Supplement: What Is the Age of Reason?

J!"# $%&%, N!'()* &%-&$

WHAT IS THE AGE OF REASON?

B, S-./0 A123425, J678 C. D3/9:655, X2;/<3 G2=2/>, 28? D2;/? L2/=968*

Introduction

Most U.S. households have accumulated significant middle-age adults make fewer financial mistakes than assets by retirement, but these assets are often ac- younger or older adults. The third section explores companied by significant liabilities. Including net possible policy responses to help older individuals home equity, households with a head age ($-%* had more effectively manage their finances. The final a median net worth of $!)#,*"" in !""%, according section concludes that the best way forward is not to the Survey of Consumer Finances (SCF).& At the yet clear, stressing that further research is needed on same time, the SCF reports that *' percent had debt several key questions. secured by a residential property, !( percent had installment loans, and )% percent carried credit card balances from month to month. Overall, about two- Cognitive Decline Among thirds of these households had at least one form of debt. This brief raises the question of whether older Older Adults households have the ability to manage their increas- A tendency for cognitive ability to decline with age ingly large and complex balance sheets. is evident from both cross-sectional surveys (which The first section of this brief documents the de- look at a population sample at a given point in time) cline in cognitive function that occurs as individuals and from longitudinal surveys (which follow a sample age. The second section describes new evidence from over time). This section summarizes findings from &" different financial transactions indicating that both types of surveys.

* Sumit Agarwal is a senior ﬁnancial economist with the Federal Reserve Bank of Chicago. John C. Driscoll is a senior economist with the Board of Governors of the Federal Reserve System. Xavier Gabaix is the Martin J. Gruber Professor of Finance at New York University. David Laibson is a Harvard College Professor at Harvard University. This brief was adapted from a paper published by Brookings Papers on Economic Activity (Agarwal et al., !""#). Xavier Gabaix and David Laibson acknowledge support from the National Science Foundation (DMS-"$!%$&'). Laibson acknowledges ﬁnancial support from the National Institute on Aging (R"&-AG-"!&($", R"&-AG-&(($, and P)"-AG-"&!'&"). The views expressed in this brief are those of the authors and do not represent the policies or positions of the Board of Governors of the Federal Reserve System, the Federal Reserve Bank of Chicago, or the Center for Retirement Research at Boston College. ! Center for Retirement Research

Cross-sectional Evidence F/1-3< B. R20<9 6C D<.<80/2 28? C618/0/;< I.D2/3.<80 W/076-0 D<.<80/2 /8 07< U8/0

40%*"% !*% F/1-3< @. T3<8?9 /8 C618/0/;< A=/5/0, =, A1< !&% 20%!"% )#% '*" $% !#% &(% Pattern comparison 0%"% %#$(# Matrix reasoning Word recall 71-79%&-%# 80-89'"-'# ≥90#" Spatial relations Age $"% Notes: Cognitive impairment without dementia is deﬁned

Percentile here as a Dementia Severity Rating Scale score of ( to &&. !%#$)& Sources: Plassman et al. (!""%); and Plassman et al. (!""').

&(!" For most tasks, we hypothesize that net performance is hump-shaped with respect to age, with a rise in crystallized intelligence partly oﬀsetting a decline !"#$% !" )" *" $" (" %" '" #" in ﬂuid intelligence. In other words, cognitive performance improves from youth to middle age, at which Chronological age point it peaks before beginning a steady decline. Sources: Salthouse (!""$); and Salthouse !"&" (forthcoming). Consequently, middle-age adults may be at a decision- making sweet spot.

One important pathway that influences perfor- Longitudinal Evidence mance in older adults is neurological problems. The prevalence of both dementia and cognitive impair- The cross-sectional evidence on age-based patterns ) ment without dementia rises rapidly with age. For in cognitive function is confounded by two effects, example, Figure ! shows that dementia in the United which may work in opposite directions: &) cohort States increases from an estimated $ percent of the effects; and !) selection effects. First, in any cross- population at ages %&-%# to )% percent at ages #" and section of subjects, the older subjects not only are above. Similarly, the estimate for a less severe form older, but also were born in different cohorts than of cognitive impairment that does not involve demen- the younger subjects. And these various cohorts may tia rises from &( percent at ages %&-%# to )# percent have different characteristics; for example, people at ages #" and above. All told, about half of adults born in &#!" may generally have lower levels of cogni- in their '"s suffer from either dementia or cognitive tive function than those born in &#$" or &#%" due impairment without dementia. to fewer educational advantages. Second, selection Age-driven declines in fluid intelligence, however, effects result from differences that may occur in the are partly offset by age-related increases in crystallized type of people in each cohort who respond to the intelligence – sometimes called experience or knowl- survey. For example, older adults have relatively more edge. Most day-to-day tasks, such as buying the right health problems (both physical and cognitive), and amount of milk at the grocery store, rely on both fluid the less healthy are likely to drop out of surveys. and crystallized intelligence. Issue in Brief )

