Wednesday, January 25, 2012

Safe Retirement Income with TIPS and a deferred annuity

Coincidentally, a message I've been seeing repeatedly from many sources during the past week is the reminder that a strategy of using a bond ladder to provide one's retirement income is still exposed to longevity risk. That is, using TIPS to provide retirement income needs over 30 years may have a very low failure rate, for instance, but with principal exhausted after 30 years the failure rate jumps up to 100% should you happen to live into a 31st year past retirement.

In the new January/February 2012 issue of the Financial Analysts Journal (the research journal for the CFA Institute), authors Stephen C. Sexauer, Michael W. Peskin, and Danield Cassidy propose a solution of building a TIPS ladder for 20 years and buying a deferred annuity which kicks in after 20 years to provide income for the rest of one's life.  Their article is called, "Making Retirement Income Last a Lifetime."

Before getting to the substance of the article, just a brief note that this article would have surely not been published had someone along the line (authors, referees, editor) done a bit of homework. First, it is rather odd that this article has the same title as one of the most famous studies on sustainable retirement spending, the December 2001 article by John Ameriks Robert Veres and Mark Warshawsky in the Journal of Financial Planning. Second, the message and contents of this new article are nearly identical to S. Gowri Shankar's article, "A new strategy to guarantee retirement income using TIPS and longevity insurance" from a 2009 issue of Financial Services Review. In a number of ways, Professor Shankar's article is actually superior for a do-it-yourself retiree as it explains more detail about how to construct the TIPS bond ladder, and separately the article provides details about how a financial company may go about packaging and selling this product.  The only difference for the new article is that it talks about making this strategy the benchmark whereby to consider alternative strategies.  The numbers in each article are also a bit different because Prof. Shankar assumes the strategy is implemented in April 2008, whereas Sexauer et al. assume an implementation date of September 2010.

That being said, what is the strategy being discussed, and is it a good idea?

The other day I included this picture at my blog:

Both articles start with a premise that inflation-adjusted annuities (SPIAs) will be non-starters for most people. Retirees seem to just not want to give up control of their assets for the purposes of getting a guaranteed income.

So the objective of the articles becomes: how can you get guaranteed income with as little annuity as possible?  Also, both articles look for strategies which are much safer than the diversified portfolios of classic safe withdrawal rate studies.  As Sexauer et al. say in a footnote, "we do not want to minimize the probability of shortfall -- we want to eliminate it."
In looking for alternatives, these articles move us into the center circle of my diagram: less inflation protection, but more control over assets. At the September 2010 prices, Sexauer et al. determine that a 65 year old retiree could spend 12% of their assets to purchase a deferred annuity, and then use the other 88% of assets to construct a bond ladder of TIPS to provide precisely 20 years of inflation-protected withdrawals. Anyone then living past 20 years would have no assets to bequest because they've spent the 88% of their wealth already and the deferred annuity represents a non-refundable payment. Sexauer et al. determine that this would support a safe real withdrawal rate of 5.118% over the first 20 years of retirement. Things get more dicey after that.

This is a viable strategy to be considered by retirees.  I'm not so convinced that this strategy is going to be better for most people than using inflation-adjusted immediate annuities instead.  I'm getting closer to the point where I am going to be able to make a detailed comparison between different strategies, and so I don't know the answer yet, but I wonder if something like using 50% of assets for an inflation-adjusted annuity and using 50% of assets in a diversified portfolio may be able to provide more satisfying outcomes for many retirees while still leaving sufficient control over assets to satisfy that aspect of the issue. 

Though I think it is a viable strategy, as you consider retirement income  let me just provide what I think are two important concerns for this particular strategy that you should consider:

1. Extending from the comment made by my wise friend Bob Seawright about my above illustration, this particular strategy really is only providing an illusion of control over those assets used to purchase the TIPS ladder.  It is true that if you die early, the remaining assets can be left to an heir.  But you don't exactly have all that much control over potential emergency needs to increase spending in one year.  You can spend more in an emergency, but the way the TIPS ladder is constructed, it means you are spending what you needed in the future and you are going to have a gap of lower spending until you hit the 21st year when the deferred annuity kicks in. The TIPS ladder is constructed to exhaust your portfolio after 20 years.  You could construct it differently to have more wealth left at the 20 year point for emergencies, but this would obviously lower the withdrawal rate you are able to use.

