the future of retirement for private sector employees …€¦ · a dc plan impacts retirement...

The opinions expressed in this presentation are those of the speaker. The International Societydisclaims responsibility for views expressed and statements made by the program speakers.

Jack VanDerhei, Ph.D., CEBSResearch DirectorEmployee Benefit Research InstituteWashington, D.C.

The Future of Retirement for Private Sector Employees in the U.S.

Key Points from Today’s Presentation1. Projections of how long Baby Boomers and Gen Xers will need to work for a

reasonable chance of adequate retirement income under various scenarios

• Review of Retirement Security Project Model (RSPM)

• Retirement Readiness Ratings measures those “at risk” for inadequate retirement income

• What can be done for those “at risk”?

• Increase savings during accumulation phase

• What if the additional savings rate is too high to be feasible?

• Deferring retirement age

• We know mathematically that this will improve the probability of adequacy

• But for how many households and at what retirement ages?

• RSPM simulations for percentage of households with adequate retirement income by retirement age

• 50, 70 and 80 percent of simulated life paths

• Value of DC participation after age 64

2. Summary of the causes of inadequate retirement savings and measurement of how proper education and/or plan design will increase employees’ probability of a successful retirement.

3. An assessment of retirement policy in the US: does it facilitate people who want to save for retirement?

5B-1

Brief Chronology of the EBRI/ERF Retirement Security Projection Model™

• 2001, Oregono Simulated retirement wealth vs. ad hoc thresholds

for retirement expenses• 2002, Kansas and Massachusetts

o Full stochastic retiree model: Investment and Longevity risk, Nursing home and home health care costs

o Net housing equity• 2003, National model

o Expanded to full national sample• 2004, Senate Aging testimony

o Impact of everyone saving another 5 percent of compensation

• 2004, EBRI Policy forumo Impact of annuitizing defined contribution/IRA

balances• 2006, EBRI Issue Brief

o Evaluation of defined benefit freezes on participants

• 2006, EBRI Issue Brief o Converted into a streamlined individual version

for the ballpark estimate Monte Carlo• 2008, EBRI policy forum

o Impact of converting 401(k) plans to automatic enrollment

• 2009, Pension Research Councilo Winners/losers analysis of defined benefit

freezes and enhanced defined contribution employer contributions provided as a quid pro quo

• 2010, EBRI Issue Brief (April)o Impact of modification of employer

contributions when they convert to automatic enrollment for 401(k) plans

o 2010, EBRI Issue Brief (July)o Updated model to 2010, included

automatic enrollment for 401(k) planso 2010, EBRI Notes (Sept and Oct)

o Analyzes how eligibility for participation in a DC plan impacts retirement income adequacy

o Computes Retirement Savings Shortfalls for Boomers and Gen Xers

o 2010, Senate HELP testimonyo Analyzes the relative importance of

employer‐provided retirement benefits and Social Security

o 2011, EBRI Issue Brief (February)o Analyzes the impact of the 2008/9 crisis in

the financial and real estate markets on retirement income adequacy

Modeling Innovations in the EBRI/ERF Retirement Security Projection Model

• Pension plan parameters coded from a time series of several hundred plans.

• 401(k) asset allocation and contribution behavior based on individual administrative records

o Annual linked records dating back to 1996

o 2009: More than 20 million employees in 50,000 plans.

• Stochastic modeling of nursing facility care and home based health care.

5B-2

Retirement Income

• Limited to income produced by

• Public and private retirement plans (including IRAs)

• Social Security

• Housing equity• Not used in the baseline results shown today

• Original baseline assumes retirement income commences at age 65 • Today’s presentation will analyze the value of deferring retirement for baby boomers and gen Xers

Retirement Expense Assumptions

• Decomposed total expenditures for retirees into:o Those that are deterministic:

Food, apparel and services, transportation, entertainment, reading and education, housing, and basic health expenditures.

o Those that are stochastic: Home health care and nursing home care.