Such a pattern would work in the opposite direction The HRS results paint a clear picture of declin- from the cohort effect by raising the cognitive func- ing cognitive function with age. They also suggest tion level of the respondents in older age groups. that selection effects may be more important than In light of these problems, it is useful to analyze cohort effects. Since older cohorts tend to have fewer data that follow individuals longitudinally. The Health educational advantages, cohort effects are predicted to and Retirement Study (HRS) is an excellent source for cause the naive profiles to fall more steeply than the such analysis of cognitive variables.4 Since it began control profiles. Selection effects, in contrast, should in &##!, the HRS has surveyed every two years a cause the naive profiles to fall less sharply than the nationally representative sample of more than !!,""" control profiles, since the individuals with the poorest Americans over the age of $".$ cognitive function tend to exit the sample. Selection Our study examined several HRS questions bias seems to be more important in the HRS data, involving tasks measuring cognitive function. For since our controlled profiles are consistently steeper example, we studied an immediate word recall task in than our naive profiles. which the interviewer reads a list of &" simple nouns, and the respondent is immediately asked to recall as many of them as possible, in any order. At age $&, the Financial Performance and average performance is (.! words out of &". By age #", the average (controlled) performance is )." words Age: The Inverse U-Shape out of &". This section summarizes evidence from our full study We analyzed the responses to the HRS questions showing that the prices people pay in different credit using two parallel tracks. First, we undertook a transactions vary by age, exhibiting an inverse U- “naive” analysis that simply plots mean performance shaped curve pattern. The main example highlighted by age, ignoring the potential role of cohort and selec- below covers credit card balance transfer offers. tion effects. Second, we conducted a “controlled” Other transactions we examined include home equity analysis that traces out the performance trajectory us- loans and lines of credit, car loans, mortgages, and ing only intra-individual differences. Figure ) shows several other types of credit card transactions. In each the results for the word recall task; results for other case, we conducted a regression analysis that identi- questions showed a similar pattern. fies age effects and controls for observable factors that might explain the patterns by age.( The results consistently confirmed an inverse U-shaped pattern. F/1-3< F. R<9-509 6C I..

Some (about one-third) identify this optimal strategy rates paid are lowest for customers in their late *"s immediately, before making any purchases with the or early $"s. Overall, the difference in interest rate new card. Others (slightly more than one-third) never outcomes between those at the peak performance age identify the optimal strategy during the teaser-rate and older ages ranged from about &" to $" basis points. period. The remaining third discover it after one For each transaction studied, we then estimated or more billing cycles as they observe their surpris- the point in the life span at which financial mistakes ingly high interest charges. These borrowers make are minimized. The mean age is $).) years, and the purchases for one or more months and then have a standard deviation calculated by treating each study as “eureka” moment, after which they implement the a single data point is *.) years. optimal strategy.' One possible explanation for the inverse U-shaped Figure * plots the frequency of “immediate eureka pattern of performance is a combination of two age- moments” for each of five age groups. The pattern based effects discussed above: diminishing returns to shows a pronounced inverted U-shape, with adults experience (crystallized intelligence) and the age- age )$-** most likely to adopt the optimal strategy im- based decline in fluid intelligence. Relatively young mediately. Conversely, the corresponding data for the borrowers tend to have low levels of crystallized “no eureka” group have the opposite pattern, indicat- intelligence but a high degree of fluid intelligence, ing that the greatest frequency of confusion occurs whereas older borrowers tend to have high levels of among younger adults and older adults.# crystallized intelligence but relatively lower fluid intelligence.&&

F/1-3< J. P<3:<80 6C B63364<39 W76 I..D<3/<8:< 2 “E-3

*")%&% dling ﬁnancial decisions, a range of policy responses are discussed below, in order from least to most paternalistic. Each approach has pros and cons. )"(%&% Recognizing that strong regulatory interventions have the potential to generate large social beneﬁts but also

!"'%&% large social costs, even the least intrusive approaches should be subjected to a careful cost-beneﬁt analysis.&!

