TRENDS AND ISSUES
ANNUITIES: NOW, LATER, NEVER?
Benjamin Goodman, Director, Actuarial Consulting Services TIAA-CREF Michael Heller, Vice President, Actuarial Consulting Services TIAA-CREF October 2006
10/06 10/06
TRENDS AND ISSUES
[Link]
EXECUTIVE SUMMARY With the growing prominence of defined contribution plans as primary retirement funding vehicles, there is a concern that participants should be making appropriate decisions in drawing income from their account balances. One major question relates to the role of life annuities as an effective means of receiving retirement income. This paper examines the financial efficacy of receiving income through a life annuity versus systematic withdrawals from a participant's account. As a basis for comparison, we contrast the receipt of life annuity income to equivalent withdrawals made from an investment account, assuming that both the life annuity income and the investment account earnings reflect the same investment return (net of expenses). This form of comparison is done for both a typical fixed annuity and for a variable life annuity. Based on hypothetical future investment earnings rates, these illustrations show that the individual who utilizes systematic withdrawal is projected to run out of funds some time before reaching his or her life expectancy. In other words, there is more than a 50% chance that the individual would run out of money! This comparison is then repeated for the variable life annuity, but this time based on actual S&P 500 investment experience over two different time periods: 1965 - 1989 and 1975 - 1999. Using real investment experience, we again see that, had an individual withdrawn (systematically) the same income as would have been payable from a variable life annuity, that individual would have run out of money before reaching his or her life expectancy. In other words, whether we use hypothetical or real investment experience, a comparison of life annuities versus systematic withdrawals (of equivalent amounts) demonstrates the financial efficacy of annuitization. Put another way, one can say that a life annuity is financially engineered to maximize the amount of living income to an individual. Having demonstrated the value of utilizing a life annuity, we then go on to examine the efficacy of delaying the start of annuity income to a later age. Using examples based on both fixed and variable annuities, we show that the cost for delaying annuitization by five years (assuming a retirement age of 65) is about a 5% reduction in future income, while a delay of ten years might result in a 15% reduction in lifetime income. Its also noted that these relative reductions would be smaller in respect to individuals who retire at earlier ages. Still, while some people might feel that a 5% income reduction might be a fair trade for being given an extra five years to decide whether or not to annuitize (during which
TIAA-CREF INSTITUTE TRENDS AND ISSUES 2
10/06
TRENDS AND ISSUES
[Link]
time certain health issues may arise), there is clearly much less value in waiting ten years or more. The paper also explores the potential impact on income if individuals postpone annuitization during a period of rising interest rates. Assuming interest rates rise by 0.25% each year, a five-year delay in annuitization results in a 7% gain in income. However, the latter result reflects the assumption that interest rates are the same both in the accumulation and annuity payout stage. In reality, some companies may use higher interest rates in their payout annuities, to reflect the higher rates generally associated with long-term fixed-income investments. If payout annuity interest rates are, indeed, higher than accumulating interest rates, the value of deferring annuitization may be lessened, unless interest rates rise significantly over a relatively short time period. Finally, we point out that the issue of postponing annuitization in a rising interest rate environment has no relevance to variable life annuities, as income under the latter type of annuity is always adjusted by the investment experience of the underlying assets. TABLE OF CONTENTS
Introduction ................................................................................................................................ 3 The Life Annuity Principle Fixed Account................................................................................ 3 The Life Annuity Principle Variable Account ........................................................................... 5 Delaying The Annuity Simple Case......................................................................................... 6 Delaying The Annuity Interest Rates Rise............................................................................... 7 Conclusion ................................................................................................................................. 8 About The Authors ..................................................................................................................... 9 Exhibit 1: Life Annuity Principle Fixed Annuity ...................................................................... 11 Exhibit 2a: Life Annuity Principle Variable Annuity At 7% ..................................................... 12 Exhibit 2b: Life Annuity Principle S&P 1965-1989................................................................. 13 Exhibit 2c: Life Annuity Principle S&P 1975-1999................................................................. 14 Exhibit 3: Delaying The Annuity Fixed .................................................................................. 15 Exhibit 4: Delaying The Annuity Variable .............................................................................. 16 Exhibit 5: Delaying The Annuity With A Rise In Interest Rates (Fixed) .................................... 17 Exhibit 6: Delaying The Annuity With A Rise In Interest Rates (Fixed) .................................... 18 Exhibit 7: Rising Interest Rates ................................................................................................ 19