• Performed annual simulations on U.S. families with a retiree to determine if each retiree would: o Require home health care, o Enter a nursing home, o Die, or o Continue to survive without incurring any of these stochastic

health costs.

5B-3

EBRI 2010 Retirement Readiness Ratings vs. National Retirement Risk Index (NRRI)

early boomers late boomers gen xers

NRRI w 2009 market 41% 48% 56%

NRRI w LTC 52% 64% 71%

RSPM baseline 47% 44% 45%

0%10%20%30%40%50%60%70%80%

Percentage of population “at risk” for inadequate retirement income, by age cohort (baseline assumptions)

Sources: EBRI/ERF Retirement Security Projection Model ™ version 100504e; “The National Retirement Risk Index: After the Crash,” Center for Retirement Research at Boston College, October 2009; “Long-term Care Costs and the National Retirement Risk Index,” Center for Retirement Research at Boston College, March 2009

Impact of “salary” on at risk probability

Lowest 2 3 Highest

Baseline RSPM 76% 51% 35% 20%

0%

10%

20%

30%

40%

50%

60%

70%

80%

Percentage of population “at risk” for inadequate retirement income, by age‐specific remaining career

income quartiles (baseline assumptions)

Source: EBRI/ERF Retirement Security Projection Model ™ version 100504e

5B-4

Impact of age and future years of eligibility for participation in a defined contribution plan on at risk probabilities

early boomers late boomers gen xers

0 54% 53% 60%

1‐9 38% 40% 45%

10‐19 27% 32%

20+ 20%

0%10%20%30%40%50%60%70%

Percentage of population “at risk” for inadequate retirement income, by age cohort and future years eligible for participation in a

defined contribution plan (baseline assumptions)

Source: EBRI/ERF Retirement Security Projection Model ™ version 100504e

Analyzing the importance of retirement age

• Figures 20‐22 of the EBRI July 2010 Issue Brief presents additional savings (expressed as a percent of compensation) needed to achieve various probabilities of success for retirement age at 65• Unfortunately, the results for many combinations of age/income

cohorts would be too high to be feasible

• We have always known that the other possibility may be to defer retirement age• Assumes this is feasible

• I.E., no health problems for worker or spouse, job still available

• The Retirement Confidence Survey has consistently found that a large percentage of retirees leave the work force earlier than planned

• 45 percent in 2011

• Easy to do stylized examples, but what would the impact be for baby boomers and gen Xers?

5B-5

Assumptions for the new analysis• Updated certain financial market and employee behavioral assumptions from the 2010 RSPM

• Started with a set of assumptions that would be most favorable to deferring retirement age (will relax these in future work)

• Wages

• No age/wage curves after 64; grows at average national wage growth

• Job change, disability, unemployment

• None after 64

• Nursing home or home health care expenses for the worker

• None prior to retirement

• Defined contribution plans

• Employee contributions continue at the age 64 contribution rate

• Any auto escalation for AE plans is turned off at this age

• Employer contributions

• Match rates are assumed to remain constant

• Nonelective contributions continue at the age 64 contribution rate

• Social Security

• Initial receipt deferred until retirement age or age 70

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

65 67 69 71 73 75 77 79 81 83 85

Percentage

of Households

Retirement Age

Percentage of Baby boom and Gen X Households Simulated to Have Adequate* Retirement Income for at Least 50 Percent of Simulated Life

Paths After Retirement Age by Pre‐Retirement Income Quartiles

lowest

2

3

highest

Source: EBRI Retirement Security Projection Model® versions110410i.* An individual or family is considered have "adequate” retirement income in this version of the model if their aggregate resources in retirement are sufficient to meet aggregate minimum retirement expenditures defined as a combination of deterministic expenses from the Consumer Expenditure Survey (as a function of income) and some health insurance and out‐of‐pocket health related expenses, plus stochastic expenses from nursing home and home health care expenses (at least until the point they are picked up by Medicaid). The resources in retirement will consist of Social Security (either status quo or one of the specified reform alternatives), account balances from defined contribution plans, IRAs and/or cash balance plans, annuities from defined benefit plans (unless the lump‐sum distribution scenario is chosen) and (in some cases) net housing equity (either in the form of an annuity or as a lump‐sum distribution). This version of the model is constructed to simulate "basic" retirement income adequacy; however, alternative versions of the model allow similar analysis for

