Monday, February 18, 2013

Guest Post: Ken Steiner on how Actuaries Think About Retirement Income

Developing A Retirement Spending Strategy--An Actuarial Approach


Ken Steiner, Fellow, Society of Actuaries, Retired

Ken Steiner worked as a pension actuary at Watson Wyatt. He and Steve Vernon (who now writes at CBS MoneyWatch and is quite popular with a number of my blog readers) were colleagues there. Steve wrote very positively about Ken's website, How Much Can I Afford to Spend in Retirement, in Steve's excellent book, Money for Life. As you will see, actuaries are no fans of the 4% rule either, and it is a pleasure to allow Ken to introduce his approach, which is detailed more completely in the longer articles at the top of his webpage. Though the dynamic approach he describes makes it more difficult to budget future spending amounts, the reality is that those seeking to fund their retirements from a diversified portfolio of volatile assets must also expect to change their spending as they see how their portfolio does. In terms of research I've reviewed in the past, this approach has the most in common with the "mortality-adjusted constant probability of failure" approach.

Retirees have many different objectives in retirement.  Certainly not all of these objectives are financial in nature.  However, retirees generally have two potentially conflicting financial goals: (i) spend enough each year to maintain a certain standard of living throughout retirement and (ii) not spend so much that accumulated savings run out prior to death.  One possible solution to this puzzle is to invest all (or substantially all) accumulated savings not intended to be bequeathed in lifetime income insurance products (immediate annuities, deferred annuities and other similar life insurance products).   Frequently, annual spending may be increased using this approach as a result of the survivorship premium built into annuity pricing.  

100% investment in lifetime income products may be a good solution for some retirees but they can also limit a retiree's ability to spend accumulated savings on unforeseen expenses, such as purchase of a new car or significantly large medical expenses, and these products may not provide sufficient protection against inflation.  For these reasons (and many others), many retirees do not purchase lifetime income products or alternatively, invest only part of their accumulated savings in such products.  

My background is in the pension actuarial field.  I'm a retired pension actuary, not an investment expert or a qualified financial planner by any means.  I will leave the subject of investments (including investments in lifetime income insurance contracts) to other more qualified individuals, such as the author of this blog.  For the purpose of this article, however, I will focus on how to best solve the financial problem presented above by the two conflicting financial goals assuming that the retiree does not fully annuitize her accumulated savings.  I will also assume that maintenance of a person's standard of living is ideally achieved by developing a spending strategy that results in level real dollar spendable amounts from year to year from all sources (Social Security, immediate annuities, deferred annuities and withdrawals from accumulated savings, part-time work, etc.).

For many years, retirees and some financial planners have utilized the "4% Rule" to determine annual amounts that may be withdrawn from accumulated savings.  More recently, "safe" withdrawal rates other than the initial 4% rate have been developed using Monte Carlo techniques and different assumptions for future expected experience.  These safe withdrawal rate (SWR) approaches generally involve applying the developed percentage to accumulated savings in the first year of retirement and increasing the amount withdrawn in subsequent years by the increase in measured inflation over the previous year, irrespective of actual investment performance during the previous year.  Proponents of SWRs argue that they are simple to apply and subject to relatively low risk of failure (i.e., running out of money).  

I'm not a big fan of SWRs.  First of all, they aren't really all that simple.  Secondly, as proven by the author of this website, they aren't necessarily that safe.  Closer inspection of the assumptions used to develop SWRs shows that adjustments to the safe withdrawal percentage are anticipated to be made for different assumed longevity and different investment mixes.  Other experts indicate that the SWRs should be periodically adjusted for actual experience, but adjustment details are never very clear.  In addition to not reflecting actual experience as it emerges, SWRs make no adjustment for the possible existence in a retiree's investment mix of fixed dollar immediate or deferred annuities (footnote 1), do not reflect possible bequest motives and make no adjustment for deviations from the spending strategy.

Footnote 1:  A reasonable withdrawal strategy should be coordinated with other fixed income sources of retirement income.  All other things being equal, the larger the amount of immediate fixed income annuity in a retiree's investment portfolio, the smaller the initial withdrawal from accumulated savings, as relatively larger amounts of accumulated savings will be necessary in later years to effectively provide for indexing of the immediate fixed annuity.  In addition, the availability of large amounts of deferred annuity income will generally permit larger initial withdrawals from accumulated savings.  

