RETIREMENT RESEARCH November 2016, Number 16-18 COGNITIVE AGING AND ABILITY TO WORK By Anek Belbase and Geoffrey T. Sanzenbacher* Introduction Working longer is an effective way to boost individu- second section explains why declining fluid intel- als’ retirement security. Thus, understanding who ligence tends not to impede work ability. The third can work longer and who may struggle is a key issue section looks at the minority of workers who may for researchers and policymakers. Some studies find struggle to remain productive and why. The final sec- that age-related declines in physical abilities can limit tion concludes that experience helps many workers in those in physically demanding jobs from working into skilled jobs stay productive and workers in less skilled their late 60s.1 But the effects of changes in cognitive jobs might have more fluid intelligence than their abilities on work have received less attention. At first job requires. However, two groups are vulnerable to glance, it appears that a decline in “fluid” intelligence age-related decline: those in jobs where accumulated – the capacity to process new information – and an knowledge cannot offset demand for fluid intelligence apparent relationship between fluid intelligence and and those who experience cognitive impairment. job achievement could pose a barrier to working   longer. However, “crystallized” intelligence – accu- mulated knowledge – increases with age, and cogni- Age and Job Performance tive reserves can offer spare capacity against declining fluid intelligence. As a result, studies comparing the Decades of research show that age does not reliably productivity of young and old workers find that age is explain the variation in productivity among workers a crude and unreliable predictor of performance.2 ages 20-65. While individual studies – which typi- This brief – the second in a series of three – re- cally examine a subset of workers in specific occupa- views the research literature to assess how cognitive tions – can show relatively strong correlations (both aging affects the ability to work during ages 50-70. positive and negative) between age and productivity, The first brief provided a primer on cognitive aging meta-studies that aggregate data across the individual and the third brief will examine how it affects retirees’ studies and apply results to the whole population find ability to manage their money from ages 70-90. practically no correlation. Several of these meta- The discussion proceeds as follows. The first studies have examined the relationship between age section documents that age is not generally related and productivity by standardizing variables across the to productivity across a variety of occupations. The samples used in prior research and analyzing the re- sulting “super” sample (see Table 1 on the next page). * Anek Belbase is a research fellow at the Center for Retirement Research at Boston College (CRR). Geoffrey T. Sanzen- bacher is a research economist at the CRR. 2 Center for Retirement Research Table 1. Correlation between Age and How Do Most Workers Stay Productivity, by Subjective and Objective Measures Productive? Correlation between Year Authors No. of age and productivity Two cognitive factors explain why most workers samples Subjective Objective remain productive despite a decline in fluid intel- 1986 Waldman & Avolio 40 -0.14 0.27 ligence: 1) crystallized intelligence – knowledge that 1989 McEvoy & Cascio 96 0.06 0.06 accumulates with age – can offset declines in fluid intelligence; and 2) cognitive reserves, or fluid intelli- 2003 Sturman 167 0.02 0.08 gence that exceeds job demands, can provide workers 2008 Ng and Feldman 118 0.02 0.03 with a buffer. The first factor tends to particularly help skilled workers, while the second factor tends to Source: Authors’ review of the literature. benefit unskilled workers more. For example, Sturman (2003) collected 115 stud- Benefits of Crystallized Intelligence ies with 167 samples that measured productivity for workers ages 17-64. These samples were catego- Older workers can make up for declining fluid intel- rized by the type of performance measure, whether ligence by drawing on their crystallized intelligence. the samples included job tenure, and whether they When these workers were younger, their fluid intelli- contained workers in jobs with relatively complex gence was at its peak, which made it easier to acquire requirements. After aggregating the individual job-related skills. For example, pharmacists need a samples based on these dimensions, and controlling great deal of fluid capacity in pharmacy school and for tenure and other factors, Sturman found practi- early in their careers to learn the facts, concepts, and cally no correlation between age and job performance. procedures needed to do their jobs well. Their fluid This finding applied using either subjective evalua- intelligence