- make assumptions about the future and crunch some numbers to develop a spending budget,
- compare your expected spending with your budget, and
- make necessary adjustments to your spending plans (or increase your income, if possible).
Since we started this blog in 2010 (and for many prior years when we were working as pension actuaries and weren’t blogging), we believed that the problem of how much to spend in retirement was a classic actuarial problem that could benefit from an actuarial solution. However, almost since the inception of this blog, we have heard from individuals (including some other actuaries) that the Actuarial Approach we advocate and make available in our Actuarial Budget Calculator (ABC) workbooks is either:
- too complicated for the average Joe because it employs complicated actuarial principles and present values, or
- it is not complicated enough because it doesn’t use stochastic assumptions to model future experience.
Before diving into our relatively brief defense of the Actuarial Approach, we remind our readers that we are retired actuaries and not financial advisors. As such, we do not give investment advice. Since many individuals (like the Washington Post columnist and many financial advisors) believe investing is inextricably linked with how much you can spend in retirement, the Actuarial Approach may perhaps start out with a strike (or two) against it. Based on basic financial economics principles, however, we do not believe this to be a flaw. In fact, we are skeptical of approaches (like some Monte Carlo models) that increase implied current spending levels (for a given probability of success) based on expectations of higher future investment returns associated with investment in more risky assets.
Background
Over the years, we have encountered several misperceptions about what we call the Actuarial Approach. The five most common ones (including the two addressed in this post) are discussed in our post of November 6, 2017. We are not going to repeat them or our responses to these misperceptions here.
The Actuarial Budget Benchmark (ABB) produced in our workbooks is a current year spending budget “data point” developed by comparing the market value of an individual’s (or couple’s) assets with the approximate market value of their future spending liabilities (i.e., the theoretical cost of purchasing currently available inflation-adjusted annuity contracts to cover future spending). Determining the market value of future spending liabilities and the market value of assets involves calculating present values of future payment streams. Our workbooks use deterministic assumptions consistent with inflation-adjusted annuity pricing (if default assumptions are used) to calculate these present values. The workbooks do not use stochastic assumptions. Under the Actuarial Approach, it is anticipated that a current year spending budget will be recalculated each year based on new data (and possibly revised assumptions).
In many ways, the development of an annual ABB is similar to the approach and process used to determine annual contribution requirements for defined benefit pension plans. In fact, the only significant difference is that we recommend using assumptions for future experience consistent with inflation-adjusted annuity purchase rates rather than expected returns on invested assets in order to increase the odds that spending budgets will increase in the future if and when expected returns do materialize and the ABB is recalculated.
The total ABB for the current year is separated into
- a spending budget for non-recurring expenses and
- a spending budget for recurring expenses.
As noted above, the same asset/liability comparison utilizing present values and deterministic assumptions about the future is employed by pension plan actuaries in their annual calculations of actuarially determined contribution requirements and by Social Security actuaries in their annual calculation of the system’s long-range actuarial balance. While stochastic assumptions may be utilized for other purposes, they are generally not utilized to determine annual pension plan contribution requirements or to annually measure Social Security’s long-range financial condition.
Investment advisors do use stochastic assumptions and Monte Carlo modeling to help pension plan sponsor clients develop investment strategies and to assess the risks that future experience will deviate from the deterministic assumptions used to develop the plan’s annual contribution requirements. Monte Carlo modeling can be an effective tool for these purposes (if assumptions about the future are reasonable). Generally, however, pension plan sponsors do not go through this Monte Carlo modeling exercise every year, and, as noted above, they generally do not use Monte Carlo modeling to determine annual pension contribution requirements.
Is the Actuarial Approach Not Complicated Enough?
Much like the services performed by investment advisors for pension plan sponsors, financial advisors have used stochastic assumptions and Monte Carlo modeling to help their clients develop investment strategies and to assess risks inherent in various investment and spending strategies. Many advisors, however, are convinced that stochastic assumptions and Monte Carlo modeling always trump deterministic “simple spreadsheet” models for all purposes. We disagree. Unlike the Actuarial Approach, these MC models do not generally develop a current year spending budget. They typically develop probabilities that total spending will exceed $X each year that clients can use to infer a total spending budget for the current year. The MC models also generally do not distinguish between non-recurring and recurring expenses and generally do not permit entry of different future increase assumptions for different types of expenses or of different levels of spending after the first death within a couple. The MC models also do not generally anticipate an annual valuation process to make sure that spending is automatically adjusted to remain on track to accomplish client objectives.
While we acknowledge that the use of stochastic assumptions and MC models can be helpful in personal financial planning when it comes to developing an investment strategy or in assessing the risks associated with various investment/spending strategies, we remain convinced that, analogous to the experience with pension plans and Social Security, the Actuarial Approach (and its annual valuation adjustment process) is better than MC modeling at developing a robust spending budget, despite the fact that (or because) it may employ deterministic assumptions. It should also be noted that stochastic assumptions could be used with the Actuarial Approach, but then we would not be able to make our simple workbooks available to individuals (especially DIYers) who simply want to develop a reasonable spending budget.
Is the Actuarial Approach Too Complicated?
On the other end of the spectrum, we have heard from retirement experts that most individuals aren’t smart enough to understand the basic actuarial concepts and present values used in the Actuarial Approach. Therefore, financial advisors and retirement experts continue to advocate “simpler” systematic withdrawal strategies or approaches (SWPs) like the 4% Rule or the IRS RMD approach. We prefer to believe that with a little bit of education and effort, most individuals can develop a better spending budget using our relatively straight-forward ABC workbooks and default assumptions. These workbooks enable users (including those who may not fully understand present values) to perform the present value calculations required under the Actuarial Approach.
Conclusion
We believe the same basic actuarial principles used for pension plan funding and measurement of Social Security’s financial condition can be used to develop a reasonable annual spending budget for individuals and couples. Instead of being “too complicated” or “not complicated enough,” we believe the Actuarial Approach may be “just right” for you, and we encourage you to give it a try.