TIAA-CREF INSTITUTE TRENDS AND ISSUES 2
10/06
TRENDS AND ISSUES
[Link]
INTRODUCTION During the accumulation years, participants in defined contribution retirement plans need to be sure that they are making adequate contributions to their plan and that these contributions are appropriately allocated among various asset classes. Upon retirement, participants have to decide how to generate income from their account balances. One of the most important questions is whether or not to purchase a life annuity, for all or a portion of ones retirement income. As a life annuity guarantees payment of income as long as the participant lives, it can provide significant peace of mind. As we will demonstrate below, the life annuity also provides the maximum amount of living income. However, there are many retirees who are hesitant to purchase life annuities for a number of reasons, including: the loss of control of retirement assets; the loss of liquidity and flexibility; and the potential for early death (and consequently early loss of principal.) During the past few years there has been yet another reason for not purchasing a fixed life annuity: the belief that interest rates are currently low and that one can achieve better results by waiting to purchase the annuity when rates will be higher. (Whether rates will, in fact, be higher in a few years may be a subject for another paper.) The purpose of this paper is three-fold: (1) to demonstrate that a life annuity is financially engineered to maximize the amount of living income payable to retirees; (2) to show the potential impact of deferring annuitization to a later age; and (3) to quantify the impact of delaying an annuity purchase during periods of rising interest rates. THE LIFE ANNUITY PRINCIPLE FIXED ACCOUNT Exhibit 1 shows the financial impact of two retirement income strategies for a hypothetical 65-year-old: purchasing a fixed life annuity versus taking systematic withdrawals from the retirement accumulation. To make this an apples-to-apples comparison, we assume that each years withdrawal is equal to the life annuity payment, and that both accounts have the same net investment earnings rate. As you can see, the 65-year-old who takes systematic withdrawals will run out of money in the 22nd year, even though life expectancy is 23 years. This exhibit demonstrates two important aspects of the life annuity principle: (1) There is a greater than 50% chance that the retiree will run out of money if he or she takes a withdrawal equal to the life annuity payment. In other words, there is a
TIAA-CREF INSTITUTE TRENDS AND ISSUES 3
10/06
TRENDS AND ISSUES
[Link]
greater risk (mathematically) of outliving income than the risk of dying young and losing the accumulation. (2) A life annuity allows a retiree to maximize income. Since most retirees want to be sure that they do not outlive their income, they would have to plan on withdrawals lasting quite a bit longer than just their life expectancy. For example, a 65- year-old might want to plan on having income last for at least 30 years, even if his life expectancy is only 23 years. Assuming a 5% interest rate, the withdrawal amount for this retiree would only be $6,195 per year instead of the $7,390 that the life annuity offers. Of course, the person utilizing systematic withdrawal could have a better result if the investment fund earned more than the interest rate underlying the life annuity. However, it would have to be a lot more. To achieve 30 years of withdrawals of $7,390, the account would have to earn over 6.8%, or almost 2% per year higher than the life annuity. Note that this life annuity principle applies at all interest rates. If we assumed a 7% interest rate instead of 5%, the annuity payment would be higher, as would the assumed withdrawal amount and earnings rate of the non-annuity account. The balance would again be depleted before life expectancy, and the non-annuity account would again have to earn 2% per year higher than the annuity interest rate to allow the money to last for 30 years. The most often quoted downside of an annuity purchase is the potential for early death and consequent loss of principal. As is evident in Exhibit 1, if our hypothetical retiree has chosen a life annuity and died at any time during the first 22 years of retirement, there would be no death benefit payable to the estate. Under the systematic withdrawal scenario, there would be a sizable death benefit available, particularly in the earlier years. This differential in death benefits is, essentially, one of the primary trade-offs between a life annuity and non-annuity alternatives. Of course, the risk of not purchasing an annuity is the potential for outliving income. The concern regarding a potentially early death (and early loss of principal) can be managed and even eliminated by selecting a life annuity with a guaranteed period, which assures the annuitant that payments will continue to the estate for a minimum number of years, regardless of when the annuitant dies. This extra benefit feature comes at the cost of a slightly reduced level of income.