helps them quickly build this storehouse tions of employees’ work by managers and coworkers of knowledge. As pharmacists grow older and more or objective measures of their work based on quantity experienced in their jobs, their crystallized intelli- and quality of output. Using a similar methodology, gence offsets declines in fluid intelligence because: Ng and Feldman (2008) also found a negligible cor- 1) the amount of information they need to learn relation between age and productivity.3 declines, reducing the need for fluid intelligence; and Since most productivity studies compare young 2) the steady accumulation of knowledge over time and old workers at one point in time, the minimal makes up for their loss of fluid capacity. Figure 1 overall changes in productivity may simply be due to cohort differences (like older workers having a stronger work ethic) or selection biases (like produc- Figure 1. Ability to Finish New York Times Cross- tive workers sticking around while unproductive word Puzzle by Age, in Standard Deviation Units workers retire). But studies of cognitive ability that 1 follow workers over time also show that productivity is maintained with age, even among workers who experience decreases in their capacity to process new 0 information.4 So, it appears that the majority of older Z-score workers can remain productive despite age-related losses in fluid intelligence. The questions are: how -1 do these workers maintain productivity and is anyone left behind? -2 20 30 40 50 60 70 80 Age Source: Salthouse (2012). Issue in Brief 3 provides an everyday example of how performance of Table 2. Fluid and Crystallized Intelligence a task that requires both fluid reasoning ability and for Participants in Air Traffic Controller knowledge – completing a crossword puzzle – can Simulation improve over time as knowledge offsets declining fluid intelligence. Fluid Crystallized Type of participant intelligence intelligence A Cognitive Buffer Old non-controller Low Low Old controller Low High While skilled workers can use knowledge to offset declining fluid intelligence, many workers – particu- Young non-controller High Low larly those in jobs that involve simple or routine tasks Young controller High High – may also have more fluid intelligence than their Source: Nunes and Kramer (2009). job requires. For example, many clerical positions require workers to perform routine activities that be- come automatic with time, leaving such workers with only needed half the time – 10 seconds – but the enough fluid capacity in “reserve” to act as a buffer younger controllers only needed 6 seconds. Inter- against decline. estingly, the older controllers performed no better than younger individuals who were not controllers (see   Figure 2). So, because of the high fluid intelligence Why Are Some Workers More demands of the task, old controllers – even with their decades of experience – could not out-perform young Vulnerable to Decline? people with no background in the field. Not surpris- ingly, then, air traffic controllers are a prime example While most workers can stay sharp on their jobs as of an occupation in which cognitive aging often leads they age, two types of workers may struggle to main- to early retirement; controllers employed by the Fed- tain their productivity: 1) those in high-skill occupa- eral Aviation Authority are actually required to retire tions with intense demand for fluid intelligence; and at age 56.6 2) those who experience unusually severe cognitive decline. Figure 2. Problem-Solving Time for Old and Young Air-Traffic Controllers and Non-Controllers High Fluid Intelligence Requirements 30 Problem-solving time (seconds) The interplay of fluid and crystallized intelligence is evident from a study of simulated performance of a 20 typical task faced by air traffic controllers. The study 20 included four groups of participants: 1) old non-con- trollers; 2) old controllers; 3) young non-controllers; 10 10 and 4) young controllers. Table 2 shows how each 10 group rates on fluid and crystallized intelligence rela- 6 tive to the task. To assess each group’s performance, research- 0 ers measured how long it took them to re-route Old non- Old Young non- Young two planes on a collision course, so the quicker the controller controller controller controller solution, the better. Not surprisingly, the older non- controllers had the worst performance, requiring 20 Source: Nunes and Kramer (2009). seconds to complete the task. The older controllers 4 Center for Retirement Research Another subset of workers who are vulnerable Conclusion to declining fluid intelligence are those who end up switching occupations. In this case, job-switchers Working longer is a powerful way to improve retire- have no crystallized intelligence and must start from ment preparedness, so understanding individuals’ ca- scratch to develop the specific knowledge and skills