Sustainable Spending for Retirement

What s Different About Retirement? RETIREMENT BEGINS WITH A PLAN TM Sustainable Spending for Retirement Presented by: Wade Pfau, Ph.D., CFA Reduced earnings capacity Visible spending constraint Heightened investment risk Unknown longevity Spending shocks Compounding inflation Declining cognitive abilities Key Retirement Risks Pre-Retirement vs. Retirement 1

Longevity Risk Wealth Glidepath: Fixed Returns Wealth Glidepath Over a 30-Year Retirement For a 6.3% Initial Withdrawal Rate, 5.1% Compounded Real Return, Inflation Adjustments Using SBBI Data, 1926-2016, S&P 500 and Intermediate Term Government Bonds Research Real Life Use Source: Society of Actuaries, 2012 Individual Annuity Mortality tables, with projections for 2016 Market Risk: Fixed vs. Random Returns Wealth Glidepath Over a 30-Year Retirement For a 6.3% Initial Withdrawal Rate, 50/50 Asset Allocation, Inflation Adjustments Using SBBI Data, 1926-2016, S&P 500 and Intermediate Term Government Bonds Real Life Use Managing Sequence Risk Spend Conservatively Spending Flexibility Reduce Volatility Buffer Assets Avoid Selling at Losses 2

Planning Horizon (Age) 3/14/2017 Retirement Income Philosophies Prioritization Among Goals Investment Approach Measuring Retirement Success Safe Withdrawal Rate Probability-Based Focus on overall lifestyle spending goals Focus on a total returns investing for the entire financial portfolio to maximize returns for an acceptable level of volatility Focus on determining an acceptable probability of failure for the overall retirement plan The historical record suggests that 4% or 4.5% is about as bad as it gets Safety-First Distinguish retirement spending goals between essential needs and discretionary expenses Match assets to different goals so that risk levels are compatible. Allows a wider role for holding individual bonds and using income annuities. Seeks to minimize harms in worst-case scenarios by avoiding failure for essential needs Unknown and unknowable. Risky assets are inherently risky and the historical success of a strategy does not provide sufficient confidence Safe Withdrawal Rates & The 4% Rule William Bengen Journal of Financial Planning, October 1994 PMT(rate,nper,pv,fv,type) PMT Formula rate: use bond yields, or higher number to amortize the upside nper: fixed conservatively at start of retirement, or dynamically updated with life expectancy throughout retirement pv: fixed at start of retirement, or dynamically updated throughout Relationship between Interest Rates, Planning Horizon, and Sustainable Spending for a 65-year old with $1 million Fixed Return 0% 0.5% 1.5% 2.5% 3.5% 4.5% 5.5% 70 $200,000 $202,000 $205,999 $209,997 $213,992 $217,982 $221,968 75 $100,000 $102,259 $106,832 $111,472 $116,175 $120,937 $125,751 80 $66,667 $69,019 $73,837 $78,797 $83,889 $89,104 $94,432 85 $50,000 $52,404 $57,385 $62,583 $67,982 $73,566 $79,317 90 $40,000 $42,440 $47,550 $52,952 $58,622 $64,535 $70,663 95 $33,333 $35,800 $41,024 $46,612 $52,533 $58,748 $65,218 100 $28,571 $31,060 $36,388 $42,152 $48,308 $54,804 $61,588 105 $25,000 $27,508 $32,933 $38,865 $45,244 $52,003 $59,071 110 $22,222 $24,747 $30,266 $36,359 $42,950 $49,954 $57,281 3

