In traditional retirement, planning the annual investment return is assumed to be a constant number, such as 8.5 percent per annum. This number depends on the asset allocation and on assumptions about the returns of the different asset classes. The outcome of the computation is typically presented as the expected wealth values over the anticipated period of retirement.
The problem with this approach is that investment returns are neither known nor constant over time. While investing is about risk, retirement calculators presenting single scenarios treat risk either as a certainty, or at best, as a 50/50 proposition: 50/50 you will do better or worse than the expected outcome. Investing is not an exact science; no one knows the precise return of different asset classes over any given number of years. Investment returns are random variables, characterized by expected values (or averages), standard deviations, and, more generally, probability distributions. For this reason, projections of an investment program's possible results should also be expressed in terms of probabilities. For example:
There is a 95 percent chance you won't run out of money in retirement.
There is a 50 percent probability you will accumulate at least $3.1 million. There is a 25 percent chance you will have $5.2 million or more. But there is also a 10 percent chance you will have $400,000 or less.
To arrive at these types of conclusions it is necessary to use what are known as Monte Carlo (MC) simulations.
MC simulations require a set of assumptions regarding time horizon, initial investment, asset allocation, withdrawals, savings, retirement income, rate of inflation, and correlation among the different asset classes.
In MC simulation programs, the growth of an investment portfolio is determined by two important inputs: portfolio average expected return and portfolio volatility, represented by the standard deviation measure. Based on these two inputs, the MC simulation program generates a sequence of random returns from which one return is applied in each year of the simulation. This process is repeated thousands of times to calculate the likelihood of possible outcomes and their potential distributions.
MC simulations also provide another important benefit: They allow investors to view the outcomes of various strategies and how marginal changes in asset allocations, savings rates and withdrawal rates change the odds of these outcomes.
For example, after examining the output, an investor might decide she is taking more risk than needed to achieve her goals. It's a relatively easy fix: lowering the equity allocation and/or lowering the exposure to risky asset classes. If she is not taking enough risk to provide acceptable odds of success, the decisions get more difficult. She must (1) take more risk than she would otherwise like to take, (2) lower the goal, (3) save more (lowering her current lifestyle), or (4) or accept and live with the estimated risk of failure.
The benefits of using an MC simulation are seen in this example. Assume the input results in an 80 percent chance of success (not running out of money while still alive). The simulation shows the impact on those odds of increasing (or decreasing) savings by $X a month. If that change increased the odds of success to 85 percent, the investor might decide that it is worthwhile to reduce current consumption to improve the odds of success by that amount. If it only raised the odds of success to 81 percent, the investor might draw a different conclusion.
The output can also be analyzed to see how changes in the asset allocation impact the odds of success. If increasing the equity allocation from 70 to 80 percent increased the odds of success from 80 to 90 percent, an investor might decide it was worth the extra risk of more equity ownership (and the extra stomach acid it was likely to produce along the way). If it only increased the odds of success to 81 percent, there might be a different decision. Changes in withdrawal rates that impact future lifestyle can be analyzed. For example, if a 4 percent withdrawal rate produced a 95 percent chance of success, and a 5 percent withdrawal rate lowered the odds of success to 90 percent, an investor might choose to raise the withdrawal rate, accepting a somewhat lower likelihood of success in return for greater consumption. If that decision is made and the risks do materialize, the investor must be prepared to accept an even lower lifestyle in the future.
The simulation program can be used to look at how delaying retirement by X years impacts various issues, such as the need to save, the withdrawal rate, or the required equity allocation. The same analysis can be done for earlier retirement. Investors can determine if an extra year of working is worth the greater lifestyle now and/or in the future, or how that extra year of work impacts the need to take risk. For example, each extra year of work might allow for a reduced need to save $X per year, a Y percent increase in the withdrawal rate, or a reduction in the equity allocation of Z percent.
The right answers are unique to the individual and their ability, willingness, and need to take risk. For some people, an 85 percent chance of running out of money will be perfectly acceptable. For others, anything short of 99 percent might be unacceptable. The decisions are all preferences driven by personal choice. Here are some guidelines:
The more risk averse the investor, or the lower the marginal utility of wealth, the more emphasis should be placed on the odds of not running out of money rather than those of creating a large portfolio.
