Research DialoguesIssue Number 60

August 1999

A publication of the TIAA-CREF Institute

In this issue:

Introduction

Previous Studies of RetirementIncome Choices at TIAA-CREF

The Retirement Decision

Income Options and the Choice toBegin Receiving Income

Changes in Annuitization Rates

Conclusions

Appendix: Calculating the OptionValue of Postponing Annuitization

The Retirement Patterns and Annuitization Decisions of a Cohort of TIAA-CREF Participants

© 1999 TIAA-CREF

pants holding relatively more equities intheir retirement portfolios) also have beenmost likely to delay the annuitizationdecision.

Introduction

Over the last ten years, TIAA-CREF hasgreatly expanded participants’ options forreceiving retirement income from theiraccumulated assets. Participants can usecurrently available options singly or incombination to customize many aspectsof their retirement income, includingtheir asset allocation, the amount of peri-odic income they expect to receive, thetiming of income distributions, the scopeof the income guarantee, as well as theamount of wealth available for bequests.(Of course, all participants are still con-strained by their available resources, legalrequirements, and the performance of thefinancial markets.)

Previous research has shown that as thenumber of TIAA-CREF income optionshas expanded, there has been an overalldecline in the number of TIAA-CREF participants choosing to begin lifetimeannuity income from their plan balances(King, 1996; TIAA-CREF, 1998). Severalexplanations for this decline have beenoffered, such as the following:

● Participants may be choosing towork longer and retire later than inprevious years. Evidence from theBureau of Labor Statistics indicatesthat the rate of decline in the aver-age age of retirement in the UnitedStates may have slowed or flattenedout in the late 1980s and early1990s. Moreover, those who retire

early are increasingly continuing towork after “retirement” (Gendell,1998; Herz, 1995). Additionally,when age-based mandatory retire-ment ended for tenured faculty in1994, many TIAA-CREF partici-pants for the first time had theopportunity to postpone retirementbeyond age 70. If relatively fewerolder individuals are retiring nowthan in the past, one would expectfewer annuitizations to occur.

● Before TIAA-CREF introduced theTransfer Payout Annuity (TPA) andthe Minimum Distribution Option(MDO) contract in 1991, some par-ticipants may not have wanted topurchase a life annuity when theyreached age 701/2 but had no otheralternative. Many participants maynow be using the TPA, MDO, andother new options as alternatives tothe life annuity.

● As a result of the unexpected andunprecedented rise of the stockmarket in recent years, many retir-ing and retired TIAA-CREF partic-ipants have experienced largeincreases in the value both of theirretirement plan assets and, impor-tantly, of their other financial assets.Because these individuals are nowwealthier than they may have antic-ipated, it is possible that their needor desire to begin retirementincome streams may have beenreduced. Many may now wish todefer, as long as possible, the tax lia-bilities that withdrawal of retire-

This issue of Research Dialogues exam-ines the retirement patterns and annuitiza-tion decisions of a large cohort ofTIAA-CREF participants (more than200,000 individuals) as they passedthrough retirement ages during 1987–1996. We use these data to illustrate,analyze, and attempt to explain the declinein the number of annuitizations amongTIAA-CREF participants over thisperiod. Along the way, we present statisticsshowing trends in retirement behavior ofthis cohort of TIAA-CREF participants.We also briefly describe and discuss someof the considerations that might be affectingparticipants’ decisions regarding the annu-itization of accumulated retirement wealth,focusing on the effects of increased retirementwealth. (An appendix to this issue exam-ines, in detail, some of the trade-offsinherent in delaying the start of life-annuity income.) Finally, we assess thehypothesis that those whose wealth hasincreased the most over this period (partici-

18429AA 3/9/00 7:01 AM Page 1

Page 2 Research Dialogues

ment assets will impose, preferringinstead to use other assets to fund atleast the first few years of retirement.Increases in wealth may also havepushed a large number of individualsover a “wealth threshold,” enablingthem to consider bequeathing a sub-stantial portion of their wealth totheir heirs or other beneficiaries.

In this article, we first briefly discussprevious analyses of TIAA-CREF partici-pants’ income choices and explain howthe current study’s cohort approach rep-resents an improvement in methodology.We then review the various explanationsgiven for changes in income-choicebehavior by examining the retirementpatterns and retirement income electionsmade from 1987 to 1996 by the cohort ofTIAA-CREF participants who werebetween the ages of 49 and 71 in 1986.

We focus on three factors in particular:first, we estimate age-specific retirementrates among these individuals over thisperiod, and we present and analyze thesedata in detail. Second, we examine changesin the fraction of individuals who retire atspecific ages and begin to receive incomefrom a life-annuity product. Finally, weexamine the relationship between assetallocation and the annuitization decision.We use the data to evaluate the hypothesisthat the “annuitization rate” has declinedthe most among those for whom wealthhas increased the most (those with greatestexposure to the stock market).

Previous Studies of Retirement Income Choices at TIAA-CREF

Previous reports from TIAA-CREF pro-vide a great deal of information regardingtrends in the income options that TIAA-CREF participants have chosen over theyears. In Research Dialogues, No. 48(King, 1996), we reviewed the choicesmade by TIAA-CREF participants whobegan to receive income in each year from1978 to 1994. The report documentedthat men beginning a life-annuityincome stream have increasingly beenmore likely to choose two-life annuities,while women have been more likely to

choose a single-life annuity. It also docu-mented the decline that has occurredsince 1978 in the relative number of first-time annuitizations at age 65 amongTIAA-CREF participants. In 1978, ofthose starting a life annuity, 41.7 percentstarted at age 65. Since then, individuals

purchasing an annuity for the first timehave been more likely to do so at agesboth later and earlier than 65. In 1994,only 20.8 percent of life annuities issuedwere started at age 65. (As of 1997, 18.5percent of first-time annuitizations wereby individuals age 65.)

The September 1998 issue of theBenefit Plan Counselor reported addition-ally that the number of TIAA-CREF par-ticipants choosing to convert theiraccumulated assets into any type of peri-odic income stream has declined in recentyears. In 1997, those choosing to start alife annuity (11,700) were roughly 30percent fewer than in 1988 (17,100). Atthe same time, the number of participantsover 55 with unannuitized TIAA-CREFassets increased by 47 percent from 1992to 1997, while the number age 65 andolder has risen 72 percent. The article alsoreported the increasing use of TIAA-CREF’s nonannuity income options bythose who did elect to start receiving dis-tributions: While 80 percent of thosestarting an income stream in 1987 con-verted their entire accumulation to a lifeannuity, only 60 percent did so in 1997.

Both these articles suggest that therehave been large changes over time in therate at which individuals are retiring andstarting to use TIAA-CREF’s annuityand other income options. However,because these studies use cross-sectionaldata, measurement of the size of suchchanges is not possible. For example,while the number of participants over age

55 with unannuitized assets has grownby 47 percent since 1992, it is also thecase that the population of all TIAA-CREF participants with unannuitizedassets has increased by roughly 42 per-cent (from 1.2 million in 1992 to 1.7million in 1997). This raises obvious

questions: How much of the 47 percentincrease among the over-55 population isa result of population growth, and howmuch is a result of the decision of partic-ipants older than age 55 not to retire orbegin retirement income?