"%& &"% Disclosure

"%%& &'"*+,-+') to !* !$'$+,-+() to )* )$($+,-+)) to ** *$)$+,-+.) to (* Over/012+.$ ($ Legislation to strengthen disclosure requirements has recently been introduced in many different domains, Age Source: Authors’ calculations. including mutual fund fees, *"&(k) fees, and mort- gage origination fees. However, we are skeptical that improved disclosure will be effective in improving Other Financial Choices financial choices. Even for cognitively healthy popula- tions, additional disclosure and consumer educa- The other financial transactions we examined show a tion make surprisingly little difference in financial similar inverted U-shaped pattern with performance choices. For example, in one recent study, employees peaking in middle age.&" For example, with respect with low saving rates were randomly assigned to a to home equity loans and lines of credit, younger and treatment in which they were paid $$" to read a short older consumers have a greater tendency to mis- explanation of their *"&(k) plan, including a calcula- estimate the value of their home, which results in a tion of how much money they would personally gain less favorable interest rate on their loans. With other by taking full advantage of the employer match. Rela- transactions involving credit cards, car loans, and tive to a control group, this group did not significantly mortgages, the results suggest that fees and interest increase its average *"&(k) saving rate.&) Issue in Brief $

Financial “Driving Licenses” fects. An alternative approach would require that new financial products obtain explicit regulatory approval Another set of proposals would require that individu- before being marketed. als pass a “license” test before being allowed to make Either approach would be socially costly, particu- nontrivial financial decisions, such as opting out of &* larly the explicit approval approach. Introducing a “safe harbor” investment products. Such proposals regulatory regime would delay the release of new would need to overcome several logistical problems. products, increase costs for financial services firms, Can a test be devised that reliably separates qualified and discourage innovation. But this approach could from unqualified investors, without generating too also prevent the marketing of socially undesirable many false negatives or false positives? Can it be products. The net social benefit is not easy to evaluate. administered at a reasonable social cost? Who would be required to take the test? Would such a test be politically feasible if it primarily targeted older adults? Conclusion What would be the impact on older individuals themselves? Older adults experience substantial declines in cognitive function over time. And evidence indicates Mandatory Advance Directives that, after peaking in middle age, the ability to make effective financial decisions declines. In response to One direct way to address the impact of cognitive this problem, several policy approaches are possible decline on financial decision making would be to and government intervention is probably desirable, require older adults to put in place a financial advance although the ideal form of intervention remains un- directive before reaching a certain age so that the clear. Economic behavior among older adults is still management of their assets could be transferred to poorly understood. Moreover, even if older adults are a third party in the event of their incapacity. Entirely making substantial financial mistakes, it is not clear new legal protections might also be created. For what a well-intentioned policymaker should do. example, a fiduciary could be appointed to approve all Before the best solutions can be identified, more “significant financial transactions” involving the prin- research – including field experiments – is needed. cipal’s funds after the principal reaches a designated Researchers and policymakers should consider sev- age. As an alternative to a fiduciary-based model, eral questions as they wrestle with the issue. These the principal could place his assets in a safe harbor questions include the magnitude and prevalence of (which would contain certain restrictions on the types losses due to poor financial decision making; which of asset holdings allowed and on draw-down rates). demographic characteristics predict poor decision Mandating advance directives would pose several making; the extent to which people anticipate or problems. First, it might be perceived by some older recognize their own cognitive decline; the efficacy adults as an unfair restriction targeted against them. of financial education; the efficacy of third parties Second, the imposition of a fiduciary would create such as advisors or family members; and the market transaction costs. Third, any attempt to define a safe response to the current situation. harbor would be politically contentious, doubtless giv- ing rise to a great deal of lobbying. An independent agency would probably be needed to partly insulate the safe harbor regulations from political pressure.KL

Regulatory Approval

Instead of primarily targeting individual investors, regulations could instead target the financial products themselves. One such regime would mimic the regulatory model currently used for nutritional supplements: new financial products would be allowed in the market without specific formal approval in advance but would be monitored for adverse ef- ( Center for Retirement Research