Lower odds of success can be tolerated the more options an investor is both able and willing to exercise if the risks do "show up."
In the following examples applying the above principles, we will consider two investors, identical with one exception.
Application: The exception: one investor owns a second home. This investor has the ability to sell that home should the "left tail" (worst-case result) of the potential distribution of returns show up, so he can accept lower odds of not running out of money than the other investor. However, if the investor is unwilling to exercise that option—the grandchildren live in that area and a high value is placed on owning the property—the option doesn't matter. Only the options that investors are both able and willing to exercise should be considered.
Application: The exception: one investor is both willing and able to lower his need for cash flow from the portfolio should the "left tail" of the potential distribution of returns show up. That investor can accept lower odds of not running out of money more than one who is either unable or unwilling to lower his spending requirements.
Application: The exception: one investor is both willing and able to extend her planned time in the work force. The other places a higher value on her ability to retire. The investor who is both willing and able to continue to work longer can accept more risk, but she should consider the need for disability insurance should health prevent her from exercising the option to work longer.
Application: The exception: one investor has a long-term care policy. This investor can accept more investment risks than an investor without a policy.
If you do not have access to MC simulators, it is still possible to estimate an appropriate safe rate of withdrawal. Financial planners and investment advisers often use "the 4 percent rule." The rule comes from the study "Retirement 'Spending': Choosing a Sustainable Withdrawal Rate."[30] The authors found that a portfolio of 50 percent stocks and 50 percent bonds had a 95 percent historical success rate for a thirty-year horizon when using a 4 percent withdrawal rate, increasing that rate with inflation. Because the three finance professors who authored the study were from Trinity University, it is often referred to as the "Trinity Study."
The success or failure of your retirement spending plan mainly depends on three things: time horizon, spending amount, and the returns of your investment portfolio. The good news is that you have some control over all three. The bad news is that you are not in complete control over any of them.
The time horizon of your retirement is a function of your retirement age and life expectancy. Most people have at least some control over their retirement age. As for spending, you may have some discretionary spending. While you cannot control investment returns, you can decide on an asset allocation giving you the expected return required to support your needs.
Table 16.1 addresses time horizon by looking at mortality statistics. The percentages shown are the probability an individual at age sixty-five will survive to a particular age. The table is from a great resource on retirement planning issues, Moshe Milevsky's Are You a Stock or a Bond?
Table 16.1. Mortality Table at Age Sixty-Five
To Age: | Female (%) | Male (%) |
|---|---|---|
Source RP2000 Mortality Table, IFID Centre Calculations. | ||
70 | 94 | 92 |
75 | 85 | 81 |
80 | 72 | 66 |
85 | 56 | 46 |
90 | 35 | 24 |
95 | 16 | 8 |
Table 16.1 can be a useful resource when planning the length of your retirement. An individual's personal health situation must also be taken into account when estimating the time horizon. Because the cost of being alive without sufficient assets to support an acceptable lifestyle is so high, we prefer to err on the side of a longer than necessary time horizon. For most retirees, leaving a larger estate than planned is preferable to running out of money in the last years of life and being forced to rely on relatives or selling a home.
Table 16.2 is a guideline for establishing a safe withdrawal rate. The results are based on our MC simulation work. The table summarizes the results at different ages for an individual whose equity allocation ranges from 30 to 50 percent. For those in the early phase of retirement we typically plan through age ninety.
Based on this table, a fifty-five-year-old should not take more than an inflation-adjusted 3 percent per annum from his portfolio. Thus, an individual with a portfolio of $1 million should not withdraw more than 30,000 the first year. He can adjust that amount each year for inflation. The table demonstrates that as time passes—and the horizon shortens—the safe withdrawal-rate percentage increases.