We attempt to answer these questionsby using different data and taking a dif-ferent approach from that of the previousstudies. Instead of examining the choicesof only those who decided to begin anincome stream in each year, we examinethe behavior of a large group, or cohort,of TIAA-CREF participants over a periodof time. In particular, we examine thechoices made by the entire 1986 popula-tion (over 200,000 individuals) of TIAA-CREF participants who were between 49and 71. The fundamental difference ofthis approach is that it explicitly recog-nizes that participants’ decisions not tobegin an income stream are as importantas their choosing to do so. After all,before deciding which income option tochoose, participants must make a choicewhether to begin receiving income.

The Retirement Decision

It is important to recognize that at anygiven time there are two possibly verydifferent groups of individuals withunannuitized assets at TIAA-CREF:those who are making current contribu-tions under their retirement contracts,and those who are not. Knowing an indi-vidual’s “contribution status” is usefulbecause it tells us something about his or

There have been large changes over time in the rate at which individuals are retiring and starting to use

TIAA-CREF’s annuity and other income options.

18429AA 3/9/00 7:01 AM Page 2

Research Dialogues Page 3

her retirement status. Almost all thosemaking contributions to their retirementcontracts are not retired. On the otherhand, we can’t draw such conclusionsabout the group not making contribu-tions: Some may be working, while somemay not. Some participants may simplybe on temporary leave, while others mayhave permanently changed jobs and nowwork at an institution where TIAA-CREFis not available. Still others may have infact retired but are using other sources ofincome to their expenses.

Using historical account data for acohort of participants, we can observewhen participants stop making contribu-tions on their TIAA-CREF contracts.Since TIAA-CREF plan contributions aretypically related to employee salary, thepoint at which contributions stop shouldbe a good proxy for determining whenindividuals are “retiring,” at least amongthe older individuals who are the mem-bers of this cohort.1

We divide the ten-year span for whichwe have data into three periods (1987–

1990, 1991–1993, and 1994–1996) andcalculate the number of individuals whostop contributing in each period at eachage from 60 through 72 as a percentage ofthose of each age who were contributorsin each period.2 We call this percentagethe “retirement rate” of individuals of agiven age in each period. Figure 1 presentsthese retirement-rate data for men andwomen.

For both men and women, the datashow peaks at typical retirement ages of62–63, 65–66, and 70–71, suggestingthat sizable fractions of cohort membersworked and made contributions up untilthese ages and then retired. There is someindication of difference in retirementrates for men and women. At each agebelow 69, in each of the three periods,women in the cohort were slightly morelikely to retire than were men. For exam-ple, roughly 30 percent of women work-ing in the year that they turned 65 retiredin that year. (As shown in Figure 1, theretirement rates were 30.5 percent in1987–1990, 29.6 percent in 1991–1993,

and 30.5 percent in 1994–1996.) At thesame time, roughly 24 percent of menworking at this age retired in the yearthey turned 65; the retirement rates for65-year-old men were 25.4 percent, 24.0percent, and 23.4 percent in each of thethree periods respectively. The differencesbetween men and women are not large(on the order of 5 percent to 6 percent)and do not appear to have been growingover time.

There are several possible explanationsfor gender differences in retirement rates,one of which is occupational differences.In the higher education population of1986 (from which the cohort is drawn),men were more likely to be tenured pro-fessors than were women. It seems reason-able to assume that professors might be(at all ages) less likely to retire from workthan individuals employed in other posi-tions. In addition, married couples mayretire together. To the extent that marriedwomen are younger than their husbands,women would therefore appear to beretiring earlier than men.

72717069686766656463626160

60%

50%

40%

30%

20%

10%

0%

Age

Perc

ent o

f age

gro

up

Source: TIAA-CREF Institute Research.[Note: The retirement rate is the percent of contributors whose contributions to TIAA-CREF plans stop at each age in selected time periods.]

Figure 1Retirement Rates, by Age

1987–1990 1991–1993 1994–1996

72717069686766656463626160

60%

50%

40%

30%

20%

10%

0%

Perc

ent o

f age

gro

up

Age

Women Men

18429AA 3/9/00 7:01 AM Page 3

Page 4 Research Dialogues

The most striking feature of the datais the over 20 percent decline in the maleretirement rate at age 71 from the period1987–1990 to the period 1994–1996.Figure 1 shows that in 1987–1990, 51.2percent of men who were making contri-butions in the year they turned 71 retiredat some point during that year. In 1994–1996, only 28.6 percent of 70-year-oldsmaking contributions retired. Thechanges from 1987–1990 to 1994–1996for men age 70 (-9.3 percent) and 72 (-19.7 percent) are also significant.While changes among individuals overage 68 are quite large, the changes inretirement rates at ages younger than 69are very small for both men and women—in almost all cases, retirement rates havechanged by less than 2 percent, althoughfor men, the retirement rates at ages 65 and66 have fallen by roughly 2 percent to 3percent. These data strongly suggest thatamong this cohort of individuals there hasbeen little or no change in the rate at whichindividuals under age 69 are retiring.

The dramatic decline in the retire-ment rate for men at ages 69 to 71 islikely to be the result of the end of age-based mandatory retirement rules fortenured faculty in higher education,which occurred in January 1994. Furthersupport for this speculation is the factthat the decline in the retirement rates atages 69 and above among women in thecohort (relatively fewer of whom aretenured professors) has been muchsmaller than among men, although stilllarge. These data on changes in retire-ment rates following the removal ofmandatory retirement rules are consis-tent with other data regarding theimpact of the ending of mandatory retire-ment, discussed in detail in ResearchDialogues, No. 58 (Clark and Hammond,1998). The article describes preliminarydata from two other empirical studiesshowing that retirement rates among fac-ulty members ages 70 or more haveindeed fallen since 1994.

Income Options and the Choice to Begin Receiving Income

Table 1 provides a basic description of theincome options generally available

to TIAA-CREF participants.3 The list ofcurrent retirement income optionsincludes several nonannuity optionsadded since 1989, each of which becameavailable to cohort members (and allother participants) as time passed.

The list of many choices, however,masks the very stark, basic choice that par-ticipants must make when they decide totake distributions from their retirementaccumulations: whether or not to purchasethe insurance provided by the life annuity.None of the nonannuity options (mini-mum distributions, lump-sum with-drawals, systematic withdrawals, andperiod-certain annuities) provides aninsurance component guaranteeing thatassets distributed via these mechanismswill not be exhausted before death.

The decision whether or not to annu-itize an accumulation is one of the mostsignificant financial decisions facingTIAA-CREF participants, and indeed alldefined contribution plan participants.There are substantial risks that must beconsidered and counterbalanced: Thechoice to annuitize an accumulation canmitigate the significant risk that an indi-vidual will otherwise outlive assets neces-sary to provide for living expenses inretirement. However, annuitization mayreduce the amount of wealth immedi-ately available for unanticipated healthcosts or other emergencies. Annuitizationalso lowers the amount of wealth thatwould otherwise be left to beneficiariesor heirs.