Planning Horizon (Age) Planning Horizon (Age) 3/14/2017 Relationship between Interest Rates, Planning Horizon, and Sustainable Spending for a 65-year old with $1 million 4% Rule: 30-year time horizon Real fixed compounded return: 1.3% Fixed Return 0% 0.5% 1.5% 2.5% 3.5% 4.5% 5.5% 70 $200,000 $202,000 $205,999 $209,997 $213,992 $217,982 $221,968 75 $100,000 $102,259 $106,832 $111,472 $116,175 $120,937 $125,751 80 $66,667 $69,019 $73,837 $78,797 $83,889 $89,104 $94,432 85 $50,000 $52,404 $57,385 $62,583 $67,982 $73,566 $79,317 90 $40,000 $42,440 $47,550 $52,952 $58,622 $64,535 $70,663 95 $33,333 $35,800 $41,024 $46,612 $52,533 $58,748 $65,218 100 $28,571 $31,060 $36,388 $42,152 $48,308 $54,804 $61,588 105 $25,000 $27,508 $32,933 $38,865 $45,244 $52,003 $59,071 110 $22,222 $24,747 $30,266 $36,359 $42,950 $49,954 $57,281 Relationship between Interest Rates, Planning Horizon, and Sustainable Spending for a 65-year old with $1 million Fixed Return 0% 0.5% 1.5% 2.5% 3.5% 4.5% 5.5% Seeking Risk Premium 70 $200,000 $202,000 $205,999 $209,997 $213,992 $217,982 $221,968 75 $100,000 $102,259 $106,832 $111,472 $116,175 $120,937 $125,751 80 $66,667 $69,019 $73,837 $78,797 $83,889 $89,104 $94,432 85 $50,000 $52,404 $57,385 $62,583 $67,982 $73,566 $79,317 90 $40,000 $42,440 $47,550 $52,952 $58,622 $64,535 $70,663 95 $33,333 $35,800 $41,024 $46,612 $52,533 $58,748 $65,218 100 $28,571 $31,060 $36,388 $42,152 $48,308 $54,804 $61,588 105 $25,000 $27,508 $32,933 $38,865 $45,244 $52,003 $59,071 110 $22,222 $24,747 $30,266 $36,359 $42,950 $49,954 $57,281 Basis for the 4% Rule Basis for the 4% Rule: Asset Allocation Maximum Sustainable Withdrawal Rates For 50/50 Asset Allocation, 30-Year Retirement, Inflation Adjustments, No Fees Using SBBI Data, 1926-2016, S&P 500 and Intermediate Term Government Bonds 4

Connection between SAFEMAX and Stock Allocation Connection Between Median Remaining Real Wealth After 30 Years and Stock Allocation Trinity Study Portfolio Success Rates Inflation-Adjusted Withdrawals For Various Withdrawal Rates, Asset Allocations, and Retirement Durations Using Ibbotson s Stocks, Bonds, Bills, and Inflation Data, 1926-2016, S&P 500 and Intermediate-Term Government Bonds Portfolio Success Rates for a 4% Withdrawal Rate 30-Year Retirement, Inflation Adjustments Using SBBI Data, 1926-2016, S&P 500 and Intermediate Term Government Bonds 5

Sustainable Withdrawal Rates Allowing for 10% Failure Probability For a 30-Year Retirement, Inflation Adjustments Efficient Frontier Based on SBBI Data, 1926-2016, S&P 500 and Intermediate-Term Government Bonds Underlying Philosophy of 4% Rule Objective: Overall lifestyle spending goal Failure = depletion of financial portfolio Desire smooth spending & appetite for market risk Confidence in the Historical Record as Precedent Market Risk Management: Total returns investing strategy; Asset allocation to minimize failure probability Longevity Risk Management: Conservative planning horizon (30 Years) sufficiently beyond life expectancy Issues with 4% Rule Assumptions Longevity risk remains Market/sequence risk could create new worst outcome At the same time, spending is usually too conservative Need to consider entire household balance sheet Assumes a constant (inflation-adjusted) spending need Unique cause of sequence risk More efficient to adjust spending based on portfolio performance Real world retirees may reduce spending with age Requires asset & distribution management into advanced age The 4% Rule - Ten Important Issues 6