As a pension actuary, I spent years determining contribution requirements and accounting costs for defined benefit pension plans sponsored by my clients.  About ten years ago, I realized that the same process that we actuaries used for these purposes could be applied by individuals who "self-insure" their own retirement.  After I retired, I put together these ideas in a website, How Much Can I Afford to Spend in Retirement?

Visit my website for more detailed discussion of the actuarial process I recommend and how to use the calculation spreadsheets that reside there.  Also included are links to articles (many from the author of this blog) and other sites relevant to withdrawal strategies.  An abbreviated description of the actuarial process and calculation spreadsheet tool contained in the website follows.

The most important part of the withdrawal strategy presented in my website is not the somewhat unsophisticated Excel spreadsheet (Excluding Social Security 2.0) that can be used to determine a spendable amount payable from accumulated savings, but the relatively simple actuarial process to be followed each year to adjust for experience different than assumed, deviations from the spending budget and changes in assumptions.  It is this annual (or more frequent) adjustment process that is key to keeping a retiree's withdrawal strategy on track with her financial objectives.  

The spreadsheet requires input of accumulated savings, any immediate fixed life annuity income, any fixed deferred annuity income (and the period of deferral), the expected rate of return on accumulated savings, the retiree's expected remaining lifetime, desired amounts at death to be left to heirs, the expected annual inflation rate and the desired rate of increase in payments each year (which may or may not be the same as expected inflation).  Social Security and other inflation-indexed retirement income is excluded from the calculation (hence the name of the spreadsheet) and is added in by the retiree at the end of the process to determine a gross spending budget for the year.  Once these items are input in the spreadsheet, the program calculates the spendable amount payable for the year from all sources (excluding Social Security and other inflation-adjusted income) and the spendable amount payable from accumulated savings.  Two run-out tabs show the decumulation of  accumulated savings based on the input items.  One of the run-out tabs shows amounts in nominal dollars and the other tab shows real inflation-adjusted dollars.  If the input desired increase in annual payments is equal to input expected inflation, the total expected spendable amounts will be the same for all future years in the inflation adjusted run-out. 

The run-outs are based on exact realization of input deterministic assumptions and exact amounts withdrawn each year.  The one thing that we do know for sure about the future is that actual experience will not exactly follow assumed experience.  As Yogi Berra said, "It's tough to make predictions, especially about the future."  Therefore, it is critical to periodically revisit this process.
At the beginning of each year, I pull out my retirement budgeting file, which includes print-outs of the input page and run-out page from the calculation spreadsheet from my website that I ran the previous year, and pull together current data to determine how much accumulated savings I have left.  I re-run the spreadsheet based on current data and revised assumptions, if necessary, and decide whether and how to smooth the expected spendable amount based on last year's results to reflect current data and assumptions.   Bam--I have this year's spending budget.  I then print out the results for the current year and put it in my files to revisit next year.    The entire process takes maybe ten minutes once I have gathered all the financial data.  

Is the process perfect?  No.  Input of unrealistic assumptions into the spreadsheet will produce unrealistic results.  As suggested in my original paper, your assumptions are probably overly optimistic if running the calculation spreadsheet with no annuity amounts, no amounts to be left to heirs and no desired increases in withdrawals produces a higher annual spendable amount than the annual amount you could receive by using your accumulated savings to purchase a fixed immediate life annuity.  In addition, if you want to make sure that you don't run out of money before you die, you are going to have to be somewhat flexible with respect to your goal of maintaining constant real dollar retirement income from year to year.  Finally, no simple calculation spreadsheet is going to anticipate everyone's specific situation.  If you find a better calculation spreadsheet elsewhere that better meets your specific needs, you should use it.

I thank Dr. Pfau for giving me the opportunity to briefly present my thoughts on withdrawal strategies.  I close this guest blog with several quotes from a recent article in the January/February 2013 issue of Money magazine regarding developing a reasonable withdrawal strategy in retirement.

"Best move: "Recalculate your withdrawals every year to take into account your current account balances and the fact that your nest egg doesn't have to support you for as long." 

"With a decision this big, you don't want to blindly stick to the 4% rule or any other rigid system..."

"As a practical matter, though, recalculating your withdrawal rate this way can be quite complicated. So unless you're working with a financial planner capable of doing the number crunching for you, your best bet is to go to an online tool like T. Rowe Price's Retirement Income Calculator every year, plug in your most up-to-date information, and adjust your withdrawals up or down as necessary." 

I couldn't agree more with this advice from Money magazine. And the online tool "like" T. Rowe Price's that I recommend can be found on my website.