TIAA-CREF INSTITUTE TRENDS AND ISSUES 4
10/06
TRENDS AND ISSUES
[Link]
THE LIFE ANNUITY PRINCIPLE VARIABLE ACCOUNT The same principle also applies to a variable life annuity. The initial payment from a variable life annuity is calculated based on a specific mortality table and an Assumed Interest Rate (AIR). Each year, the payment changes based on the actual return of the underlying fund as compared to the AIR. If the fund return is higher than the AIR, the payment will increase and vice-versa (if lower.) The formula for the change is:
(1 + Fund Return) divided by (1 + AIR), less 1
For example, assuming an AIR of 4%, and an actual return of 10%, the increase will be:
(1.10 / 1.04) - 1 = 0.05769 or 5.769%.
Assuming an AIR of 4%, and an actual return of 5%, the change in payment will be:
(0.95 / 1.04) - 1 = - 0.08654 or -8.654%.
Exhibit 2a compares the variable life annuity to systematic withdrawal from a fund that earns 7% each year. As is the case with a fixed annuity, we see that by making a withdrawal from the investment account equal to the amount paid from the variable life annuity, the retiree will eventually run out of money before attaining ones life expectancy. Furthermore, as both the variable annuity and systematic withdrawal funds are invested in the same underlying assets, it would be impossible to outperform the annuity return by 2% per year (unless the annuity issuer had a very high expense charge.) Exhibit 2b and 2c compare variable life annuity payments to systematic withdrawal using actual S&P 500 investment return history. The first exhibit assumes the individual retired in 1965, while the latter assumes a retirement date in 1975. In both examples, the systematic withdrawal scenario resulted in the full depletion of investment funds before life expectancy was reached. In other words, there was more than a 50% chance that the retiree would have outlived his or her income. Thus, the life annuity principle applies to all variable life annuities, regardless of the variation in the investment return of the underlying funds.
TIAA-CREF INSTITUTE TRENDS AND ISSUES 5
10/06
TRENDS AND ISSUES
[Link]
DELAYING THE ANNUITY SIMPLE CASE As demonstrated with these exhibits, the life annuity offers a greater living income than one can safely take on their own, due to the annuity carriers ability to spread mortality gains and losses among a large group of annuitants. The (theoretical) remaining account balances of those who die are used, in essence, to pay income to those who continue to live. This concept is sometimes referred to as the benefit of survivorship. Thus, those who know they are going to live a relatively long time are best off joining the annuity pool immediately, to take advantage of those who join the pool and die at a relatively young age. Of course, no one truly knows how long he or she is going to live. In fact, those who know they are not going to live very long, due to disease or extreme illness, are advised not to purchase a life annuity. However, assuming an individual has decided to receive retirement income through a life annuity, a question often arises as to when such income should begin. In other words, one can begin receiving life annuity income immediately upon retirement or, alternatively, take systematic withdrawals from an investment account for a number of years and then begin the life annuity income. In particular, individuals who retire at a relatively young age often question the appropriateness of immediately beginning life annuity income, as they will be locking in a fair amount of their retirement account proceeds for a potentially very long time period. They are also aware that, the younger ones age, the lower the income payable from a life annuity (per $1,000 of account balance.) Lets first look at a simple example of a 65-year-old retiree who waits 5 or 10 years to purchase a life annuity. Again, in order to make a fair, apples to apples comparison, we assume that she takes withdrawals from the account equal to the life annuity payment, and that the account earns the same interest rate as the annuity. As you can see on Exhibit 3, waiting will negatively impact her future income. Delaying 5 years results in lifetime income after age 70 of $7,025 instead of the $7,390 that the age 65 annuity produces. This 5% loss might be considered a minimal cost of the call option that the delay allows. The retiree might die during these 5 years, or may have other needs that require emergency funds that would not have been available had she purchased a life annuity at age 65. A 10-year delay is more costly, resulting in a reduction of income of over 15% ($6,239 instead of $7,390). A delaying strategy will not work for variable annuities either. As shown on Exhibit 4, with a 4% AIR and 7% investment return, a 5-year delay will cost the retiree about 5%,
TIAA-CREF INSTITUTE TRENDS AND ISSUES 6