#1 Is US historical data representative of future? Is U.S. Historical Data Sufficiently Representative? Time period since 1926 is too short (less than 3 independent 30 year periods) The 20 th century United States is a rather remarkable and unparalleled period in world history should this experience really be projected forward? Current market conditions matter much more for sustainable spending rates than historical averages We really have very little basis for knowing the suitability of our simulation input assumptions Maximum Sustainable Withdrawal Rates for Retirees Global Returns Dataset, 1900-2015 Asset Allocation: 50% Stocks & 50% Bills Importance of Current Market Conditions Sources of bond returns: Initial Bond Yield Subsequent Yield Changes Sources of stock returns: Dividend Income Growth of the Underlying Earnings Changes in the Valuation Multiples Placed on Those Earnings 7

10-Year Treasury Yields at the Start of Each Year, 1871-2017 Robert Shiller s PE10 Robert Shiller s Cyclically-Adjusted Price Earnings Ratio (PE10) In January of Each Year, 1880-2017 8

Importance of Current Market Conditions 2 approaches to incorporate current market conditions: 1) Regression analysis to explain sustainable withdrawal rates with retirement date market conditions 2) Incorporate current conditions (instead of historical averages) into input assumptions for Monte Carlo simulation analysis Actual and Predicted Maximum Sustainable Withdrawal Rates For 50/50 Asset Allocation, 30-Year Retirement Period Estimated for Years 1881-1987, Explanatory Variables: PE10 and 10-Year Treasury Rate Portfolio Success Rates for a 4% Withdrawal Rate Rolling vs. Monte Carlo Simulations For a 30-Year Retirement, Inflations Adjustments #2 Integrating Pre- and Post-Retirement Periods 9

Safe Savings Rates Traditional wealth accumulation targets do not provide the most effective framework for retirement planning. Even if they did, financial market volatility makes it hard to judge whether one is on track to meet such targets. Rather, integrate the pre-retirement and post-retirement periods together to see what must be done prior to retirement in order to accumulate enough to afford one s desired retirement expenditures. Minimum Necessary Savings Rate to Accumulate 12.5x Final Salary Maximum Sustainable Withdrawal Rates For 50/50 Asset Allocation, 30-Year Work Period, 30-Year Retirement Period Relationship Between Valuations, Savings Rates, and Withdrawal Rates For 50/50 Asset Allocation, 30-Year Work Period, 30-Year Retirement Period overvalued undervalued overvalued undervalued Minimum Necessary Savings Under Isolated and Integrated Approaches For 50/50 Asset Allocation, 30-Year Work Period, 30-Year Retirement Period, 50% Replacement Rate 10

#3 Earning the Index Returns Portfolio Underperformance and Poor Investment Decisions Impact of a 1% Annual Account Underperformance Historical Simulations 1966 retirees are the source of our SAFEMAX With market index returns, 1966 retiree: 4.03% withdrawal rate. With 1% underperformance, 1966 retiree: Withdrawal rate falls by 0.48 percentage points to 3.56%. Reduced spending power by 11.9%. Across the historical period, fees caused the maximum sustainable withdrawal rates to fall on average by 0.65 percentage points or 11% of spending power. Buying High and Selling Low One example: high valuations result in a 0% stock allocation, medium valuations result in 50% stocks, and low valuations result in 75% stocks. This strategy lowers sustainable withdrawal rates compared to a fixed 50/50 allocation in almost all years. The SAFEMAX falls to 3.06%, compared to a 3.95% SAFEMAX with the fixed 50/50 allocation. (This analysis was with a different dataset) 22.5% spending reduction at the SAFEMAX; 29 years in which sustainable withdrawal rate falls below 4% Maximum Sustainable Withdrawal Rates, 50% Stock Allocation and Market Mistiming Asset Allocation, 30-Year Retirement 11

Role of Taxes Standard safe withdrawal rate assumption is a taxdeferred financial portfolio #4 Taxes Taxable portfolio: Taxes must also be paid on reinvested dividends, interest, and capital gains as they accrue. This limits the chance for the portfolio to earn compound growth on those removed tax payments. William Bengen estimated: 20% effective tax rate reduces his SAFEMAX from 4.15% to 3.67% #5 Bequests and a Margin of Safety SAFEMAX for Three Scenarios, 50/50 Portfolio The classic case, no wealth after 30 years: 4.03% Nominal value of retirement date wealth is preserved after 30 years: SAFEMAX = 3.77% Real inflation-adjusted value of retirement date wealth is preserved after 30 years: SAFEMAX = 2.72% 12