When using Table 16.2 to make withdrawal rate decisions, keep in mind it is not a "one size fits all" solution. If your health is such that you are likely to live well beyond age ninety or die much sooner than ninety, you should adjust your withdrawal rate accordingly. Those with other options to exercise (reduce discretionary spending, access home equity) can be more aggressive when choosing a withdrawal rate. Those without such options should consider a slightly more conservative withdrawal rate than the table indicates. The withdrawal rates shown are no guarantee a portfolio will not be depleted: They only indicate a high likelihood it will not occur.
Having decided on the appropriate withdrawal rate, you should choose the most tax-efficient location from which to make the withdrawals.
As discussed in Chapter 10, proper location of assets is an important part of the winning investment strategy. Another is using the most tax efficient withdrawal sequence to fund retirement. Should a retiree first withdraw funds from the taxable account, then the traditional IRA [401(k), 403(b), and other tax-deferred accounts], and finally the Roth IRA? Would another sequence be preferable? Two studies, Stephen M. Horan's "Withdrawal Location with Progressive Tax Rates"[31] and William Reichenstein's "Tax-Efficient Sequencing Of Accounts to Tap in Retirement"[32] provide answers.
The solutions are based on two key principles. The first: "Returns on funds held in Roth IRAs and traditional IRAs grow effectively tax exempt, while funds held in taxable accounts are usually taxed at positive effective tax rates." The second principle, discussed in Chapter 14, is that only "part of a traditional IRA's principal belongs to the investor. The IRS 'owns' the remaining portion. The objective is to minimize the government's share of the principal."[33]
Since the winning strategy is withdrawing funds from the account with the higher tax rate, the general rule of thumb is to first withdraw from taxable accounts. In addition, since we want to fund expenses with the most tax-inefficient assets, the sequencing of withdrawals from taxable accounts should first be from fixed income (bond) holdings. There are some exceptions to these rules:
The main source of income is derived from tax-sheltered accounts. Withdrawals should be made from these accounts until taxable income at least reaches the lowest tax bracket. This exception should apply to any year when taxable income is low. For example, a retiree will likely be in a low tax bracket in years when she has large medical expenses, perhaps due to a stay in a nursing home.[34] For retirees with large balances (several million) in their IRAs, it may be appropriate to withdraw from those accounts until income reaches the 28 percent bracket.[35]
The beneficiaries of the IRA will be in a higher tax bracket than the owner.
Retirees who have substantial unrealized gains on taxable assets can await the step-up in basis at death. Such retirees should withdraw funds from retirement accounts before liquidating the appreciated asset. An example would be a terminally ill person, because the effective tax rate on the capital gains will be zero if they await the step-up in basis at death.
By delaying withdrawals from an IRA, the greater required minimum distributions (RMD) will cause the tax rate to rise to higher levels. In addition, if a retiree is in a low tax bracket but doesn't need to withdraw funds from the IRA to meet spending needs, she should consider a conversion to a Roth IRA. Thus, the conventional wisdom to delay withdrawals from traditional IRAs for as long as possible is not always correct.
Many individuals face the decision of whether to withdraw from a traditional IRA or a Roth IRA. According to Reichenstein:
Withdrawing funds from the traditional IRA makes sense in years when the retiree is in a low tax bracket and if the retiree's beneficiary will be in a higher tax bracket. Withdrawing funds from a Roth IRA instead of a traditional IRA makes sense in years when the retiree is in a high tax bracket and if the retiree's beneficiary will be in a lower tax bracket. In addition, Roth IRA withdrawals may also be preferred if the retiree expects to have large deductible medical expenses later in retirement—if the deductions lead to low tax rates. Withdrawals from a Roth may also be preferred if the retiree wishes to leave funds to a charity.[36]
Reichenstein found that sequencing strategies are sensitive to the following:
The higher the tax rate, the greater the advantage in first withdrawing from a taxable account. If capital gains taxes were increased the advantage would increase, as it would if an investor held an actively managed mutual fund that, due to turnover, was tax inefficient.
The relative advantage of the "taxable account first" strategy is greatest when the portfolio is roughly evenly divided between retirement accounts and taxable accounts.
The relative advantage of the taxable first strategy increases with the asset's rate of return.[37]