10/06
TRENDS AND ISSUES
[Link]
and a 10-year delay about 15%. Of course the delay strategy is a good idea for the estate of one who dies during the 5-year or 10-year period. Its important to note that the relative impact of delaying annuitization is more related to age than to the number of years of delay. If we considered the impact on a 55-year-old retiree, we would see that a 5-year delay (from ages 55 to 60) would only result in a 2% cut in future income, while a 10-year delay would reduce future life annuity income by about 6%. In essence, there is a much greater reduction in future income associated with postponement of annuitization past age 70 as opposed to postponement at earlier ages. DELAYING THE ANNUITY INTEREST RATES RISE As we mentioned earlier in this paper, recently some retirees have been delaying the purchase of a fixed life annuity because they feel that interest rates are very low, and they want to wait for these rates to increase before they lock in the annuity. We have already shown in Exhibit 3 that a 5-year delay strategy causes about a 5% loss of income from age 70 and on. However, that exhibit assumed that interest rates were unchanged during the entire period. What if interest rates rise during the 5-year or 10year delay? First we will address an increase only in the life annuity interest rate. In Exhibit 3, we had shown that a 5-year delay in annuitization (to Age 70), would have resulted in a remaining account balance of $84,755. In order for this account balance to generate the same level of income as otherwise payable at age 65 ($7,390 per year), the life annuity interest rate would have to be at least 5.65% per year. Thus, if the retiree thinks interest rates are headed upwards, it does not take much of an increase for this strategy to work. However, to generate a payment of $7,390 at age 75 (reflecting a 10-year delay in annuitization), the life annuity interest rate would have to be 7.67%. This does not seem like a good bet, as one would be relying on an increase of more than 250 basis points to break even. Of course, if life annuity interest rates increase, it is likely that the accumulation rate would increase as well. In Exhibit 5 we show results assuming rates climb by 25 basis points each year for 5 years. The accumulation rate is 5% during the first year, but then rises to 6% by the 5th year. The annuitization interest rate (start of Year 6) is 6.25%. As you can see, there is much to gain by delaying five years income that is more than 7%
TIAA-CREF INSTITUTE TRENDS AND ISSUES 7
10/06
TRENDS AND ISSUES
[Link]
higher than that payable to the age 65 annuitant. For the 10-year delay, assuming rates remain at 6.25%, the result is again higher than the income payable to the retiree who annuitizes at age 65, in addition to the added flexibility. Thus, we can say that if one truly knows that rates are on the rise, it can be advantageous to delay annuitization if the annuity being considered is a fixed life annuity. Note that these illustrations assume that the 5% initial interest rate applies to both accumulating and payout annuities. In actuality, annuity carriers may use a higher life annuity interest rate than the interest rate applied to investment accounts in the accumulation stage. This differential in fixed-annuity interest rates may simply reflect the underlying investments. Given the relatively long periods of investment, it is common to invest the principal underlying life annuities in long-term fixed-income instruments, while accumulating annuity funds may be invested in shorter term securities. Thus, unless there is an inverted yield curve, long-term investments yield higher rates and the rate used to determine life annuity income may often be higher. Nevertheless, even if accumulating rates are 100 basis points lower than the payout rate, an increase of 25 basis points a year for 5 years will still result in a slightly larger payment for the 5-year delay strategy. However, a 10-year delay in annuitization will result in somewhat lower payments from age 75 until death. These results are shown on Exhibit 6. Furthermore, even if rates continue to climb 25 basis points a year (indefinitely), the account will be depleted in 23 years if the retiree never elects to purchase a life annuity (see Exhibit 7). Its important to note that this issue of the potential impact of increasing interest rates does not apply to a variable life annuity. Assuming that the variable life annuity and the alternative fund are invested in the same underlying assets, any changes in interest rates will have the same relative impact on both the underlying assets and the consequent rates of return. So, if an individual is very concerned about the potential impact of rising interest rates, he or she may be best served by taking a combination of systematic withdrawals and variable life annuity income for a few years, followed perhaps by staged purchases of fixed-income life annuities. CONCLUSION While many people are reticent to give up control of their retirement assets for the promise of a lifetime stream of income, this paper demonstrates the financial tradeoff implicit in the purchase of a life annuity. If one compares the receipt of life annuity