Portfolio Real Return Portfolio Real Return 3/14/2017 The 4% rule assumes only a few asset classes. #6 Broader Portfolio Diversification William Bengen s initial research: large-cap stocks and intermediate-term government bonds What matters for sustainable withdrawal rates are the interaction of portfolio returns and volatilities. Including more asset classes can allow for different portfolio characteristics and potentially a portfolio with better return/volatility characteristics can be found. Large-Cap Stocks & Intermediate-Term Government Bonds (based on historical data) Optimal Portfolio: Large-Cap: 45% ITGB: 55% Withdrawal Rate: 4.3% 10 9 8 7 6 5 4 3 2 1 0-1 -2 Figure 2 Sustainable Withdrawal Rates with 10% Failure Rate for 30-Year Retirement Horizon 8 5 6 7 4 5 6 3 4 5 4 3 2 3 2 1 2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Standard Deviation of Portfolio Real Returns Adding Small-Cap Stocks Optimal Portfolio: Large-Cap: 15% Small-Cap: 27% ITGB: 58% Withdrawal Rate: 4.6% 10 9 8 7 6 5 4 3 2 1 0-1 -2 Figure Sustainable Withdrawal Rates with 10% Failure Rate for 30-Year Retirement Horizon The Effect of Adding Small-Capitalization Stocks to the Portfolio Mix 8 5 6 7 4 5 6 3 4 5 4 3 2 3 2 1 2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Standard Deviation of Portfolio Real Returns 13

#7 Actual Retiree Spending Patterns Go-Go (65-74) Slow-Go (75-84) No-Go (85+) Blanchett s Retirement Spending Smile The Estimated Path of Real Retirement Spending for a $100,000 Initial Budget at 65 Understanding the Path of Real Retirement Spending vs. Constant Inflation- Adjusted Spending Strategy 14

Maximum Sustainable Withdrawal Rates (MWRs) Cases: (1) Blanchett s Retirement Spending Smile, (2) Constant Inflation-Adjusted Spending Strategy For 50/50 Asset Allocation, 30- Year Retirement Duration, Inflation Adjustments #8 Adjusting Spending to Market Returns Using SBBI Data, 1926-2016, S&P 500 and Intermediate- Term Government Bonds Variable Spending in Retirement Extreme #1: Constant Inflation-Adjusted Spending (no adjustment to portfolio balance) Extreme #2: Fixed Percentage of Remaining Balance (full adjustment to portfolio balance) Variable spending strategies seek compromise between these extremes by avoiding too many spending cuts while also protecting against the risk that spending must subsequently fall to uncomfortably low levels. Constant Inflation-Adjusted Spending Amounts Advantages: -Smooth consumption path Disadvantages: -Wealth can be depleted 15

Constant Percentage of Remaining Assets Variable Spending Strategies in Retirement Advantages: -Since spending adjusts, cannot deplete wealth -Spending can increase on upside -Helps mitigate sequence of returns risk -More efficient in the sense of using a greater percentage of assets Disadvantages: -Variable and unpredictable spending path Decision Rule Methods Bengen's Constant Inflation-Adjusted Spending (1994) Bengen's Fixed-Percentage Withdrawals (2001) Bengen's Floor-and-Ceiling Withdrawals (2001) Guyton and Klinger's Decision Rules (2006) David Zolt's Target Percentage Adjustment (2013) Actuarial Methods RMD Spending Rules PMT Formula (ex. Waring and Siegel (2015); Steiner (2014); Bogleheads) Monte-Carlo PMT Formulas: Frank, Mitchell, and Blanchett Age-Based 3D Model (2011, 2012a, 2012b); Blanchett, Maciej, and Chen Mortality- Updating Constant Probability of Failure (2012); David Blanchett's Simple Formula (2013) Annuitize the Floor & Invest for Discretionary Evaluating Variable Spending Strategies - Problems with Failure Rates - Focus is only on initial spending rate Pre-failure spending is ignored Some variable spending strategies technically cannot fail Portfolio failure ignores available of income sources outside portfolio Failure rate ignores upside potential No single number can summarize all of the characteristics of a retirement income strategy! Evaluating Variable Spending Strategies - XYZ Formula - The retiree willingly accepts: X% probability that Wealth falls below $Y (in inflation-adjusted terms) by year Z of retirement 16