It is important to recognize that forthe cohort members we study (as well ascurrent participants) the irreversibility of the decision to start an annuity gener-ates an incentive, or an “option value,” topostponing annuitization until moreinformation about possible future out-comes can be obtained.4 In other words,because uncertainty about the future maybe greater at time t than it is at time t + 1,there is a value to postponing annuitiza-tion until t + 1. For individuals consideringthe purchase of a life annuity, the valuablenew information generally relates tochanges in health status and theprospects of living longer than average.

Of course, there are also costs inherent inpostponing the annuitization decision.Whether the benefits of waiting exceedthe costs is a very important question—and one that is sometimes quite difficultto answer. (An appendix to this issue ofResearch Dialogues discusses some of theconsiderations involved in determiningthe option value of waiting to purchasean annuity.)

While the option value is an impor-tant factor for all individuals to considerwhen deciding whether to begin annu-ity income, the existence of the optionvalue cannot by itself explain why annu-itization rates have fallen. To cause thischange, something must have increasedthe value of waiting to purchase anannuity over the period we study. Wehypothesize that the substantial increasesince 1987 in the value of equities mayhave played just such a role in enhanc-ing the option value of waiting to annu-itize. Several separate factors may beparticularly important.

First, there may be a simple “wealtheffect”: Because the financial marketshave performed so well, individuals arenow generally wealthier than they mayhave expected when they began savingfor retirement. This extra wealth maymake them more confident that they willhave enough resources to prosperthroughout even the very longest retire-ment. As a result, they may feel less needto insure themselves against longevity risk.

Second, since participants’ accumula-tions are larger, the cost of guessingwrong about longevity after having pur-chased an annuity may be perceived ashigher. Although participants do notneed to annuitize their entire accumula-tions at once, and can purchase annuityoptions that will provide some benefitseven in the case of an early death, someparticipants may not realize this. Theymay simply avoid starting an annuityaltogether, since they don’t wish to “giveup” the option to use their larger accu-mulations in other ways.

Third, the tax benefits of postponingthe withdrawal of retirement plan assetsthrough an annuity, or indeed throughany mechanism, may have increased.

18429AA 3/9/00 7:01 AM Page 4

Research Dialogues Page 5

Individuals retiring in recent years mayhave other after-tax assets that, like theirretirement accumulations, have greatlyappreciated in value as the stock markethas risen. Bodie and Crane (1997) presentdata on a subset of TIAA-CREF partici-pants that indicates that while TIAA-CREF assets make up the largest part ofthe financial wealth of most participants,many individuals have substantialamounts of other assets. For individualswho intend to spend the majority of boththeir after-tax savings and their retire-ment wealth in their retirement, it maymake sense from a tax standpoint to use

after-tax savings first, deferring incometaxes on retirement accumulations aslong as possible. The tax issues involvedare in general quite complicated, and indi-vidual situations may differ dramatically.

Finally, interest rates generallydeclined in the 1980s. As a result, forindividuals starting TIAA traditionalannuities, starting annuity income pay-ments (per dollar annuitized) have beengenerally lower than they were in theearly eighties. (All CREF life annuitiesused a constant assumed interest rate of 4percent throughout this period, so initialincome payments were not affected by

interest rate changes.) Individuals maynot understand or be aware that allTIAA-CREF annuities have at least somevariable component (TIAA traditionalannuities guarantee an interest rate of 2.5percent, with dividends as declared eachyear; there are no investment return guar-antees for CREF annuities). The lowerstarting income levels for TIAA annuitiesmay have made the TIAA life annuityless attractive than it was when prevail-ing interest rates were higher.

All these factors may have played animportant role in the decline in the fre-quency of annuitizations over the period1987–1996. In the next two sections, weexamine the changes in annuitizationrates over time and then attempt to assesswhether these changes are at all related toincreased retirement wealth.

Changes in Annuitization Rates

Figure 2 illustrates, for three different periods (1987–1990, 1991–1993, and1994–1996), the percentage of retirees ofeach age who started their first life annu-ity at the time they retired.5

The age patterns in the data are fairlystable over time, with peaks at ages 62,65, and 70–71. The data indicate thatthe annuitization probability is highestfor those who retire at age 70–71 in eachof the periods analyzed. For both menand women, the data show a decline overtime in annuitization rates at all ages.The decline appears to have been slightlylarger for older individuals. For men age70, the annuitization rate fell from 83.4percent in the period 1987–1990 to 57percent in 1994–1996, while at age 62,the decline was from 66.3 percent to 52.1percent. The drop was greatest for bothmen and women from the 1991–1993period to the 1994–1996 period. Overall,the declines were greater in magnitudefor men (there was an average decline of17 percent for women and 20 percent formen over all ages shown from the1987–1990 period to the 1994–1996period). While some drop in annuitiza-tion rates at the older ages (particularlyabove age 70) is perhaps a predictableresult of the introduction of the

Income Options

Life Annuity (since 1918)Provides income for the life of the annuitant (or annuitants, if a two-life annuity is pur-chased). An irrevocable contract between TIAA-CREF and the annuitant(s), this option isthe only one that provides insurance against the risk that the annuitant(s) may livelonger than their assets would otherwise support. Additional options can be elected (at a specified cost) that will ensure that payments will continue for at least a set minimumnumber of years (i.e., assuring that payments would be made to a designated beneficiaryeven in the case of an early death of the annuitant(s)).

Nonannuity Options

Minimum Distribution Option (MDO) (since 1991)Provides an amount of income just sufficient to avoid penalties that the federal govern-ment assesses on individuals who do not use the assets accumulated in tax-deferredretirement accounts to provide income in retirement. While lump-sum withdrawals fromaccumulated assets are allowed, and conversion to an annuity is generally possible,selection of particular income calculation options can preclude the later selection of a lifeannuity.

Systematic withdrawals and transfers (SWAT) (since 1994)The participant specifies a desired schedule of payments, and regular withdrawals ortransfers are made from their accumulated assets according to the schedule. Paymentscan be stopped or changed at any time, but will otherwise be made as long as there areassets left to fund them.

Interest Payment Retirement Option (IPRO) (since 1989)For individuals who don’t yet want to purchase an annuity, but wish to begin receiving sys-tematic income payments from amounts accumulated in the TIAA traditional annuity. Theinterest credited to TIAA traditional annuity accumulations is distributed to the participantas an income payment, while the principal balance of the accumulation remains undistrib-uted and must later be annuitized or converted to a Minimum Distribution contract.

Retirement Transition Benefit (lump-sum withdrawal) (since 1990)For individuals who may need larger amounts of cash immediately following retirement,most TIAA-CREF retirement plans allow a lump-sum withdrawal of at least a portion ofaccumulations.

Transfer Payout Annuity (TPA) (since 1991)This option provides for regular payments to be made from TIAA traditional annuityaccumulations to an annuitant or beneficiaries over a ten-year period. The payments aremade regularly over the set period, regardless of how long the annuitant lives. (Thisoption is available only for TIAA traditional annuity accumulations.)

Table 1Income Options Currently Offered by TIAA-CREF

18429AA 3/9/00 7:01 AM Page 5

Page 6 Research Dialogues

Minimum Distribution Option contractin 1991, the reason for large declines atearlier ages is more difficult to explain.

