Monday, March 12, 2012

How Do Spending Needs Evolve During Retirement?


My new monthly column is now available at Advisor Perspectives. It is, "How Do Spending Needs Evolve During Retirement"

A lot of retirement withdrawal rate research uses the assumption that retirees wish to maintain constant inflation-adjusted withdrawals throughout their retirements. Spending does not either tend to increase or decrease as they age.  Ty Bernicke challenged that in a 2006 article, and one point I make in the article is that with Bernicke's spending assumptions, the worst-case historical sustainable withdrawal rate increased from 4.15% to 5.55%.  But I think he is going too far in the "retirees reduce their spending" direction as a general baseline assumption. My article explores more about the assumption and also talks about Somnath Basu's age banding research.


And for my blog readers, I have an exclusive deleted scene :) Actually, the article was getting to be on the long side, and I just had trouble finding a way to fit in the following section in a manner that flowed well with the rest of the article.  Here it is:

William Bengen and a Prosperous Retirement
For one other look at this issue, we can refer to William Bengen, whose follow-up studies on safe withdrawal rates do touch upon many important issues. In his article “Sustainable Withdrawals” from Harold Evensky’s and Deena Katz’s Retirement Income Redesigned, he investigates the inflation-adjusted spending assumption by considering an active phase for retirees between 65 and 74, then a transitional phase from 75 to 84, and a passive phase after age 85. In the active and passive phases, retirees maintain constant inflation-adjusted withdrawals. But in the transition phase between 75 and 84, retiree spending trails the CPI by 3 percentage points each year.  By spending less later in retirement, retirees should be expect to spend more early on, and William Bengen shows that with these spending reduction assumptions, the initial SAFEMAX (lowest sustainable withdrawal rate in history)  withdrawal rate increased from 4.15% to 4.59%. For a retiree planning to use the appropriate SAFEMAX, this implies a 9.6% reduction in the required retirement date nest-egg to fund retirement expenses.

11 comments:

  1. As a financial planner who works with hundreds of clients living in retirement, I agree with your conclusion that many, if not most, see no decrease in their income requirements - what they see is a change in the use of that income. As you point out, health-related expenditures increase significantly while discretionary spending diminishes. I believe that discretionary spending reductions are not entirely voluntary; if fact, as more people fear running out of money, these reductions are likely to increase in order to offset health costs and preserve one's nestegg. I think your research is on the right track - much more so that the others mentioned in your article.

    John F Evans, MBA, CPA, CFP, CRPC

    ReplyDelete
    Replies
    1. Thank you John. I'm always very glad to hear from practicing planners. Fear of running out of money could be a very powerful factor that limits spending in ways that would be hard to realize by just looking at the data. The data might show wealth holding steady, but that doesn't mean people are voluntarily cutting expenses. Thanks again.

      Delete
  2. Hi Wade

    I liked your article and agree with the concerns about whether or not the spending decline is fully voluntary. The impact of indexation is notable in Australia where the Age pension (social security) is now indexed to wages to preserve the standard of living. When this was changed there were a lot of discussion of how the CPI-indexing led to (involuntary) declines in the standard of living.

    ReplyDelete
    Replies
    1. Aaron, thanks, and thanks also for sending me that article before. I'm sorry I didn't respond. I still need to read it all.

      What you are saying is interesting. Generally wages increase faster than prices. So indexing benefits to wages would allow the standard of living of retirees to rise with the rest of the population rather than holding steady at the retirement date level. That would help make sure that retirees do not fall behind everyone else. But it must be extra expensive to index benefits to wages instead of prices, right?

      Basically, the US Social Security system indexes to wages until retirement and then to prices after retirement. That means older retirees will tend to have lower benefits than younger retirees.

      Delete
  3. Here's one possible way for retirees to mitigate the differential inflation rate for health care - buy more health care equities whose returns will rise along with the costs to a consumer. In fact, generalizing that idea, a retiree could build their portfolio to match his/her spending profile.

    ReplyDelete
    Replies
    1. That sounds like it will have lots of basis risk, i.e. the price of health care equities will not rise in precise coordination with the price of health care. Have you seen some research papers about this? Thank you for sharing.

      Delete
  4. Cannot say I've seen any research that supports the notion, no.

    ReplyDelete
  5. I'm kind of stuck on Fig 3 showing only a few thousand dollars a year in health care expenses. This must not count the cost of Assisted Living, which can easily be two or three times the entire gross pension/SS income of a 75+ year old.

    If an average cost for assisted living is presumed to be just $40k, and only 25% of elderly are in it, that still makes the average health care cost $10k.
    Bongleur

    ReplyDelete
  6. Bongleur,

    I'm under the impression that one of the limitations of the dataset is that it doesn't include individuals living in institutionalized settings, which would include nursing homes and the like. Actually, that could be a big problem resulting in underestimates about retiree spending, which is a reason I don't want to go too far in assuming less spending with age.

    ReplyDelete
  7. Wade,

    I am a practicing financial planner and agreed wholeheartedly with these insights into the differing sector inflation rates in retirement periods. Would you happen to know of current planning software that might allow a planner to incorporate such variables?

    Thanks.

    ReplyDelete
    Replies
    1. Kevin,

      Thank you for the feedback, and I'm sorry that I do not really know what software might be good for this. This might be a good question to post at the FPA Linkedin group, for instance.

      Best wishes, Wade

      Delete