Endnotes

& This figure excludes defined benefit pension &" Our analyses of these other transactions rely on wealth. data from various proprietary data sets from financial institutions. These data sets are described in Agarwal ! Cattell (&#'%); Salthouse (!""$); and Salthouse et al. (!""#). (!"&" forthcoming). && Alternatively, the inverse U-shaped pattern could ) Plassman et al. (!""%); Plassman et al. (!""'); Ferri also be influenced by cohort or selection efforts. In et al. (!""$); and Fratiglioni, De Ronchi, and Agüero- fact, we find no evidence for either cohort or selection Torres (&###). effects that could explain our results, but our data do not allow us to definitively rule them out. * Ofstedal, Fisher, and Herzog (!""$); and McArdle, Smith, and Willis (!"&" forthcoming). &! For a more expansive discussion of possible policy responses, see Agarwal et al. (!""#). $ See Ofstedal, Fisher, and Herzog (!""$) for a com- plete description of the cognitive scales in the HRS. &) See Choi, Laibson, and Madrian (!""'). Examples of other studies that show a similar lack of effective- ( Our analysis is part of a recent literature that stud- ness from providing more information can be found ies the effects of aging and cognitive function on the in Choi, Laibson, and Madrian (!""$); Madrian and use of financial instruments (see, for example, Willis, Shea (!""&); and Beshears et al. (!""#). !""%; and McArdle, Smith, and Willis !"&" forthcoming), which in turn is part of a broader literature on &* Alesina and Lusardi (!""(). household finance (Campbell, !""(). See Agarwal et al. (!""#) for more details on the literature. &$ The financial reform legislation currently under consideration includes a new consumer financial % We use a proprietary panel data set with data from protection bureau within the Federal Reserve. several large financial institutions, later acquired by a single financial institution that made balance transfer offers nationally. The offers were not made condi- tional on closing the old credit card account. The data set contains information on &*,%#' individuals who accepted such balance transfer offers over the period January !""" through December !""!.

' We thank Robert Barro of Harvard University for drawing our attention to this type of potentially tricky ﬁnancial product. We also note that changes in regulation proposed in May !""' by the Federal Reserve, the National Credit Union Administration, and the Oﬃce of Thrift Supervision would forbid banks from applying payments solely to transferred balances.

# We also check for the possibilities that the relatively old and the relatively young might have lower levels of debt or less access to credit than the middle-aged. We ﬁnd that neither credit card debt nor the number of open credit cards varies in economically or statistically signiﬁcant ways with age. Issue in Brief %

References

Agarwal, Sumit, John C. Driscoll, Xavier Gabaix, Those Taught: Financial Education in the Work- and David I. Laibson. !""#. “The Age of Reason: place.” Working Paper. Chicago, IL: University of Financial Decisions Over the Life Cycle and Chicago. Implications for Regulation.” Brookings Papers on Economic Activity !: $&-&&%. Washington, DC: McArdle, John J., James P. Smith, and Robert Willis. Brookings Institution Press. !"&" (forthcoming). “Cognition and Economic Outcomes in the Health and Retirement Survey.” Alesina, Alberto, and Annamaria Lusardi. !""(. In Explorations in the Economics of Aging, edited by “Driving Your Financial Investment? Perhaps a David Wise. Cambridge, MA: National Bureau of License Is Needed.” Vox of Dartmouth XXV(!), Economic Research. July !*. Ofstedal, Mary Beth, Gwenith G. Fisher, and A. Beshears, John, James J. Choi, David Laibson, and Regula Herzog. !""$. “Documentation of Cog- Brigitte C. Madrian. !""#. “How Does Simpli- nitive Functioning Measures in the Health and ﬁed Disclosure Aﬀect Individuals’ Mutual Fund Retirement Study.” HRS Documentation Report Choices?” Working Paper &*'$#. Cambridge, DR-""(. (March) Ann Arbor, MI: University of MA: National Bureau of Economic Research. Michigan.