Certainly some of the drop in annuiti-zation rates is simply a result of the intro-duction of alternatives to the life annuity.Those participants who would have likedto begin receiving retirement income in aform other than the life annuity were ableto do so through an increasing variety ofoptions in the late 1980s and early 1990s,and certainly some of those individualshave chosen alternative income options.

However, if participants are simplysubstituting other forms of income forthe life annuity, then we should see eitheran increase or no change over time in therate at which they are electing to receiveincome of any form from their retirementannuities when they retire. Using thecohort data, we can check to see whetherthis is in fact the case. Figure 3 shows therates at which retiring individuals electedany form of income from their TIAA-CREF annuity contracts. (Individuals areconsidered to take income if they start a

life annuity, take a cash withdrawal, startincome under an IPRO or MDO, start aperiod-certain annuity, systematic with-drawals, or a Transfer Payout Annuity thatpays a cash benefit.)

The data show that among both menand women retiring at age 60 or 61, therehas in fact been little or no change overallin the rate at which individuals elect toreceive income. Coupled with the dataregarding the decline in annuitizationsamong these individuals, the data suggestthat younger retirees have shifted to otherforms of distribution to receive incomeimmediately upon retirement. An inter-esting question that we may be able toanswer in the future is what fraction ofthese individuals ultimately choose to takean annuity or other form of distributionlater on in retirement (say after ten years inretirement). This question can’t beanswered now, since those retiring at age60 in 1994–1996 have only been retired afew years.

The data also show that amongwomen retiring at ages 62 to 68 and men

retiring at ages 64 to 68, there has been adecline in the rate at which individualsare choosing to receive income of anykind from their accumulations. Forwomen age 65, the fraction of retireesstarting to receive income within a yearof retirement fell from 88.6 percent in1987–1990 to 80.6 percent in 1994–1996. For men, the decline was from89.5 percent in 1987–1990 to 82.1 per-cent in 1994–1996. Declines at agesaround 65 are similar in size, averagingroughly 5 percent to 7 percent. Thisbehavior suggests that a portion of theobserved decline in annuitization rates atretirement is a result of participants’electing not to take any distributions atall from their retirement assets, asopposed to a shift away from the lifeannuity as a form of distribution.

Finally, among those retiring at age69 and above, there has been almost nochange over time in the rates at whichindividuals begin to receive income fromtheir retirement assets. As shown inFigure 3, nearly all of those retiring after

72717069686766656463626160

90%

80%

70%

60%

50%

40%

30%

Age

Perc

ent o

f age

gro

up

Figure 2Rates of Annuitization among Retirees, by Age

1987–1990 1991–1993 1994–1996

72717069686766656463626160

90%

80%

70%

60%

50%

40%

30%

Perc

ent o

f age

gro

up

Age

Women Men

Source: TIAA-CREF Institute Research.[Note: The annuitization rate is the percent of retirees starting first life-annuity income within one year of retiring.]

18429AA 3/9/00 7:01 AM Page 6

Research Dialogues Page 7

72717069686766656463626160

100%

90%

80%

70%

60%

50%

Age

Perc

ent o

f age

gro

up

Figure 3Rates of Income Receipt among Retirees, by Age

1987–1990 1991–1993 1994–1996

72717069686766656463626160

100%

90%

80%

70%

60%

50%

Perc

ent o

f age

gro

up

Age

Women Men

Source: TIAA-CREF Institute Research.

age 69 receive some form of distributionwithin the next calendar year. Forwomen, the rates of income receipt for those retiring in the year they turn 71were 98.6 percent in 1987–1990, 99.1percent in 1991–1993, and 97.4 percentin 1994–1996. For men, the rates were97.8 percent, 99.2 percent, and 97.7 per-cent, respectively. Certainly the mostimportant factor in this consistency isthat federal minimum distributionrequirements require that at least a min-imum amount of income be distributedbeginning at age 701/2 for virtually allretirees. For those who do not have a lifeannuity, TIAA-CREF’s MDO contractallows them to meet these requirementswith the smallest amount of withdrawalspossible. (For a detailed description andanalysis of the minimum distributionrequirements, see Warshawsky, 1998.)

However, the fact that annuitizationrates among retirees ages 69 and abovehave dropped does not necessarily indi-cate that individuals over this age whowant to receive income from their retire-

ment assets have chosen the MDO infavor of the life annuity. Since everyretiree over 701/2 must begin some kindof distribution, it is possible that many ofthose using the MDO contract are thosewho would simply prefer to delay with-drawals even longer but have no choice.The fact that there has been an increase inthe fraction of individuals at ages 65 to68 who are delaying receipt of incomesuggests that many of those over 69might also like to delay income, but wouldface severe tax penalties for doing so.

One way to test the hypothesis thatindividuals are postponing annuitizationdecisions as a result of unanticipatedincreases in wealth is to examine thebehavior of those individuals for whomthe increases in wealth have been rela-tively larger. We can identify the individ-uals who have had systematically largerincreases in retirement wealth in recentyears by examining data on asset alloca-tion in unannuitized contracts. Thoseindividuals who have higher allocations inequity-based accounts will, by virtue of

the extremely favorable recent stock mar-ket results, have experienced greaterincreases in retirement wealth than others(assuming all else the same). If increasedwealth plays a role in the decision toannuitize, we should therefore see that theannuitization rates for those individualswith higher equity allocations have fallenfurther since 1987–1990 than annuitiza-tion rates among individuals with lowerequity allocations.

Figures 4a and 4b present data onannuitization probabilities at retirementfor the cohort, broken out by the percentof the total unannuitized accumulationheld in equities in the year prior to retire-ment. The data show that the decline inthe fraction of individuals purchasing anannuity from 1987–1990 to 1994–1996was generally slightly higher for bothmen and women with more than 79 per-cent of their unannuitized accumulationsin equities than it was for individualswith lower equity allocations. For exam-ple, for men age 65 with between 20 per-cent and 39 percent equities, the fraction

18429AA 3/9/00 7:01 AM Page 7

Page 8 Research Dialogues

72717069686766656463626160

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

1987–19901991–19931994–1996

72717069686766656463626160

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%72717069686766656463626160

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

72717069686766656463626160

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

72717069686766656463626160

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

Figure 4aAnnuitization Rates for Female Retirees by Age, Time Period, and Equity Allocation at Retirement

0%–19% Equity 20%–39% Equity

40%–59% Equity

80%–100% Equity

60%–79% Equity

Perc

ent o

f ret

irees

sta

rting

firs

t life

ann

uity

Perc

ent o

f ret

irees

sta

rting

firs

t life

ann

uity

Perc

ent o

f ret

irees

sta

rting

firs

t life

ann

uity

Perc

ent o

f ret

irees

sta

rting

firs

t life

ann

uity

Perc

ent o

f ret

irees

sta

rting

firs

t life

ann

uity

Age

Age

Age

Age

AgeSource: TIAA-CREF Institute Research.