TIAA-CREF INSTITUTE TRENDS AND ISSUES 8
10/06
TRENDS AND ISSUES
[Link]
income to equivalent withdrawals made from an investment account (assuming the same net underlying investment return), one can readily demonstrate that the latter account will be exhausted some time before reaching ones life expectancy. This type of comparison works equally well for both fixed and variable life annuities and at any interest rate or series of year-to-year investment returns (in the case of a variable annuity.) The implication of these comparisons is that, if one tries to self-annuitize and draw down the same level of income as payable under a life annuity, there will be more than a 50% chance that he or she will run of out of funds while still alive. Consequently, an individual who does not annuitize will need to lower the amount of yearly income withdrawn from his or her retirement account in order to avoid the possibility of financial ruin. Thus, one can readily conclude that a life annuity, be it fixed or variable, provides the highest level of living income available to a retired individual. Once convinced of the efficacy of choosing a life annuity for all or a portion of ones retirement income, a common follow-up question is: when should income begin? Assuming no expected significant change in interest rates, we see that a five-year delay, from age 65 to age 70, results in about a 5% loss in future income. Given the relatively small magnitude of this potential loss, some individuals may weigh this loss against the greater flexibility associated with non-annuitization. However, delaying the start of life annuity income from age 65 to age 75, a ten-year delay, can easily result in a 15% loss in future income. Given the sharper increase in mortality rates after age 70, one can conclude that it pays to begin life annuity income no later than at age 70. If one firmly believes that interest rates are almost certain to increase significantly in the very near future, there may be good reason to postpone annuitization for at least a few years. However, this will depend on various factors, including how fast and how far interest rates change, and the rate of return associated with the investment account versus the underlying rate of return of the life annuity. Of course, the potential impact of change in the interest rate environment should not affect the timing of annuitization with a variable annuity.
ABOUT THE AUTHORS Benjamin Goodman is Director, Actuarial Consulting Services of TIAA-CREF. His primary focus is on the development of participant advice and the pricing of retirement plan services. Benny graduated from Touro College in 1984 with a B.S. degree in
TIAA-CREF INSTITUTE TRENDS AND ISSUES 9
10/06
TRENDS AND ISSUES
[Link]
computer science with a specialization in mathematics. He is a Fellow of the Society of Actuaries and a Member of the American Academy of Actuaries. Michael Heller is Vice President, Actuarial Consulting Services of TIAA-CREF. He is responsible for managing a number of actuarial functions, primarily focused on providing advice and assistance in the design and funding of both defined benefit and defined contribution retirement plans. Michael graduated from the City College of New York in 1972 with a B.S. degree in mathematics. He is a Fellow of the Society of Actuaries, a Member of the American Academy of Actuaries and an Enrolled Actuary.
TIAA-CREF INSTITUTE TRENDS AND ISSUES 10
10/06
TRENDS AND ISSUES
[Link]
EXHIBIT 1: LIFE ANNUITY PRINCIPLE FIXED ANNUITY
TIAA-CREF INSTITUTE TRENDS AND ISSUES 11
10/06
TRENDS AND ISSUES
[Link]
EXHIBIT 2A: LIFE ANNUITY PRINCIPLE VARIABLE ANNUITY AT 7%
TIAA-CREF INSTITUTE TRENDS AND ISSUES 12
10/06
TRENDS AND ISSUES
[Link]
EXHIBIT 2B: LIFE ANNUITY PRINCIPLE S&P 1965-1989
TIAA-CREF INSTITUTE TRENDS AND ISSUES 13
10/06
TRENDS AND ISSUES
[Link]
EXHIBIT 2C: LIFE ANNUITY PRINCIPLE S&P 1975-1999
TIAA-CREF INSTITUTE TRENDS AND ISSUES 14
10/06
TRENDS AND ISSUES
[Link]
EXHIBIT 3: DELAYING THE ANNUITY FIXED
TIAA-CREF INSTITUTE TRENDS AND ISSUES 15
10/06
TRENDS AND ISSUES
[Link]
EXHIBIT 4: DELAYING THE ANNUITY VARIABLE
TIAA-CREF INSTITUTE TRENDS AND ISSUES 16
10/06
TRENDS AND ISSUES
[Link]
EXHIBIT 5: DELAYING THE ANNUITY WITH A RISE IN INTEREST RATES (FIXED)
TIAA-CREF INSTITUTE TRENDS AND ISSUES 17
10/06
TRENDS AND ISSUES
[Link]
EXHIBIT 6: DELAYING THE ANNUITY WITH A RISE IN INTEREST RATES (FIXED)
TIAA-CREF INSTITUTE TRENDS AND ISSUES 18
10/06
TRENDS AND ISSUES
[Link]
EXHIBIT 7: RISING INTEREST RATES
TIAA-CREF INSTITUTE TRENDS AND ISSUES 19