pacity for stretching out their work years is important. that are needed. Such workers could struggle to At first glance, recent research suggests cognitive remain productive compared to younger workers be- aging might hinder working longer due to declin- cause age erodes their fluid intelligence and, thus, the ing fluid intelligence. However, studies comparing capacity to learn new skills, particularly if the skills the productivity of workers in their 20s to workers in are unrelated to their existing knowledge base.7 their 30s, 40s, 50s and 60s find that age is unrelated to performance in most occupations.13 As people age, Cognitive Impairment improvements in knowledge appear to largely offset declines in fluid intelligence, and the amount of fluid Workers who begin to develop a cognitive impairment capacity that most workers have through their late 60s in their 60s may not be able to work as long as they seems to offer a sufficient buffer against any declines. planned due to a declining ability to meet job de- However, some groups of older workers are mands. A cognitive impairment typically starts out as vulnerable to cognitive decline. Workers in jobs that a mild, and often temporary, condition that primarily require a high degree of fluid intelligence (which can affects fluid intelligence, judgment, and reasoning include those who end up shifting careers) or who ability. However, over half the cases of mild impair- experience a cognitive impairment are likely to have ment progress to dementia, which erodes all cognitive trouble extending their worklives. As policymakers functions irreversibly.8 consider ways to encourage working longer, they may While the incidence of Alzheimer’s disease (the want to pay close attention to the potential impact of most common type of dementia) in one’s 50s and any proposed changes on such workers, as well as the early 60s is low – less than 4 percent of people under effects on disability and unemployment programs age 65 – it rises to 15 percent of 65-74 year olds.9 that could see increased demand. Dementia develops gradually, so many who end up having it in their 70s may already have developed significant symptoms in their 60s.10 This situation can go unnoticed because people who develop it typi- cally lack awareness of their declining abilities.11 The increasing risk of dementia with age suggests two types of workers would benefit from regular screen- ing after age 65: workers in cognitively demanding occupations; and those in jobs where errors could significantly harm others, like surgeons. Screening could protect the public from harm and potentially al- low the cognitively impaired to qualify for retirement under the disability program.12 Issue in Brief 5 Endnotes 1 For research on physical abilities and aging, see 10 For example, see Manly et al. (2008) and Katz et Holden (1988); Lund, Iversen, and Poulsen (2001); al. (2012). Important factors affecting reporting of and Karpansalo et. al. (2002). the prevalence of mild cognitive impairment include the type of assessment used, the type of dataset, and 2 Kuncel, Ones, and Sackett (2010). cohort effects. 3 While these meta-analyses do not show a general 11 Okonkwo et al. (2008) correlation between age and productivity, they do find correlations when skilled and unskilled work- 12 Adler and Constantinou (2008). ers are examined separately. For unskilled workers, performance is positively correlated with age early 13 Kuncel, Ones, and Sackett (2010). in a worker’s career, until about 30 or 40, and nega- tively correlated with age thereafter (e.g., Sturman 2003). One reason might be that physical capacity could be more important than cognitive capacity in determining the productivity of unskilled workers. In contrast, the productivity of skilled workers continues to be positively correlated with age throughout their careers, including their 50s and 60s. This finding is particularly true when other age-related factors like motivation and tenure are taken into account; see Ng and Feldman (2008). Note that while Ng and Feld- man examine workers between 17 and 59, other stud- ies suggest that performance declines in one’s 60s relative to peak performing years, even among skilled workers. 4 Salthouse (2012). 5 Salthouse (2012). 6 Nunes and Kramer (2009). Controllers with excep- tional skills and experience are allowed to work up to age 61. 7 Schaei and Willis (2016). 8 Alzheimer’s Association (2014). 9 Alzheimer’s Association (2014). 6 Center for Retirement Research References Adler, Robert G. and Conn Constantinou. 2008. McEvoy, Glenn M. and Wayne F. Cascio. 1989. “Cu- “Knowing – or Not Knowing – When to Stop: mulative Evidence of the Relationship Between Cognitive Decline in Ageing Doctors.” The Medical Employee Age and Job Performance.” Journal of Journal of Australia 189 (11): 622-624. Applied Psychology 74(1): 11-17. Alzheimer’s Association. 