18429AA 3/9/00 7:01 AM Page 8

Research Dialogues Page 9

72717069686766656463626160

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

72717069686766656463626160

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%72717069686766656463626160

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

72717069686766656463626160

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

72717069686766656463626160

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

Figure 4bAnnuitization Rates for Male Retirees, by Age, Time Period, and Equity Allocation at Retirement

0%–19% Equity 20%–39% Equity

40%–59% Equity

80%–100% Equity

60%–79% Equity

Perc

ent o

f ret

irees

sta

rting

firs

t life

ann

uity

Perc

ent o

f ret

irees

sta

rting

firs

t life

ann

uity

Perc

ent o

f ret

irees

sta

rting

firs

t life

ann

uity

Perc

ent o

f ret

irees

sta

rting

firs

t life

ann

uity

Perc

ent o

f ret

irees

sta

rting

firs

t life

ann

uity

Age

Age

Age

Age

Age

1987–19901991–19931994–1996

Source: TIAA-CREF Institute Research.

18429AA 3/9/00 7:01 AM Page 9

Page 10 Research Dialogues

starting an annuity within a year ofretiring was 81.1 percent in 1987–1990and 62.3 percent in 1994–1996, a dif-ference of 19 percent. At the same time,for men age 65 with between 80 percentand 100 percent equities, the change inthe annuitization rate was from 76.2percent in 1987–1990 to 40 percent in1994–1996, a difference of over 36 per-cent. The pattern is more uniformamong women: Figure 4a clearly showsthat the gap between the line for1987–1990 and the line for 1994–1996is much larger for women in the 80 per-cent to 100 percent equities category thanit is for the other categories. The factthat the result shows up much moreclearly for women may be related to thefact that women are more likely to pur-chase single-life annuities than men (seeResearch Dialogues, No. 48). If women arein fact more frequently purchasingannuities for themselves only, then theymay perceive a given increase in retire-ment wealth as greater than men would,since men are more likely to purchasetwo-life annuities, in which the benefitsof any increased wealth are shared byboth annuitants.

While the data do seem to supportthe claim that the decline in annuitiza-tion rates has been higher among indi-viduals with greater equity allocations,the results do vary a great deal depend-ing on the age at which the comparisonis made. (Figure 4b shows, for example,that among men age 68, with between60 percent and 79 percent in equities,the change in the annuitization rate wasonly –6.1 percent from 1987–1990 to1994–1996. This is less than thedeclines at most ages even in the lowestequity-allocation category.) In addition,it is possible that equity allocation maybe correlated with other important vari-ables that may affect the decision topurchase a life annuity. In particular, itis likely that individuals with higherequity allocations are more likely totake financial risks. These risk-takingindividuals may be less worried aboutthe longevity risks they face at retire-ment, and so are predisposed to avoid

purchasing an annuity if an alternateoption is available.

Conclusions

As TIAA-CREF’s menu of incomeoptions has grown, participants havegreater freedom to postpone decisionsabout when to start an income stream.The data we have reviewed have detailedseveral interesting changes—and inter-esting consistencies—in the behavior ofTIAA-CREF participants over time:

● The rate at which members of the TIAA-CREF population over theage of 69—particularly men—areretiring declined dramatically from1987–1990 to 1994–1996, withmuch of the decline occurring fol-lowing the end of mandatory retire-ment policies for tenured faculty inhigher education.

● The rate at which members of theTIAA-CREF population, both menand women, retire at ages 68 andunder has remained relatively con-stant since 1987. There have beenslight declines in retirement ratesamong men at age 65, but almostno change at all at other ages.

● The frequency at which participantsare choosing to begin a life annuityimmediately following retirementhas declined steadily for both menand women of all ages as the num-ber of alternative options availableto TIAA-CREF participants hasexpanded.

● Many participants retiring at ages62 to 68 are postponing distribu-tion of their retirement assets untillater ages. A majority continue touse the life annuity as the mecha-nism for receiving income inretirement.

● Unexpected increases in retirementwealth and other assets may haveplayed an important role in thedrop in annuitization rates.Individuals may see themselves aswealthier than they expected, andthey may feel that annuitization is

too costly in terms of preventingthe use of wealth for emergenciesand limiting its availability forbequests.

● The data indicate that, in general,the decline in annuitization rateshas been slightly larger amongindividuals with greater equityallocations. This is consistent withthe hypothesis that increases inretirement wealth are at least par-tially responsible for the observeddecline in annuitizations.

Appendix

Calculating the Option Value of Postponing Annuitization

Determining the option value of delayingannuitization involves considering boththe benefits of waiting for more informa-tion and the costs of delaying.

While the benefit of acquiring infor-mation about potential longevity can belarge, there are also substantial costs topostponement. Most significantly, post-poning annuitization means forgoing atleast some current and/or future income.It also means that time must continuallybe spent paying attention to the situa-tion—so as to be sure to make the deci-sion to annuitize at the right time.

The option value of postponing annu-itization therefore tends to decline overtime, as uncertainty about the futuredecreases, the costs of postponing thedecision rise, and the value of additionalinformation declines. At some point, thebenefits of waiting an additional year areless than the cost of putting off the deci-sion, and the best thing to do is to goahead with annuitizing.

It is important to note that for indi-viduals considering the purchase of avariable annuity that allows the annui-tant to select and change the type ofinvestments funding the annuity benefit,financial uncertainty (uncertainty aboutfuture investment returns) should notaffect the timing of annuitization.Participants in this type of variable annu-ity control their asset allocation, andtherefore their exposure to financial risks,