5B-6

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

65 67 69 71 73 75 77 79 81 83 85

Percentage

of Households

Retirement Age



lowest

2

3

highest

Source: EBRI Retirement Security Projection Model® versions110410i.* An individual or family is considered have "adequate” retirement income in this version of the model if their aggregate resources in retirement are sufficient to meet aggregate minimum retirement expenditures defined as a combination of deterministic expenses from the Consumer Expenditure Survey (as a function of income) and some health insurance and out‐of‐pocket health related expenses, plus stochastic expenses from nursing home and home health care expenses (at least until the point they are picked up by Medicaid). The resources in retirement will consist of Social Security (either status quo or one of the specified reform alternatives), account balances from defined contribution plans, IRAs and/or cash balance plans, annuities from defined benefit plans (unless the lump‐sum distribution scenario is chosen) and (in some cases) net housing equity (either in the form of an annuity or as a lump‐sum distribution). This version of the model is constructed to simulate "basic" retirement income adequacy; however, alternative versions of the model allow similar analysis for replacement rates, standard of living and other ad hoc thresholds.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

65 67 69 71 73 75 77 79 81 83 85

Percentage

of Households

Retirement Age



lowest

2

3

highest


5B-7

0%

5%

10%

15%

20%

25%

65 69 75 84

Value of DC Participation After Age 64: Increase in the Percentage of Baby boom and Gen X Households Simulated to Have Adequate*

Retirement Income for at Least 50 Percent of Simulated Life Paths After Retirement Age by Pre‐Retirement Income Quartiles

lowest

2

3

highest


0%

5%

10%

15%

20%

25%

65 69 75 84



lowest

2

3

highest


5B-8

0%

5%

10%

15%

20%

25%

65 69 75 84



lowest

2

3

highest


Questions to consider

• How have consistent 401(k) participants fared?• Why is it important to look at the consistent participants?

• Given current contribution and asset allocation behavior, what is the potential for the retirement savings for those covered by 401(k) plans for an entire career?

• How would the 20/20 limits modify this?• National Commission on Fiscal Responsibility and Reform

• Cap tax‐preferred retirement contributions to lower of $20,000 or 20% of income

• Would these results differ for automatic enrollment?

5B-9

401(k) Account Balancesa Among 401(k) Participants Present From Year-End 1999 Through September 30, 2011b

$0

$20,000

$40,000

$60,000

$80,000

$100,000

$120,000

$140,000

$160,000

$180,000

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Source: Tabulations from EBRI/ICI Participant-Directed Retirement Plan Data Collection Project.a Account balances are participant account balances held in 401(k) plans at the participants' current employers and are net of plan loans. Retirement savings held in plans at previous employers or rolled over into IRAs are not included. b The analysis for 1999 through 2009 is based on a group of 1.6 million participants with account balances at the end of each year from 1999 through 2009. The values for 2010 and September 30, 2011 are EBRI estimates

>10–20 >20–30 >30

40s $98,743 $136,379

50s $116,009 $186,580 $170,384

60s $111,689 $176,281 $192,032

$‐

$50,000

$100,000

$150,000

$200,000

$250,000

Average Account Balances Among 401(k) Participants Present From Year‐End 1999 Through Year‐End 2009, by Age and Tenure

Source: Tabulations from the EBRI/ICI Participant‐Directed Retirement Plan Data Collection Project.The analysis is based on a group of 1.6 million participants with account balances at the end of each year from 1999 through 2009. Age and tenure groups are based on participant age and tenure at year‐end 2009.