The Relationship Between the Retirement Living Standard and Market Returns Spending Patterns for Retirement Acceptable Spending Volatility Evaluating Variable Spending Strategies - Full Distribution of Outcomes - 1. Initial Spending Rate 2. Range of Spending Amounts 3. Range of Remaining Wealth XYZ Formula calibrates different strategies so that their distribution of outcomes can be more fairly compared 17

Table 3 Sustainable Spending Rates from an Investment Portfolio over 30 years, For a 65-Year Old Couple Strategies Allowing for a 10% Chance That Spending has fallen below an inflation-adjusted $250 by Year 30 Using a 50/50 Portfolio of Stocks and Bonds Initial Wealth Level = $100,000 Initial Spending Spending Strategy Rate Constant Inflation-Adjusted 2.90% Spending Modified Inflation-Adjusted 3.32% Spending (Inflation - 1%) Fixed Percentage 12.38% Bengen's Floor-and-Ceiling Rule 3.35% Guyton and Klinger's Decision Rules 5.32% Zolt Target Percentage Adjustment: 3.91% No CPI Increase Required Minimum Distribution 8.30% Spending Rule (Modified) Required Minimum Distribution 3.23% Spending Rule PMT Formula with 10-Year Treasury 4.34% Yields and Dynamic Life Expectancy Annuitize Floor & Aggressive 5.21% Discretionary Spending Percentile of Real Spending Real Spending Real Spending Real Remaining Wealth After Distribution in 10 years in 20 years in 30 years 30 Years 90th $2,900 $2,900 $2,900 $276,984 50th $2,900 $2,900 $2,900 $77,111 10th $2,900 $2,900 $2,170 $284 90th $3,048 $2,768 $2,511 $268,130 50th $3,040 $2,756 $2,497 $74,916 10th $3,032 $2,744 $2,462 $475 90th $6,936 $3,038 $1,338 $10,047 50th $4,488 $1,553 $582 $4,280 10th $2,936 $836 $262 $1,856 90th $4,020 $4,020 $4,020 $220,045 50th $2,914 $2,848 $2,848 $69,177 10th $2,848 $2,848 $2,016 $230 90th $5,793 $5,867 $6,163 $99,092 50th $4,073 $3,360 $3,121 $26,370 10th $2,767 $1,971 $846 $307 90th $3,910 $3,910 $3,910 $194,779 50th $3,910 $3,624 $3,181 $41,597 10th $3,195 $2,318 $1,049 $437 90th $8,292 $5,144 $1,332 $3,584 50th $5,365 $2,630 $579 $1,527 10th $3,510 $1,415 $260 $662 90th $5,559 $8,322 $9,915 $85,206 50th $3,597 $4,255 $4,311 $36,297 10th $2,353 $2,289 $1,937 $15,740 90th $6,792 $7,323 $5,970 $46,537 50th $4,393 $3,876 $2,715 $20,772 10th $2,883 $2,109 $1,277 $9,283 90th $5,651 $5,720 $5,992 $92,222 50th $4,051 $3,387 $3,159 $24,541 10th $2,835 $2,094 $820 $286 #9 Retirement Time Horizon Notes: Analysis assumes that withdrawals are made at the start of each year, a 0.5% portfolio administrative fee is deducted at the end of each year, and market return simulations are based on capital market assumptions detailed in the appendix. Planning for Specific Time Horizon Longer time horizons guide optimal solutions toward: Lower withdrawal rates Higher stock allocations A stronger case for guaranteed income retirement products Relationship Between SAFEMAX and Retirement Duration For a 50/50 Asset Allocation, Inflations- Adjusted Spending Using SBBI Data, 1926-2016, S&P 500 and Intermediate-Term Government Bonds 18