18429AA 3/9/00 7:01 AM Page 10

Research Dialogues Page 11

Tab

le A

1P

roje

cted

In

com

e an

d A

ccu

mu

lati

on

Pat

tern

s B

ased

on

Dif

fere

nt

An

nu

ity

Sta

rt A

ges

, fo

r H

ypoth

etic

al S

ingle

-Lif

e A

nn

uit

y w

ith

No G

uar

ante

e P

erio

d

Scen

ario

A

Scen

ario

BSc

enar

io C

Scen

ario

DSc

enar

io E

Ann

uitiz

e at

65

Ann

uitiz

e at

70

Ann

uitiz

e at

75

Ann

uitiz

e at

80

Ann

uitiz

e at

85

End-

of-Y

ear

End-

of-Y

ear

End-

of-Y

ear

End-

of-Y

ear

Annu

al

Annu

alUn

annu

itize

dAn

nual

Un

annu

itize

dAn

nual

Un

annu

itize

dAn

nual

Un

annu

itize

dA

ge

Paym

ent

Paym

ent

Accu

mul

atio

nPa

ymen

tAc

cum

ulat

ion

Paym

ent

Accu

mul

atio

nPa

ymen

tAc

cum

ulat

ion

65$2

5,02

6 $2

5,02

6 $2

92,8

48

$25,

026

$292

,848

$2

5,02

6 $2

92,8

48

$25,

026

$292

,848

66

$25,

026

$25,

026

$285

,230

$2

5,02

6 $2

85,2

30

$25,

026

$285

,230

$2

5,02

6 $2

85,2

30

67$2

5,02

6$2

5,02

6 $2

77,1

18

$25,

026

$277

,118

$2

5,02

6 $2

77,1

18

$25,

026

$277

,118

68

$25,

026

$25,

026

$268

,478

$2

5,02

6 $2

68,4

78

$25,

026

$268

,478

$2

5,02

6 $2

68,4

78

69$2

5,02

6$2

5,02

6 $2

59,2

77

$25,

026

$259

,277

$2

5,02

6 $2

59,2

77

$25,

026

$259

,277

70

$25,

026

$23,

826

$0

$25,

026

$249

,477

$2

5,02

6 $2

49,4

77

$25,

026

$249

,477

71

$25,

026

$23,

826

$0

$25,

026

$239

,041

$2

5,02

6 $2

39,0

41

$25,

026

$239

,041

72

$25,

026

$23,

826

$0

$25,

026

$227

,926

$2

5,02

6 $2

27,9

26

$25,

026

$227

,926

73

$25,

026

$23,

826

$0

$25,

026

$216

,089

$2

5,02

6 $2

16,0

89

$25,

026

$216

,089

74

$25,

026

$23,

826

$0

$25,

026

$203

,483

$2

5,02

6 $2

03,4

83

$25,

026

$203

,483

75

$25,

026

$23,

826

$0

$21,

153

$0

$25,

026

$190

,057

$2

5,02

6 $1

90,0

57

76$2

5,02

6 $2

3,82

6$0

$2

1,15

3 $0

$2

5,02

6 $1

75,7

58

$25,

026

$175

,758

77

$25,

026

$23,

826

$0

$21,

153

$0

$25,

026

$160

,530

$2

5,02

6 $1

60,5

30

78$2

5,02

6 $2

3,82

6$0

$2

1,15

3 $0

$2

5,02

6 $1

44,3

12

$25,

026

$144

,312

79

$25,

026

$23,

826

$0

$21,

153

$0

$25,

026

$127

,040

$2

5,02

6 $1

27,0

40

80$2

5,02

6 $2

3,82

6$0

$2

1,15

3 $0

$1

5,39

7 $0

$2

5,02

6 $1

08,6

45

81$2

5,02

6 $2

3,82

6$0

$2

1,15

3$0

$1

5,39

7 $0

$2

5,02

6 $8

9,05

5 82

$25,

026

$23,

826

$0

$21,

153

$0

$15,

397

$0

$25,

026

$68,

191

83$2

5,02

6 $2

3,82

6$0

$2

1,15

3 $0

$1

5,39

7 $0

$2

5,02

6 $4

5,97

1 84

$25,

026

$23,

826

$0

$21,

153

$0

$15,

397

$0

$25,

026

$22,

306

85$2

5,02

6 $2

3,82

6 $0

$2

1,15

3 $0

$1

5,39

7 $0

$3

,253

$0

86

$25,

026

$23,

826

$0

$21,

153

$0

$15,

397

$0

$3,2

53

$0

87$2

5,02

6 $2

3,82

6 $0

$2

1,15

3 $0

$1

5,39

7 $0

$3

,253

$0

88

$25,

026

$23,

826

$0

$21,

153

$0

$15,

397

$0

$3,2

53

$0

89$2

5,02

6 $2

3,82

6 $0

$2

1,15

3 $0

$1

5,39

7 $0

$3

,253

$0

90

$25,

026

$23,

826

$0

$21,

153

$0

$15,

397

$0

$3,2

53

$0

91$2

5,02

6 $2

3,82

6 $0

$2

1,15

3 $0

$1

5,39

7 $0

$3

,253

$0

92

$25,

026

$23,

826

$0

$21,

153

$0

$15,

397

$0

$3,2

53

$0

93$2

5,02

6 $2

3,82

6$0

$2

1,15

3 $0

$1

5,39

7 $0

$3

,253

$0

94

$25,

026

$23,

826

$0

$21,

153

$0

$15,

397

$0

$3,2

53

$0

95 o

n$2

5,02

6 $2

3,82

6 $0

$2

1,15

3 $0

$1

5,39

7 $0

$3

,253

$0

Syst

emat

ic w

ithd

raw

als

equ

al t

o th

e ag

e-65

an

nu

ity

paym

ent

are

mad

e in

eac

h ye

ar b

efor

e th

e an

nu

ity

is s

tart

ed.

On

ce s

tart

ed, a

ll a

nn

uit

y pa

ymen

ts (

in b

old)

con

tin

ue

for

the

rest

of

the

ann

uit

ant’s

lif

e.

Ann

uity

ben

efit

leve

ls re

lativ

e to

age

-65

star

t:10

0%95

%85

%62

%13

%As

sum

ptio

ns: A

ccum

ulat

ion

at re

tirem

ent (

age

65) i

s $3

00,0

00. A

nnua

l ann

uity

pay

men

ts o

r sys

tem

atic

with

draw

als

are

mad

e on

ce a

yea

r, at

the

begi

nnin

g of

eac

h ye

ar. A

nnui

ty is

a s

tand

ard

sing

le-li

fe

annu

ity, a

ssum

ing

a hy

poth

etic

al 6

.5%

tota

l int

eres

t rat

e. M

orta

lity

base

d on

Ann

uity

200

0 (m

erge

d ge

nder

) tab

le, w

ith a

ges

set b

ack

2 ye

ars.