5B-10

Change In Average Account Balances (by Age and Tenure) From January 1, 2010 – August 1, 2011 Among 401(k) Participants with Account Balances as

of Dec. 31, 2009

Sources: 2009 Account Balances: Tabulations from EBRI/ICI Participant-Directed Retirement Plan Data Collection Project; 2011 Account Balances: EBRI estimates. The analysis is based on all participants with account balances at the end of 2009 and contribution information for that year.

22

5B-11

Lowest 2 3 Highest

baseline 9.16 11.35 13.08 13.47

alternative return scenario 4.31 5.03 5.50 5.87

‐

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

Multiple of final earnings

Median 401(k) account balances at age 67 (expressed as a multiple of final earnings) for participants currently ages 25‐29 by income quartile

Source: EBRI Retirement Security Projection Model Version 110503c.The simulated rates of return for the baseline and alternative return scenario are the same as in VanDerhei and Copeland (July 2010). This version of the analysis models 401(k) participants who are not automatically enrolled and assumes no job turnover, withdrawals or loan defaults. The full stochastic nature of the model will be included in future analysis.

Lowest 2 3 Highest

baseline 53% 65% 75% 77%

alternative return scenario 25% 29% 32% 34%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Real replacement rates

Median real replacement rates at age 67 from 401(k) balances for participants currently ages 25‐29 by income quartile

Source: EBRI Retirement Security Projection Model Version 110503c.The simulated rates of return for the baseline and alternative return scenario are the same as in VanDerhei and Copeland (July 2010). This version of the analysis models 401(k) participants who are not automatically enrolled and assumes no job turnover, withdrawals or loan defaults. The full stochastic nature of the model will be included in future analysis.

5B-12

0%

2%

4%

6%

8%

10%

12%

14%

26‐35 36‐45 46‐55 56‐65

Average percent reductions in 401(k) account balances at Social Security NRA by imposing 20/20 limits in 2012, by age and age‐specific salary quartiles

Lowest

2

3

Highest

Source: EBRI Retirement Security Projection Model Version 110503c.NB: this simulation only models the financial impact of the expected reduction in 401(k) contributions for employees who are notautomatically enrolled by imposing the new limits and does not attempt to assess behavioral modifications on the part of either the plan sponsor nor the employees assumed to be eligible for participation in the plan. The simulated rates of return are the same as in VanDerhei and Copeland (July 2010). This version of the analysis assumes no job turnover, withdrawals or loan defaults. The full stochastic nature of the model will be included in future analysis.

Would these results differ for automatic enrollment?

• The previous results assumed none of the participants were automatically enrolled

• Given the current modeling assumption of no job change, it would be very difficult to provide a valid comparison using automatic enrollment• Currently very little, if any, information that can be used to track what

automatically enrolled participants with automatic escalation of contributions would do upon job change

• For example: if a participant has already been escalated to 8 percent of compensation and upon job change was automatically enrolled into another 401(k) plan, would they “remember” where they had been or decrease contributions to the default rate of the new plan

• The EBRI/ICI Participant‐Directed Retirement Plan Data Collection Project should allow at least preliminary data to inform these assumptions with the year‐end 2010 analysis.

5B-13

SUMMARY/CONCLUSIONS

For more information• All publications are available for free download at: www.ebri.org

• The Impact of Deferring Retirement Age on Retirement Income Adequacy, June 2011EBRI Issue Brief 358– www.tinyurl.com/ebrijune11ib

• Capping Tax‐Preferred Retirement Contributions: Preliminary Evidence of the Impact of the National Commission on Fiscal Responsibility and Reform Recommendations, July 2011 EBRI Notes, Vol. 32, No. 7– www.tinyurl.com/ebrijuly11notes

5B-14

the future of retirement for private sector employees …€¦ · a dc plan impacts retirement...

Documents