Campbell, John Y. !""(. “Household Finance.” Jour- Plassman, Brenda L., Kenneth M. Langa, Gwenith G. nal of Finance (&(*): &$$)-&("*. Fisher, Steven G. Heeringa, David R. Weir, Mary Beth Ofstedal, James R. Burke, Michael D. Hurd, Guy G. Potter, Willard L. Rodgers, David C. Stef- Cattell, Raymond B., ed. &#'%. Intelligence: Its Struc- fens, Robert J. Willis, and Robert B. Wallace. !""%. ture, Growth and Action. New York: Elsevier. “Prevalence of Dementia in the United States: The Aging, Demographics and Memory Study. Neuro- Choi, James J., David Laibson, and Brigitte C. epidemiology !#: &!$-&)!. Madrian. !""$. “Are Empowerment and Educa- tion Enough? Underdiversiﬁcation in *"&(k) Plassman, Brenda L., Kenneth M. Langa, Gwenith G. Plans.” Brookings Papers on Economic Activity Fisher, Steven G. Heeringa, David R. Weir, Mary !: &$&-#'. Beth Ofstedal, James R. Burke, Michael D. Hurd, Guy G. Potter, Willard L. Rodgers, David C. Stef- Choi, James J., David Laibson, and Brigitte C. fens, John J. McArdle, Robert J. Willis, and Robert Madrian. !""'. “$&"" Bills on the Sidewalk: Sub- B. Wallace. !""'. “Prevalence of Cognitive Impair- optimal Investment in *"&(k) Plans.” Working ment without Dementia in the United States.” Paper &&$$* (revised). Cambridge, MA: National Annals of Internal Medicine &*'((): *!%-)*. Bureau of Economic Research. Salthouse, Timothy A. !""$. “Eﬀects of Aging on Rea- Ferri, Cleusa P., Martin Prince, Carol Brayne, soning.” In The Cambridge Handbook of Thinking Henry Brodaty, Laura Fratiglioni, Mary Ganguli, and Reasoning. Cambridge, England: Cambridge Kathleen Hall, Kazuo Hasegawa, Hugh Hendrie, University Press. Yueqin Huang, Anthony Jorm, Colin Mathers, Paulo R. Menezes, Elizabeth Rimmer, and Mar- Salthouse, Timothy A. !"&" (forthcoming). “Executive cia Scazufca. !""$. “Global Prevalence of De- Functioning.” In Cognitive Aging: A Primer, !nd The Lancet mentia: A Delphi Consensus Study.” ed., edited by D. C. Park and N. Schwarz. London, )(((#$")): !&&!-&%. UK: Psychology Press.

Fratiglioni, Laura, Diana De Ronchi, and Hella University of Michigan. &##!-!""(. Health and Retire- Agüero Torres. &###. “Worldwide Prevalence and ment Study. Ann Arbor, MI. Incidence of Dementia.” Drugs & Aging &$($): )($-%$. Willis, Robert J. !""%. “Cognitive Economics and Hu- man Capital.” Presidential Address to the Society Madrian, Brigitte C. and Dennis F. Shea. !""&. of Labor Economists, Chicago, May. “Preaching to the Converted and Converting

About the Center Affiliated Institutions The Center for Retirement Research at Boston Col- The Brookings Institution lege was established in &##' through a grant from the Massachusetts Institute of Technology Social Security Administration. The Center’s mission Syracuse University is to produce first-class research and forge a strong Urban Institute link between the academic community and decision makers in the public and private sectors around an Contact Information issue of critical importance to the nation’s future. Center for Retirement Research To achieve this mission, the Center sponsors a wide Boston College variety of research projects, transmits new findings to Hovey House a broad audience, trains new scholars, and broadens &*" Commonwealth Avenue access to valuable data sources. Since its inception, Chestnut Hill, MA "!*(%-)'"' the Center has established a reputation as an authori- Phone: ((&%) $$!-&%(! tative source of information on all major aspects of Fax: ((&%) $$!-"&#& the retirement income debate. E-mail: [email protected] Website: http://crr.bc.edu

The Center for Retirement Research thanks AARP, InvescoSM, LPL Financial, MetLife, Nationwide Mutual Insurance Company, Prudential Financial, State Street, TIAA-CREF Institute, and T. Rowe Price for support of this project.

© !"&", by Trustees of Boston College, Center for Retire- The research reported herein was supported by the Center’s ment Research. All rights reserved. Short sections of Partnership Program. The ﬁndings and conclusions ex- text, not to exceed two paragraphs, may be quoted without pressed are solely those of the authors and do not represent explicit permission provided that the authors are identiﬁed the opinions or policy of the partners or the Center for and full credit, including copyright notice, is given to Trust- Retirement Research at Boston College. ees of Boston College, Center for Retirement Research.