18429AA 3/9/00 7:01 AM Page 11

Page 12 Research Dialogues

Tab

le A

2P

roje

cted

In

com

e an

d A

ccu

mu

lati

on

Pat

tern

s B

ased

on

Dif

fere

nt

An

nu

ity

Sta

rt A

ges

, fo

r H

ypoth

etic

al S

ingle

-Lif

e A

nn

uit

y w

ith

Gu

aran

teed

Pay

men

ts t

hro

ugh

Age

84

Scen

ario

A

Scen

ario

BSc

enar

io C

Scen

ario

DSc

enar

io E

Ann

uitiz

e at

65

Ann

uitiz

e at

70

Ann

uitiz

e at

75

Ann

uitiz

e at

80

Ann

uitiz

e at

85

End-

of-Y

ear

End-

of-Y

ear

End-

of-Y

ear

End-

of-Y

ear

End-

of-Y

ear

End-

of-Y

ear

End-

of-Y

ear

End-

of-Y

ear

Annu

al

Valu

e of

Cer

tain

Annu

alUn

annu

itize

dVa

lue

of C

erta

inAn

nual

Un

annu

itize

dVa

lue

of C

erta

inAn

nual

Un

annu

itize

dVa

lue

of C

erta

inAn

nual

Unan

nuiti

zed

Ag

ePa

ymen

tPa

ymen

ts*

Paym

ent

Accu

mul

atio

nPa

ymen

ts*

Paym

ent

Accu

mul

atio

nPa

ymen

ts*

Paym

ent

Accu

mul

atio

nPa

ymen

tsPa

ymen

tAc

cum

ulat

ion

65$2

3,06

4$2

63,6

76

$23,

064

$294

,937

$0

$2

3,06

4 $2

94,9

37

$0

$23,

064

$294

,937

$0

$2

3,06

4 $2

94,9

37

66$2

3,06

4$2

56,2

52

$23,

064

$289

,545

$0

$2

3,06

4 $2

89,5

45

$0

$23,

064

$289

,545

$0

$2

3,06

4 $2

89,5

45

67$2

3,06

4$2

48,3

45

$23,

064

$283

,803

$0

$2

3,06

4 $2

83,8

03

$0

$23,

064

$283

,803

$0

$2

3,06

4 $2

83,8

03

68$2

3,06

4$2

39,9

25

$23,

064

$277

,687

$0

$2

3,06

4 $2

77,6

87

$0

$23,

064

$277

,687

$0

$2

3,06

4 $2

77,6

87

69$2

3,06

4$2

30,9

57

$23,

064

$271

,173

$0

$2

3,06

4 $2

71,1

73

$0

$23,

064

$271

,173

$0

$2

3,06

4 $2

71,1

73

70$2

3,06

4$2

21,4

06

$22,

921

$0

$220

,037

$2

3,06

4 $2

64,2

37

$0

$23,

064

$264

,237

$0

$2

3,06

4 $2

64,2

37

71$2

3,06

4$2

11,2

35

$22,

921

$0

$209

,928

$2

3,06

4 $2

56,8

49

$0

$23,

064

$256

,849

$0

$2

3,06

4 $2

56,8

49

72$2

3,06

4$2

00,4

02

$22,

921

$0

$199

,162

$2

3,06

4 $2

48,9

81

$0

$23,

064

$248

,981

$0

$2

3,06

4 $2

48,9

81

73$2

3,06

4$1

88,8

65

$22,

921

$0

$187

,697

$2

3,06

4 $2

40,6

02

$0

$23,

064

$240

,602

$0

$2

3,06

4 $2

40,6

02

74$2

3,06

4$1

76,5

79

$22,

921

$0

$175

,486

$2

3,06

4 $2

31,6

79

$0

$23,

064

$231

,679

$0

$2

3,06

4 $2

31,6

79

75$2

3,06

4$1

63,4

93

$22,

921

$0

$162

,482

$2

2,44

5$0

$1

59,1

06

$23,

064

$222

,175

$0

$2

3,06

4 $2

22,1

75

76$2

3,06

4$1

49,5

58

$22,

921

$0

$148

,632

$2

2,44

5$0

$1

45,5

44

$23,

064

$212

,053

$0

$2

3,06

4 $2

12,0

53

77$2

3,06

4$1

34,7

16

$22,

921

$0

$133

,882

$2

2,44

5$0

$1

31,1

01

$23,

064

$201

,274

$0

$2

3,06

4 $2

01,2

74

78$2

3,06

4$1

18,9

10

$22,

921

$0

$118

,174

$2

2,44

5$0

$1

15,7

19

$23,

064

$189

,793

$0

$2

3,06

4 $1

89,7

93

79$2

3,06

4$1

02,0

76

$22,

921

$0

$101

,444

$2

2,44

5$0

$9

9,33

7 $2

3,06

4 $1

77,5

67

$0

$23,

064

$177

,567

80

$23,

064

$84,

148

$22,

921

$0

$83,

627

$22,

445

$0

$81,

890

$20,

818

$0

$75,

955

$23,

064

$164

,546

81

$23,

064

$65,

054

$22,

921

$0

$64,

652

$22,

445

$0

$63,

309

$20,

818

$0

$58,

720

$23,

064

$150

,679

82

$23,

064

$44,

720

$22,

921

$0

$44,

443

$22,

445

$0

$43,

520

$20,

818

$0

$40,

366

$23,

064

$135

,910

83

$23,

064

$23,

064

$22,

921

$0

$22,

921

$22,

445

$0

$22,

445

$20,

818

$0

$20,

818

$23,

064

$120

,181

84

$23,

064

$0

$22,

921

$0

$0

$22,

445

$0

$0

$20,

818

$0

$0

$23,

064

$103

,430

85

$23,

064

$0

$22,

921

$0

$0

$22,

445

$0

$0

$20,

818

$0

$0

$15,

081

$0

86$2

3,06

4$0

$2

2,92

1$0

$0

$2

2,44

5$0

$0

$2

0,81

8 $0

$0

$1

5,08

1 $0

87

$23,

064

$0

$22,

921

$0

$0

$22,

445

$0

$0

$20,

818

$0

$0

$15,

081

$0

88$2

3,06

4$0

$2

2,92

1$0

$0

$2

2,44

5$0

$0

$2

0,81

8 $0

$0

$1

5,08

1 $0

89

$23,

064

$0

$22,

921

$0

$0

$22,

445

$0

$0

$20,

818

$0

$0

$15,

081

$0

90$2

3,06

4$0

$2

2,92

1$0

$0

$2

2,44

5$0

$0

$2

0,81

8 $0

$0

$1

5,08

1 $0

91

$23,

064

$0

$22,

921

$0

$0

$22,

445

$0

$0

$20,

818

$0

$0

$15,

081

$0

92$2

3,06

4$0

$2

2,92

1$0

$0

$2

2,44

5$0

$0

$2

0,81

8 $0

$0

$1

5,08

1 $0

93

$23,

064

$0

$22,

921

$0

$0

$22,

445

$0

$0

$20,

818

$0

$0

$15,

081

$0

94$2

3,06

4$0

$2

2,92

1$0

$0

$2

2,44

5$0

$0

$2

0,81

8 $0

$0

$1

5,08

1 $0

95

on

$23,

064

$0

$22,

921

$0

$0

$22,

445

$0

$0

$20,

818

$0

$0

$15,

081

$0

*The

se p

rese

nt v

alue

cal

cula

tions

ass

ume

a 6.

5% ra

te o

f int

eres

t.Sy

stem

atic

wit

hdra

wal

s eq

ual

to

the

age-

65 a

nn

uit

y pa

ymen

t ar

e m

ade

in e

ach

year

bef

ore

the

ann

uit

y is

sta

rted

. O

nce

sta

rted

, all

an

nu

ity

paym

ents

(in

bol

d) c

onti

nu

e fo

r th

e re

st o

f th

e an

nu

itan

t’s l

ife.

A

nnui

ty b

enef

it le

vels

rela

tive

to a

ge-6

5 st

art:

100%

99%

97%

90%

65%

Assu

mpt

ions

: Acc

umul

atio

n at

retir

emen

t (ag

e 65

) is

$300

,000

. Ann

ual a

nnui

ty p

aym

ents

or s

yste

mat

ic w

ithdr

awal

s ar

e m

ade

once

a y

ear,

at th

e be

ginn

ing

of e

ach

year

. Ann

uity

is a

sta

ndar

d si

ngle

-life

an

nuity

, ass

umin

g a

hypo

thet

ical

6.5

% to

tal i

nter

est r

ate.

Mor

talit

y ba

sed

on A

nnui

ty 2

000

(mer

ged

gend

er) t

able

, with

age

s se

t bac

k 2

year

s.

18429AA 3/9/00 7:01 AM Page 12

Research Dialogues Page 13

both before and after annuitizing.Whether they begin an annuity or not,they will continue to benefit from thepositive (or suffer from the negative)investment performance that results fromtheir asset allocation choices. Therefore,while uncertainty about financial mar-kets should affect the asset allocationdecision, such uncertainty should have norole in the decision whether to beginannuity income.

