Background
The assumptions used in the Actuarial Approach Excel workbooks to determine the current year spending budget are “deterministic” in that they are defined for each future year. If experience exactly follows these deterministic assumptions, future recurring spending budgets are expected to increase from year to year according to the input percentage future budget increase assumption (expected pattern). If future experience is less favorable than assumed, calculated future spending budgets will be lower than the expected pattern, and if future experience is more favorable than assumed, calculated future spending budget will be higher than the expected pattern (assuming no smoothing of the results). A deterministic model will produce just one pattern of assumed future experience, and it is quite likely that that one pattern will be wrong.
Consistent with basic financial economics theory, the deterministic assumptions we recommend as default assumptions for future investment returns, inflation and lifetime planning periods are intended to be consistent with low-risk assumptions used to price inflation-adjusted annuities. We use this market information on relatively low-risk investments to price spending liabilities and assets. We discuss the rationale for selecting these assumptions, and their importance, below.
Instead of one pattern of future experience developed in a deterministic model, Monte Carlo models will generate many patterns (or simulations) of future experience. For example, a Monte Carlo model can generate thousands of future investment return patterns based on random returns developed from normal distributions of expected returns and expected variations selected for different asset classes. Monte Carlo models can capture the random nature of future experience better than models that use deterministic assumptions. This gives the impression that such models are more realistic than deterministic models. Some might argue, however, that instead of one wrong projected pattern of future experience, a Monte Carlo model will generate thousands of future patterns that will turn out to be wrong. The many simulations will, however, enable a financial advisor using a Monte Carlo model to illustrate ranges of future experience and generate probabilities that certain spending or investment strategies can achieve specified objectives.
Purpose of Calculations
As discussed above, deterministic models tend to provide specific outcomes, and Monte Carlo models tend to provide outcome ranges and probabilities of achieving specified objectives. Deterministic approaches also enable a user to consider more complicated calculations within the model. This is one of the reasons why pension plan actuaries generally use deterministic models to determine specific contribution requirements for pension plan sponsors, and Social Security actuaries use deterministic models to measure Social Security’s financial status and to price proposed changes.
The Actuarial Approach develops a total spending budget for the year, comprised of a non-recurring expense budget and a recurring expense budget. For example, it might indicate that the non-recurring expense budget for this year is $X, the recurring expense budget for this year is $Y and the total expenses budget is $Z. The budget developed under the Actuarial Approach can also be fine-tuned by assuming different rates of increase for different types of future expenses. These are specific outcomes as opposed to a Monte Carlo model that may indicate there is a 90% probability that annual real dollar spending will exceed $A per year until the last death within the couple and an 80% probability that annual real dollar spending will exceed $B (higher than $A).
We believe that the specific spending budget outcomes produced by the Actuarial Approach are generally more useful and understandable than probabilities of achieving certain annual spending goals under Monte Carlo models.
Importance of Model Assumptions
As noted above, the default deterministic assumptions under the Actuarial Approach are intended to be consistent with assumptions used to price currently available inflation-adjusted life annuities. The reasons for selecting these assumptions as a default option include:
- These relatively low-risk assumptions are consistent with general financial economics theory and are used to determine the market value of spending liabilities
- Most retirees are more concerned with over-spending risks than under-spending risks and would rather err on the conservative side with greater potential for future increasing spending budgets rather than future decreasing spending budgets
- While investment in higher risk securities may be expected to yield higher returns, risk-adjusted returns should be unchanged
Importance of Approach Processes
There are several processes which are part of the Actuarial Approach. Perhaps the most important process is the annual valuation process. This dynamic process keeps spending budgets on track by annually remeasuring the market value of assets and spending liabilities and making appropriate adjustments to the spending budget when necessary to reflect emergence of actual experience.
Because the 10,000 simulations under a Monte Carlo model imply a 90% probability of being able to spend at least $A each year (for example) for life, there is an implication that a static $A is the “set and forget” plan and annual or periodic valuations may not be necessary.
Another process under the Actuarial Approach involves periodic stress testing of assumptions to assess risks. This will generally involve more worse-case “what-if” analysis than anticipated under a Monte Carlo model.
Ease of Understanding and Calculation
Not everyone understands probabilities. A probability in excess of 85% is frequently interpreted to be 100%, while a probability less than 15% is frequently interpreted to be zero. In addition, including stochastic assumptions in a model significantly complicates the required calculations. DIYers are much more likely to be able to understand and do their own calculations using the Actuarial Approach than their own Monte Carlo model calculations.
Social Security
In this section, we note the similarities between the Actuarial Approach and actuarial calculations performed by Social Security actuaries in their annual measurement of the program’s financial condition.
Social Security actuaries use both a deterministic model and a stochastic model to evaluate Social Security’s financial condition. Most of their communications focus on the results of their annual deterministic model. For example, in the 2019 Trustees Report, under best estimate assumptions, Social Security actuaries predict program trust fund assets will be depleted in 2035, a widely quoted number. By comparison, they briefly note on page 18 of the Trustees Report that under their stochastic model, “trust fund asset reserves will become depleted between 2031 and 2044 with a 95-percent confidence.”
In a recently prepared report, the Office of the Actuary evaluated the impact on Social Security’s long-range actuarial balance and annual balance in the 75th year of their deterministic forecast of various possible changes to the program. The estimates contained in this report are all based on their deterministic model. Readers may find the discussion in Section G—Provisions Affecting Trust Fund Investment in Marketable Securities to be of interest relative to selection of assumptions. The Social Security actuaries determined that investment of the Social Security Trust Fund would have no immediate impact on the program’s long-range actuarial balance. They indicated:
“Thus, these selections have no effect on the actuarial balance of the OASDI program. Many analysts believe the higher expected return for equities should not be included in valuations because the tendency for higher average returns is compensation for the higher volatility in equities. The low or risk-adjusted yield assumption reflects this perspective.”
Thus, in addition to preferring to use their deterministic model rather than their stochastic model to measure the adequacy of Social Security and communicate results, the Social Security actuaries support the same concept of risk-adjusted returns that we advocate for personal spending budget determinations.
Conclusion
Our mission at How Much Can I Afford to Spend in Retirement is to help people make better financial decisions relative to their retirement. We are two retired actuaries. We are not:
- financial advisors,
- investment advisors, or
- tax experts