2. There are two problems with the deferred annuity both related to the fact that it is a nominal annuity. The amount of deferred annuity you buy is such that you are "expected" to have a smooth real income flow from year 20 to year 21. Both articles calculate this using the breakeven inflation rate implied by the difference in yields for 20-year nominal and TIPS bonds.  For Sexauer et al, the assumed 20-year average inflation rate is only 1.91%.  To the extent that the actual average inflation rate over those 20 years may be more or less than that, the amount of spending you get with the deferred annuity is really unclear. For instance, if inflation follows its historical average of 3% over those 20 years, instead of 1.91%, then your spending power will drop by almost 20% in the year that the deferred annuity kicks in. This is a risk you must be aware of.  The other problem with the deferred annuity is that it is a nominal annuity, so it's real value will continue declining after the 21st year of there is inflation. 

And that's my take on this particular retirement income strategy.

Update: Readers interested in this approach should also see the Michael Edesess review of a similar PIMCO-Met Life retirement income product at Advisor Perspectives.


A blog reader requested that I share a link to an article, "10 Reasons You Really Need a Will." That's something I need to learn more about myself.


  1. I completely agree with Wade's comments about these TIPS/deferred annuity combos, and I'll just add a few observations. First, I should note that the annuity part of this combo goes by a number of different names--deferred annuity, longevity insurance, deferred income annuity, deeply deferred annuity, deferred immediate annuity (for jumbo shrimp lovers!), and probably others. Our industry has a jargon problem.

    I've raised questions from time to time about how "annuity averse" people really are--perhaps to some extent, but I have argued in an Advisor Perspectives article last summer that the sales incentives in the products may explain much of the difference in popularity of income annuities vs. products like variable annuities. If people were powerfully averse to income annuities, I would expect there to be much more of a clamor to turn Social Security into a DC plan, and much less protest by public sector workers over threats to their pensions.

    Re: the specific TIPS/deferred annuity combos, it might be worth looking at what a company called Thrive Income does. They do work for actual clients and, instead of just TIPS, look at combining a variety of asset classes with income annuities and deferred annuities. This is a much more general approach and they consider particular client needs and risk appetite. I only know them by reading about them, but their approach sounds intersting.

    1. Joe, thanks for commenting! When I first heard people using the term longevity insurance, it took me a while to realize it was the same thing as a deferred annuity. I just read your Advisor Perspectives column on the issue of who is to blame for the annuity puzzle: sellers or buyers? Very interesting. The link is:

      Thank you, Wade

    2. Actually, longevity insurance can be better characterized as a "deferred immediate annuity." This is a term coined by Harold Evensky. Unlike deferred annuities, you lose even the premium you payed for longevity insurance if you die before benefit commence. And without riders available on some longevity products, once you pay the premium and set the date that benefits will begin, you are locked in. The restrictions allow policy holders to receive more benefit from mortality credits, resulting in higher payout ratios than deferred annuities.

    3. Duncan,

      Thanks for the comment. I find the terminology very confusing, but what myself and these articles call deferred annuities seems to be the same as what you call "deferred immediate annuity." If you just wait until later to buy an immediate annuity, I guess I would just call that waiting until later. This "deferred" word is vague, since it is not clear whether you defer making the purchase or whether you make the purchase now but defer the beginning of the payments. I'm using it in the latter sense. So then, that is the same as longevity insurance, or is there still something different about longevity insurance?

      Thanks, Wade

    4. Thanks for clearing that up. The terminology has also been very confusing for me. I agree, the word "deferred" for a "deferred annuity" is very vague. I've been wondering about what a "deferred immediate annuity" was. It's nice to have that cleared up.

  2. Another great post, Wade. Thanks for keeping me up to speed!


  3. Is the expected failure rate of a SPIA purchased from a private company over a 30 years considered to be zero?


    1. LH, certainly not. But if you mean... is that an assumption in this research... I think so. I can understand that buying a 20-year deferred annuity is a little scary as there is 20 years for the company to disappear before payments begin.

  4. Good tips and tricks. Everybody wants to have a perfect 'safe' retirement income plan but a big is 'How'? Thanks for answering it.