2014. “Alzheimer’s Disease Ng, Thomas W.H. and Daniel C. Feldman. 2008. “The Facts and Figures.” Alzheimer’s and Dementia Relationship of Age to Ten Dimensions of 10(2). Job Performance.” Journal of Applied Psychology 93(2): 392-423. Holden, Karen C. 1988. “Physically Demanding Oc- cupations, Health, and Work After Retirement: Nunes, Ashley, and Arthur F. Kramer. 2009. “Ex- Findings from the New Beneficiary Survey.” Social perience-Based Mitigation of Age-Related Per- Security Bulletin 51(3): 3-15. formance Declines: Evidence from Air Traffic Con- trol.” Journal of Experimental Psychology: Applied Karpansalo, Minna, Pirjo Manninen, Timo A. Lakka, 15(1): 12-24. Jussi Kauhanen, Rainer Rauramaa, and Jukka T. Salonen. 2002. “Physical Workload and Okonkwo, Ozioma C., Virginia G. Wadley, H. Randall Risk of Early Retirement: Prospective Population- Griffith, Katherine Belue, Sara Lanza, Edward Y. Based Study among Middle-Aged Men.” Journal of Zamrini, Lindy E. Harrell, John C. Brockington, Occupational and Environmental Medicine 44(10): David Clark, and Rema Raman. 2008. “Awareness 930-939. of Deficits in Financial Abilities in Patients with Mild Cognitive Impairment: Going Beyond Self- Katz, Mindy J., Richard B. Lipton, Charles B. Hall, Informant Discrepancy.” The American ournal of J Molly E. Zimmerman, Amy E. Sanders, Joe Ver- Geriatric Psychiatry 16(8): 650-659. ghese, Dennis W. Dickson, and Carol A. Derby. 2012. “Age-Specific and Sex-Specific Prevalence Salthouse, Timothy. 2012. “Consequences of Age- and Incidence of Mild Cognitive Impairment, Related Cognitive Declines.” Annual Review of Dementia, and Alzheimer Dementia in Blacks and Psychology 63(1): 201-26. Whites: A report from the Einstein Aging Study.” Alzheimer Disease and Associated Disorders 26(4): Schaie, K. Warner, and Sherry L. Willis. 2016. Hand- 335-343. book of the Psychology of Aging: Eighth Edition. Boston, MA: Academic Press. Kuncel, Nathan R., Deniz S. Ones, and Paul R. Sack- ett. 2010. “Individual Differences as Predictors of Sturman, Michael C. 2003. “Searching for the Invert- Work, Educational, and Broad Life Outcomes.” ed U-Shaped Relationship Between Time Personality and Individual Differences 49(4): 331- and Performance: Meta-Analyses of the Experi- 336. ence/Performance, Tenure/Performance, and Age/Performance Relationships.” Journal of Lund, Thomas, Lars Iversen, and Kjeld B. Poulsen. Management 29(5): 609-640. 2001. “Work Environment Factors, Health, Lifestyle and Marital Status as Predictors of Job Waldman, David A., and Bruce J. Avolio. 1986. “A Change and Early Retirement in Physically Heavy Meta-analysis of Age Differences in Job Occupations.” American Journal of Industrial Medi- Performance.” Journal of Applied Psychology 71(1): cine 40(2): 161-169. 33-38. Manly, Jennifer J., Ming-X. Tang, Nicole Schupf, Yaakov Stern, Jean-Paul G. Vonsattel, and Richard Mayeux. 2008. “Frequency and Course of Mild Cognitive Impairment in a Multiethnic Commu- nity.” Annals of Neurology 63(4): 494-506. RETIREMENT RESEARCH About the Center Affiliated Institutions The mission of the Center for Retirement Research The Brookings Institution at Boston College is to produce first-class research Syracuse University and educational tools and forge a strong link between Urban Institute the academic community and decision-makers in the public and private sectors around an issue of criti- cal importance to the nation’s future. To achieve Contact Information Center for Retirement Research this mission, the Center sponsors a wide variety of Boston College research projects, transmits new findings to a broad Hovey House audience, trains new scholars, and broadens access to 140 Commonwealth Avenue valuable data sources. Since its inception in 1998, the Chestnut Hill, MA 02467-3808 Center has established a reputation as an authorita- Phone: (617) 552-1762 tive source of information on all major aspects of the Fax: (617) 552-0191 retirement income debate. E-mail: crr@bc.edu Website: http://crr.bc.edu © 2016, by Trustees of Boston College, Center for Retirement Research. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that the authors are identified and full credit, including copyright notice, is given to Trustees of Boston College, Center for Retirement Research. The research reported herein was performed pursuant to a grant from the U.S. Social Security Administration (SSA) funded as part of the Retirement Research Consortium. The opinions and conclusions expressed are solely those of the authors and do not represent the opinions or policy of SSA or any agency of the federal government. Neither the United States Govern- ment nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of the contents of this report. Reference herein to any specific commercial product, process or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply endorsement, recommendation or favoring by the United States Government or any agency thereof.