A numerical example of the conse-quences of delaying the decision to beginannuity income may be useful. Table A1shows five hypothetical scenarios inwhich a participant with the same initialaccumulation amount ($300,000) at age65 compares various hypothetical life-annuity starting ages. In all the scenariosin which the annuity is started after age65, we assume the participant will makesystematic withdrawals in an amountequal to the age-65 annuity payment ineach year up until starting the annuity.Given these assumptions (and the otherfinancial assumptions listed at the bot-tom of the table), the table shows thetotal single-life annuity benefit that theparticipant would hypothetically receiveafter starting the annuity. It also showsthe amount of unannuitized accumula-tion remaining in each year before start-ing the annuity.

The boldface figures at the bottom ofthe table show the relative amounts ofsurvival-contingent income that can bepurchased in each scenario. By delayingthe start of the annuity from age 65 toage 70, the participant at age 70 wouldbe able to fund up to 95 percent of theannuity income benefit that he or shewould have had if the annuity had beenstarted at age 65. (These percentages donot depend on the size of the accumula-tion at age 64. The absolute size of thepayments or withdrawals will be larger orsmaller in proportion to the initial accu-mulation, but the relative comparisonsbeing made will be exactly the same.)

We could say, then, that delaying thestart of the annuity until age 70 is at least95 percent as good as starting annuityincome at age 65. By not starting the

annuity at age 65, the participant main-tains the option to use the accumulationin other ways before turning 70. Forexample, if the accumulation is cashable,the participant could withdraw it duringthis period if extra money were neededfor an emergency. In addition, if the par-ticipant were to die between the ages of65 and 70, the entire unannuitized accu-mulation would then become the prop-erty of a chosen beneficiary. If theparticipant had purchased the hypotheti-cal single-life annuity at age 65 with noguarantee period, there would be nodeath benefit.

The exact value of a delay will varyfrom participant to participant, depend-ing on a number of factors, such aswhether the remaining unannuitized bal-ance is cashable or not, the participant’shealth, and how much he or she caresabout leaving a bequest. In light of theirindividual circumstances, each partici-pant must decide if the benefit of annu-itizing at age 65 (a 5 percent higherincome at ages 70 and beyond) is worthgiving up the option to wait.

Table A1 also shows that as the periodof delay increases, the cost increases.Participants who delay annuitizationuntil age 75 will only be able to fund 85percent of the remaining age-65 annuitypayments. A ten-year delay, therefore,“costs” the participant 15 percent of theafter-75 income. Similarly, the tableshows the severe consequences that a fif-teen-year delay could entail: Waiting tillage 80 would cost the participant 38 per-cent of the income otherwise payable atthat age. A twenty-year delay would costalmost all (87 percent) of the after-85income payable if the annuity werestarted at age 65.

Table A2 shows another calculation ofthe “cost” of delaying annuitization,under slightly different circumstances: Inthis case, the participant is considering alife annuity with a guarantee period thatwill ensure that, if the annuitant diesbefore age 85, some annuity benefits willbe paid to beneficiaries up until the participant would have been 85. Here,waiting ten years to annuitize means hav-

ing 3 percent less annuity income (and3 percent less guaranteed income) at age75. The reason the numbers are so differ-ent is that as the participant approachesage 85, less and less of the accumulationis used to purchase the guarantee forbeneficiaries.

Both tables illustrate that, initially,the costs of waiting to purchase an annu-ity are quite modest. If the benefits ofdelaying—which are specific to eachindividual—are at all significant, it maymake sense for retired individuals atyounger ages to hold off on annuitization.However, once participants are in their70s, the costs of further postponementare significant, and increase rather quickly.

[This report was prepared forResearch Dialogues by John Ameriks, research economist, TIAA-CREF Institute(jameriks@tiaa-cref.org).]

Endnotes

1The main problem with the approach is that con-tributions can stop for reasons other than retire-ment, and may begin again after having beenstopped. Indeed, among the cohort we studied,more than 10,000 individuals had a year in whichthey made no contributions under their retirementplan contracts but then later resumed contribu-tions. The data we have collected gathers togetherall TIAA-CREF contracts owned by an individual,and contributions are not considered to havestopped unless an entire calendar year has passed inwhich no contribution is received under any con-tract. This definition of retirement should elimi-nate a large number of “false retirements”—such asindividuals on summer breaks, leaves of less than ayear, or job changes.

2This percentage is basically an empirical hazardrate, measuring the likelihood of retirement ateach age in each period, conditional on not havingretired by that age in each period. We calculate ageon the last day of the year in which contributionsstopped; age will therefore tend to slightly over-state the participant’s actual age at retirement,since age, of course, will have increased betweenthe date of retirement within the year and the endof the year.

3Some income options may not be available to someparticipants, depending on individual retirementplan features and rules. This list is not intended tosubstitute for the many publications availablefrom TIAA-CREF describing its products; a com-prehensive description of TIAA-CREF income

18429AA 3/9/00 7:01 AM Page 13

Page 14 Research Dialogues

options can be found in Choosing Income Options,number 6 in the TIAA-CREF Library Series.

4The basic idea underlying the notion of an optionvalue as an aspect of irreversible investment decisionsis discussed in detail in Dixit and Pindyck, 1994.

5These data are constructed as follows: The base ofthe percentage is all individuals of each age whosecontributions stop in year t, for any reason otherthan death, and who do not have a preexisting lifeannuity. The numerator is those in the same groupwho were issued any form of life annuity as the firstannuitant in year t, t-1, or t+1.

Bibliography

Clark, Robert L., and P. Brett Hammond. “ToRetire or Not? Examining Life after the End ofMandatory Retirement in Higher Education.”Research Dialogues, no. 58 (December 1998).

Crane, Dwight B., and Zvi Bodie. “PersonalInvesting: Advice, Theory, and Evidence.”Financial Analysts Journal 53, no. 6 (Nov./Dec.1997): 13–23.

Dixit, Avinash K., and Robert S. Pindyck.Investment Under Uncertainty. Princeton:Princeton University Press, 1994.

Gendell, Murray. “Trends in retirement age in fourcountries, 1965–1995.” U.S. Bureau of LaborStatistics Monthly Labor Review (August 1998):20–30.

Herz, Diane E. “Work after early retirement: anincreasing trend among men.” U.S. Bureau ofLabor Statistics Monthly Labor Review (April1995): 13–20.

King, Francis P. “Trends in the Selection of TIAA-CREF Life-Annuity Options, 1978–1994.”Research Dialogues, no. 48 (July 1996).

TIAA-CREF. “Participants Waiting Longer toStart Income Streams; Partial Settlements,Graded Benefit Increasingly Popular.” BenefitPlan Counselor 18, no. 5 (September 1998): 1, 5.

Warshawsky, Mark J. “Distributions fromRetirement Plans: Minimum Requirements,Current Options, and Future Directions.”Research Dialogues, no. 57 (September 1998).

Additional copies of Research Dialogues can be obtained by calling 800 842-2733, extension 3038,or can be downloaded from TIAA-CREF’s Web siteat www.tiaa-cref.org

18429AA 3/9/00 7:01 AM Page 14

18429AA 3/9/00 7:01 AM Page 15

730 Third AvenueNew York, NY 10017-3206

Ensuring the future for those who shape it.SM

18429AA 3/9/00 7:01 AM Page 16