Withdrawal Rate 10 Years 15 Years 20 Years 25 Years 30 Years 35 Years 40 Years 3/14/2017 SAFEMAX (Planning Age=100) Compared with IRS RMD Percentages Withdrawal Rates, Asset Allocation, and Time Horizon Monte Carlo Simulations Figure 7 Sustainable Withdrawal Rates with 10% Failure Rate for Various Retirement Horizons Using SBBI Historical Parameters 11 10 9 8 7 6 5 4 3 2 1 0 0 10 20 30 40 50 60 70 80 90 100 Stock Allocation #10 Time Segmentation Strategies Taxonomy of Retirement Income Bond Ladders Types of Bond Ladders One-Time Ladders Fixed Short Term Fixed Medium Term Fixed Long Term Rolling Ladders Automatic Market-Based Personalized Potential Use Build a Social Security Delay Bridge Twenty-Year Bond Ladder followed by Deferred Income Annuity to cover remainder of lifetime Thirty Year Bond Ladder as Source of "Lifetime" Retirement Income Each year purchase a new bond to extend the horizon by one more year as bond matures, keeping ladder length constant over time Allow ladder length to fluctuate based on market performance. For example, extend ladder by more when the stock market has performed well or market valuations are high. But let the ladder length decrease without extension after poor market performance or low market valuations Conduct a capital needs analysis for how much wealth should remain in each year of retirement to meet goals. Extend ladder when actual wealth exceeds the requirement. Let ladder length decrease when actual wealth is falling short. 19

One Simulation for an Automatic 10-Year Rolling Ladder To support an initial $40,000 Spending Goal with a 2% COLA Dynamic Asset Allocation in Retirement Twenty Random Simulations for an Automatic 10-Year Rolling Ladder Along with the 10 th, Median, and 90 th Percentiles for the Entire Distribution One Simulation for a Market-Based 10- Year Rolling Ladder To support an initial $40,000 Spending Goal with a 2% COLA Dynamic Asset Allocation in Retirement Twenty Random Simulations for a Market- Based 10-Year Rolling Ladder Along with the 10 th, Median, and 90 th Percentiles for the Entire Distribution 20

Critical Path for Sustainable Retirement Wealth To support an initial $40,000 with a 2% COLA for 45 Years One Simulation for a Personalized 10- Year Rolling Ladder To support an initial $40,000 Spending Goal with a 2% COLA Dynamic Asset Allocation in Retirement Twenty Random Simulations for an Automatic 10-Year Rolling Ladder Along with the 10 th, Median, and 90 th Percentiles for the Entire Distribution Probability of Success for Different Investment Strategies in Retirement To Support an Initial $40,000 Spending Goal with a 2% COLA 21

The Alternative: Essentials vs. Discretionary Building a Retirement Income Strategy Sources of Investment Spending Principal Interest, Dividends Model Portfolio Capital Gains (Risk Premium) 22

Sources of Annuity Payments $43,118 Income Sources for Income Annuity Principal Interest $31,060 Assumptions: 65-Year Old Female Bond Ladder Planning Age: 100 Mortality Credits (Risk Pooling) Survival-Weighted Present Value of Cash Flows 0.5% fixed real yield curve Society of Actuaries Individual Annuitant Mortality Table Probability of Success: Funding Real $43,118 for Different Time Horizons Systematic Withdrawals Product Allocation Source: Own calculations with 100,000 Monte Carlo Simulations for stock and bond portfolios. Bonds earn a fixed real return of 0.5%. Stocks earn an arithmetic average real return of 6.5% with a 20% annual volatility. Variable Annuities with Guaranteed Living Benefit Riders (GLWBs) Income Annuities (SPIAs & DIAs) 23

THANK YOU! ANY QUESTIONS? wade@retirementresearcher.com @WadePfau (Twitter) www.retirementresearcher.com 24