Monday, July 4, 2016

Use a Rainy Day Fund to Manage Investment Risk in Retirement

Inspiration for this post comes from an article by Joe Tomlinson in Advisor Perspectives entitled, “Retirement Planning and the Impact of Investment Market Performance” and a survey entitled, “The Society of Actuaries 2015 Risks and Process of Retirement Survey.”  This post is also a follow-up to my previous post regarding using The Actuarial Approach to mitigate sequence of return risk and my post of June 27, 2015 entitled, “You Can Save During Retirement Too.”  The primary focus of today’s post is to discuss the use of a Rainy Day Fund to reduce or mitigate year to year variations in spending budgets.
In his article, fellow actuary Joe Tomlinson uses Monte Carlo modeling to “stress” several different “withdrawal” strategies including the 4% Rule and “actuarial approaches” like the one I recommend in this blog.  Joe’s models incorporate an average Equity Risk Premium assumption of 3% and T Bond returns of 0.5% to determine which withdrawal strategies are the most resilient in a “stressed” economic environment.  Of course frequent readers of my site are well aware that I advocate development of reasonable spending budgets and specifically advise against using “withdrawal strategies” to “tap one’s savings.”   Notwithstanding its emphasis on withdrawal strategies, Joe’s article is worth reading, and he reaches a number of interesting conclusions for budgeting during “stressed” economic conditions, including:

  • “Although the 4% rule is widely used in retirement research, it is not well suited to real-world retirement planning.” 
  • “actuarial approaches” don’t reduce the average failure rate [based on Joe’s unique definition of failure discussed more below], but do reduce the average shortfall associated with failure.  In addition, actuarial approaches increase average “consumption”, reduce average bequests but increase the volatility of consumption from year to year. 
  • Smoothing the actuarial approach can reduce volatility of consumption from year to year, but it does not reduce the average failure rate.  On the other hand, using a relatively large portion of ones accumulated savings to purchase an annuity can significantly reduce both volatility of consumption from year to year and also avoid Joe’s definition of failure.
Like most retirement researchers using Monte Carlo modeling, Joe assumes that retirees will spend (or consume) exactly their spending budget every year.  I call assumptions like this one “Monte Carlo reality” as opposed to “reality.”  As a practical matter, retirees frequently spend what they want or need to spend during a year without regard to their spending budget (if in fact they have one).  This brings us to the second item of inspiration for this post:  the most recent Society of Actuaries survey.  Unlike the theoretical Monte Carlo world, this survey attempts to capture what retirees actually do. 

There is a lot of good material in this survey, but I am going to focus on the strategies that retirees are actually taking to manage risks in retirement as background for the discussion that follows.  The survey results for this item are summarized in Figure 53 and for those already retired, the strategies with the highest percentage “already done or plan to do” include:

  • Eliminate all your consumer debt (86%) 
  • Cut back on spending (76%) 
  • Try to save as much as you can (74%)
The strategies with the lowest percentages for retirees include:
  • Work in retirement (30%) 
  • Buy a product or choose an employer plan option that will guarantee income for life (22%) 
  • Postpone taking Social Security (20%) 
  • Postpone retirement (12%)
Figure 59 shows that 85% of surveyed retirees would reduce expenditures significantly if they were running out of money due to unforeseen circumstances.

Figure 159 tells us that only 22% of surveyed retirees had a plan for spending down financial assets while the rest had no plan or planned to grow assets or maintain asset values in retirement.

Thus, the survey tells us that rather than buy annuities, most retirees are managing their risk in retirement by reducing their spending during poor economic times and saving some of their assets during good economic times.

Unlike failure under the 4% Rule (which requires actually running out of money), Joe defines failure under the actuarial approaches as any future year during which his sample couple’s annual spending budget drops below $70,000.  This strikes me as an arbitrary measure of failure as the couple could simply spend more than their budget in such a year or they could transfer assets from funds earmarked for discretionary spending purposes.  It is somewhat hard for me to believe that a couple with $1.5 million in assets and $40,000 in annual Social Security income is going to perceive a temporary spending budget less than $70,000 as being a failure.  Rather than looking at this as failure, it seems to me that most retirees view a temporary reduction in their spending budget as simply the price to pay for investing in risky assets rather than buying an annuity.

Joe is right, however, that, all things being equal, an actuarial approach can produce a budget that is more volatile from year to year than the 4% Rule or other safe withdrawal rate approaches.  Joe is also right that one way to dampen the budget volatility of an actuarial approach is to invest some of the assets that were invested in risky assets into less risky assets, such as annuities.  I happen to like annuities and have written favorably in this blog about the potential advantages of combining annuities with investments.  However, given their dismal endorsement in the SoA survey, it is clear that they are not everyone’s cup of tea.  Fortunately, there are other ways to either smooth budgets or to smooth spending.  

It is important to point out once again that budgets are not equal to actual spending, so when Joe says that “the tradeoff [of going to an actuarial approach] is that the year to year volatility of consumption increases significantly, he is not being 100% accurate.  In Joe’s Monte Carlo reality where budgets equal spending equals consumption, this may be true, but in reality, either budgets or actual spending can be smoothed when investment returns are volatile.  One approach, as mentioned by Joe in his article and by me in my previous post, is simply to smooth the budgets produced by the actuarial approach.  Another approach, which I discussed in my post of June 27, 2015, is to set-up and use a “Rainy Day Fund” (RDF).  Under this approach, investment gains (investment returns in excess of those expected based on the investment return assumption) are transferred to the RDF while investment losses (to the extent they are covered by the RDF) are transferred back to the fund used to determine the budget when needed.  The RDF assets are ignored for budget determination purposes and are intended to be used during periods of poor investment performance, or to be used for other expenses when and if the RDF becomes too large.

To illustrate how the RDF can work, let’s use the same example we used in the previous post (a 65-year old retiree with $500,000 of accumulated savings, $20,000 annual Social Security benefits and a first year spending budget of $41,751), but with the following sequence of actual investment returns over the next five years:  10%, 0%, 20%, -5%, 4.5%.  At the end of the first year, expected assets are $499,770, but actual assets are $526,074 (assuming the retiree spends exactly the budget amount for the year).  Therefore, the retiree will transfer the investment gain of $26,304 from the budget fund to the RDF.  Transferring this amount will keep the second year spending budget constant in real dollars at $41,751.   Because of the less than expected investment return in the second year, the retiree will transfer assets back from the RDF to the budget fund to maintain the constant $41,751 real dollar budget and the RDF will drop to $4,817 as a result.  The same process is used for subsequent years.  The graph below compares spending budgets prepared by the Actuarial Approach with no smoothing with the Actuarial Approach utilizing the RDF approach.  Amounts are shown in inflation adjusted dollars.  At the end of the fifth year, the RDF would contain $31,934.  Note that the gains and losses transferred back and forth from the RDF could also include gains and losses other than investment gains and losses and the annual gain/loss would be easily determined as the difference, if any, between actual end of year assets (including the RDF) and expected end of year assets shown in the Actuarial Budget Calculator run out tab.  If the RDF falls to zero, then budgets would have to be reduced as discussed in the previous post.

click to enlarge

This graph shows that for certain investment sequences volatility in year to year spending budgets can be avoided through the use of the Actuarial Approach with an RDF.  This can be important for those retirees who are concerned about such volatility but who may not want to spend significant amounts of their savings on annuity products to manage this risk.   Alternatively, you can live with the volatility, spend what you want and simply use the Actuarial Approach with no smoothing to periodically justify adjustments in your spending.