Retirement Study in 1992 and have been followed for next 18 years through 2010. I use data on cognition and detailed occupations in the HRS linked to a measure of occupational computerization from the O*NET data assembled by the Bureau of Labor Statistics. Beginning with Autor et al. (2003), the labor economics literature suggests that advances in computers substitute for the tasks done by many middle-skilled workers and complement those done by high-skilled individuals. Advances in computer technology tend, therefore, to lower the productivity of the middle-skilled and raise the productivity of the high skilled. Older workers face a decision of whether to invest in keeping up with new technology, shifting to another occupation or exiting from full time work into partial or full retirement. I find strong evidence that women and many men retired earlier if they are in computer-intensive occupations while, for other men it appears that computerization does not have a significant effect on retirement. Higher cognition and being in a high wage occupation appears to partially offset retirement incentives of computerization.